ArticlePDF Available

Simulation Of Trust-Based Mechanism For Enhancing User Confidence In Mobile Crowdsensing Systems

Authors:
  • Murang'a University of Technology

Abstract and Figures

With the rapid development of mobile technology and subsequent mass adoption of mobile devices, mobile crowdsensing (MCS) has gained a lot of research attention. In MCS systems, trust is a key focus in the overall improvement in the participant uptake of the sensing tasks. The trust-based scheme of MCS is studied to predict the damage level, the scores of quality-of-service (QoS), and the levels of quality-of-data (QoD) of MCS systems. Users can participate in MCS sensing based on trustworthy indicators that are related to user experience and system reputation, as well as the knowledge obtained about the MCS systems. This paper illustrates the establishment of user confidence during recruitment in MCS as it is very critical for the success of MCS systems and proposes a simulation trust-based mechanism (SiTBaM) approach. The level of MCS security is enhanced to protect the privacy of participants, so that participants can be assured that the MCS system they are working with during sensing moment is trustworthy. The application of SiTBaM in MCS is verified to yield better results as the simulations show that it offers higher QoS levels, QoD scores, as well as low damage levels in the presence of any task or many malicious users. These results were validated through comparisons with other schemes.
Content may be subject to copyright.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 1
Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier 10.1109/ACCESS.2017.Doi Number
SIMULATION OF TRUST-BASED MECHANISM
FOR ENHANCING USER CONFIDENCE IN
MOBILE CROWDSENSING SYSTEMS
Dorothy Mwongeli Kalui*1,2, Dezheng Zhang1, Geoffrey Muchiri Muketha3, Jared Onsomu4
1Department of Computer Science, University of Science and Technology Beijing, Beijing, China
2Department of Computer Science Meru University of Science and Technology-Kenya
3Department of Computer Science Muranga University of Technology-Kenya
4University of Nairobi-Kenya
Corresponding author: Dorothy Mwongeli Kalui (e-mail: dkalui@yahoo.com).
This work was supported in part by the National Research Fund (NRF), Kenya, Multidisciplinary Research Grant 2016/2017.
ABSTRACT With the rapid development of mobile technology and subsequent mass adoption of mobile
devices, mobile crowdsensing (MCS) has gained a lot of research attention. In MCS systems, trust is a key
focus in the overall improvement in the participant uptake of the sensing tasks. The trust-based scheme of
MCS is studied to predict the damage level, the scores of quality-of-service (QoS), and the levels of quality-
of-data (QoD) of MCS systems. Users can participate in MCS sensing based on trustworthy indicators that
are related to user experience and system reputation, as well as the knowledge obtained about the MCS
systems. This paper illustrates the establishment of user confidence during recruitment in MCS as it is very
critical for the success of MCS systems and proposes a simulation trust-based mechanism (SiTBaM)
approach. The level of MCS security is enhanced to protect the privacy of participants, so that participants
can be assured that the MCS system they are working with during sensing moment is trustworthy. The
application of SiTBaM in MCS is verified to yield better results as the simulations show that it offers higher
QoS levels, QoD scores, as well as low damage levels in the presence of any task or many malicious users.
These results were validated through comparisons with other schemes.
INDEX TERMS mobile crowdsensing, quality-of-Data, quality-of-service, security, trust-based scheme
I. INTRODUCTION
Mobile crowdsensing (MCS) has attracted significant focus
in the recent past making it an appealing paradigm in the user
communication sensing systems. The MCS system is a
human-driven Internet of Things (IoT) service empowering
citizens to observe the phenomena of individual, community,
or even societal value by sharing sensor data about their
environment while on the move [1]. There is an emerging
human-powered modern sensing paradigm that leverages
millions of individual mobile devices to sense, collect,
analyze urban data without the deployment of any large
number of static sensors as sensing infrastructures thus
making it low cost and of spatial-temporal coverage [2], and
this fits the category of MCS systems. MCS relies on
contributions from mobile devices (i.e. smartphones, tablets,
iPads, and wearable devices) [3] of a large number of users
or crowd. Smartphones, tablets, iPads, and wearable devices
are equipped with a rich set of sensors and deployed widely
making them an excellent source of information. This new
paradigm is more scalable and cost-effective than deploying
static wireless sensor networks for dense-sensing coverage
across large geographical areas [4]. Basically, MCS
applications focus on community sensing tasks for large-
scale phenomena that cannot easily be measured by a single
individual. Rather, these phenomena can only be measured
accurately when data are aggregated spatiotemporally from
many individuals [5].
MCS system is different from conventional sensing
solutions because it is powered using specialized networks
of sensors aimed at leveraging human intelligence to collect,
process, and aggregate sensing data using individuals’
mobile devices (e.g., using a camera to capture a specific
target), so as to realize a higher quality and more efficient
sensing solution [6]. The intelligence of humans together
with the mobility aspects will guarantee a larger coverage
and better context awareness if compared to the traditional
sensing networks. However, the participants may be
reluctant to share data that they deem to be of sensitive nature
due to privacy concerns. Mobile crowdsensing is a new
sensing paradigm that incorporates built-in sensors of mobile
devices and human intelligence to monitor, share, analyze
big and heterogeneous data about diverse phenomena [7]. A
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
typical MCS system consists of a cloud-based platform and
a large number of mobile devices or more commonly, the
smartphone users, where the platform works as a sensing
service buyer who posts the required sensing information
and recruits a set of mobile device users to provide sensing
services or to participate in the sensing campaign [8]. When
the participant is once selected by the platform, he starts to
collect the required data and sends it back to the requesting
platform. The requester initiates a crowdsensing application
that usually needs to have monetary investment so the
inferred truths can be the data provided by mobile
crowdsensing is used to design a variety of applications
according to individual or group activities to model their
behaviour and predict possible solutions for different
patterns.
MCS technology has attracted much attention since this
technology can perform sensing jobs that individual users
cannot cope up with. In the case of participatory
crowdsensing users, they can collude with each other to
mislead the system by sending fake information since they
own and control the devices used for MCS as these users may
have unknown intentions, varied capabilities and
unpredictable reliability which leads to untrustworthy data
[9]. The participants or mobile users are registered as
candidate workers to collect and contribute data through
their sensing devices [10]. When a new task arrives, the MCS
server selects some workers to complete these tasks but
results in some issues in task allocation since the various
participants possess diverse qualities on handling different
tasks, hindering efficiency as it solely depends on the
location information to calculate the distance between tasks
and workers. If large distance exists between the target
location of a task and the participant, then there will be
greater rewards for completing the task unlike when the
distance is short. In [11], it is argued that many participants
are usually reluctant to participate in MCS campaigns either
because of fear of their privacy, or because of resource (e.g.
smartphone battery, and memory) consumption, thus making
many researchers rely on voluntary participants.
In the case of large-scale MCS deployments, massive
computational resources are required for device management
and real-time data processing. Despite this challenge of
massive resource requirements, large-scale and centralized
MCS introduces other problems including
generates significant load on mobile network
creates increased traffic to cloud servers running MCS
services,
high computational cost, associated with real-time usage
scenarios, due to a large number of devices participating
in MCS tasks with frequently changing context,
increased latency of data and information propagation,
which is critical for real-time usage scenarios, and
a threat to user privacy since all user traces are collected
in a centralized manner.
The mobile edge technology is designed to enable third
parties to run their services and applications at the edge of
the mobile networks so that they can reduce these problems.
MCS is gaining a lot of familiarity in this era of mobile
technology. According to Wikipedia, mobile crowdsensing
is a mobile data-gathering technology where a large group of
individuals who have mobile devices capable of sensing and
computing collectively at the same time can share data and
extract information to measure, map, analyze, estimate, or
infer any processes of their common interest. This technique
can also be summarized to mean crowdsourcing of sensor
data from mobile devices which can be largely dispersed
from each other. A number of individuals, forming the
crowd, is committed to performing observations of real-
world phenomena of common interest through the use of
mobile phones, given their capacity to sense the environment
and other phenomena in the community (e.g. finding the total
number of people in a restaurant, or in a cinema hall given
their GPS position) [12].
Currently, smart mobile devices are becoming
increasingly popular everywhere and are equipped with very
powerful sensors that have been pervasively applied in
crowdsensing as effective tools to solve large-scale sensing
tasks in urban areas. The group owner (GO) performs the
coordination work by establishing contracts with mobile user
devices to specify the expected results and their
corresponding incentive payments [13]. These incentives can
take various forms i.e. presence/location-aware,
behavioural-aware, flat incentive, mobility-aware, and
mixed incentive [14]. The task requesters can allocate
sensing tasks to the mobile nodes through a crowdsensing
platform, eliminating the cost of deploying and maintaining
large numbers of fixed sensors [15]. However, several kinds
of crowdsensing tasks like audio, visual, or audio-visual
transmission which generates large-scale sensed data may
bring high network traffic costs to participants using a 3G,
4G and 5G networks thus affecting their satisfaction. With
the increased number of smartphone uptake which stands at
over 4.5 billion gadgets, human mobility patterns and daily
actions have increased tremendously with a possibility of
many participants taking part in MCS data collection in a
passive or opportunistic manner [16]. The advantages and
disadvantages of MCS are summarized in Table 1.
TABLE I
ADVANTAGES AND DISADVANTAGES OF MOBILE
CROWDSENSING (MCS)
Advantages
Disadvantages
More scalable,
Low deployment cost,
Large scale,
Crowd carrier mobility,
Fine-grained measurement,
Crowd powered data collection,
Spatiotemporal coverage,
Crowd powered data analysis,
Flexibility,
Savings on computational
resources,
Improved service quality,
High predictive model accuracy,
Management facilitation.
High consumption of energy,
The optimization problem of
user selection,
Takes a long time in
recruitment,
Installation/maintenance cost,
Lack of scalability,
Insufficient spatial-temporal
coverage.
Sources: [4][17][18][19][20][21]
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
This paper illustrates the evaluation of SiTBaM, a
simulator for MCS trust. The SiTBaM is specifically
designed to perform analysis and evaluation of trust in
diverse environments under MCS campaigns, and support
hybrid sensing paradigm. This simulation platform can
visualize the obtained results of trust. The rest of the paper is
organized as follows. A basic overview of the types of MCS
schemes, and challenges, opportunities, and solutions are
provided in Section 2. Details about the MCS architecture
are provided in Section 3, followed by the discussion on
privacy preservation and trust management of MCS in
Section 4. Section 5 provides the details on MCS systems
simulations and discussion. Finally, the conclusion is
provided in Section 6.
II. MOBILE CROWDSENSING
Mobile crowdsensing is gaining a lot of familiarity in this era
of mobile technology. According to Wikipedia, mobile
crowdsensing is a mobile data-gathering technology where a
large group of individuals who have mobile devices capable
of sensing and computing collectively at the same time can
share data and extract information to measure, map, analyze,
estimate, or infer any processes of their common interest.
This technique can also be summarized to mean
crowdsourcing of sensor data from mobile devices which can
be largely dispersed from each other. According to [12], a
number of individuals, forming the crowd, is committed to
performing observations of real-world phenomena of
common interest through the use of mobile phones, given
their capacity to sense the environment and other phenomena
in the community (e.g. finding the total number of people in
a restaurant, or in a cinema hall given their GPS position).
There are two common types of MCS techniques namely,
participatory MCS, and opportunistic MCS [3]. In
Participatory MCS paradigm, the user is actively involved
and is aware of the sensing through the use of the front end
applications and actively reports observations while in the
case of opportunistic MCS, the user involvement is
minimized and in some cases, none and often, an application
can be running in the background which performs sensing
and monitoring tasks with minimal or no user intervention.
However another type of MCS termed as hybrid MCS can
also exist, which harvests the benefits of both methods,
making the number of MCS methods to be three [22]. For the
task of sensing, the built-in and ubiquitous sensors of the
smartphones are used either in a participatory, or
opportunistic way depending on whether the data collection
happens with or without participant involvement [7].
The mechanism of task allocation models is based on
human Involvement; knowledge available to service
providers (SP); and spatial distribution. In so doing the MCS
tasking entities are responsible for assigning tasks to carriers,
via the task assignment models. The MCS contains four
stages in its life cycle namely: task creation, task assignment,
data collection, and data aggregation shown in Fig. 1.
Task
Creation
Data
Aggregation
Data
Collection
Task
Assignment
FIGURE 1. Life-cycle of MCS
A.
PARTICIPATORY TECHNIQUE
This is a type of MCS system where the participant is
actively involved in the collection of data. This can include
the case of photography, and filling in questionnaires. This
means that the users are self-aware about sharing data with
the other users in participatory sensing mode of collecting
data. The participants use their own mobile devices to
complete the task by collecting the data and giving the
feedback regarding the results[23], and users prefer to have
control over what and when to participate in a sensing
campaign. This method is prone to adversarial attacks
because a malicious node could easily send false information
to a service provider, and it thrives from continuous input
from the user [24] as the user voluntarily participates in the
contribution of information. Here sensor data collection is
triggered by tasks, which specify the sensing modalities like
regions of interest, and sampling context based on
application requests. The tasks are distributed to mobile
device carriers that satisfy the tasking requirements, and
people can decide to accept or refuse the task allocated. We
can find that data is collected under the “primary use”
manner in explicit sensing. Privacy in explicit sensing should
guarantee that participants maintain control over the release
of their sensitive information, for example, the degree of
granularity and data recipients.
B.
OPPORTUNISTIC TECHNIQUE
In the case of an opportunistic MCS system, the sensor data
is acquired autonomously and reported to the cloud
periodically without the user involvement [19]. In this
method, mobile devices are involved in the process of
decision making instead of the users as is the case of
participatory crowdsensing [7]. In Opportunistic Sensing,
users unconsciously participate in tasks, and their devices
can complete the task without human help e.g. as long as a
user turns on WIFI, the task of “sensing WIFI signal and
strength” is completed without people’s intervention [15].
This scheme has less reliance on active user involvement in
the process of sensing and sending information, as the data
is sensed and sent automatically. This whole process takes
place via the portable sensors that accompany the user
participant. These portable sensors can be grouped as mobile
sensors, body sensors, or vehicular sensors respectively [22].
Thus this method is majorly concerned with the passive
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
extraction of mobile sensor data, and it aims at keeping the
user involvement to the minimum [24]. In this scheme data
is contributed not for a sensing task, but for users to enjoy
online services like socializing on Facebook or Wechat,
purchasing goods on Amazon or Alibaba, etc. in this scheme,
also we get that the data is reused to enhance original
services or create new services by third parties i.e., used for
a second purpose.
C.
HYBRID TECHNIQUE
The hybrid MCS technique is collectively the best
crowdsensing system as it incorporates the benefits of the
former two methods of sensing. In this, the mobile source
nodes apply active sensing mode (agreeable participation in
data forwarding) and passive sensing mode (via opportunistic
node interaction) in the network. This is an improved
crowdsensing data collection method, as it improves the
accuracy of the data collected since some of the data that the
user may not have been willing to divulge can easily be
collected opportunistically. This method is applicable to
indicate a smooth switching and collaboration between
participatory and opportunistic models to overcome the
disadvantages of both approaches.
III. ARCHITECTURE OF MOBILE CROW-DSENSING
SYSTEM
In this section, we outline the various entities of MCS system
architecture. Like mobile cloud computing, MCS is
relatively a new technology with lots of potential
applications but no agreed-on standard architecture exists to
date [25]. This motivates this research for new and
innovative architecture to suit this review that tries to
emphasize on certain requirements such as privacy, cost,
mobility, delay, and power consumption while trying to
select the optimal compromise for the other. The architecture
of the cloud-based mobile crowdsensing system consists of
a cloud-based platform and a large number of smartphone
users [8] is shown in Fig.2. From Fig. 2, MCS architecture
contains the four major components: sensors, mobile
devices, communication infrastructure, and processing
infrastructure [26]. These components can further be
grouped into two that is the mobile data collection
components, and the web-based data server [27]. The sensors
and mobile gadgets form the mobile data collection subset,
and the web-based data server is composed of the
communication and processing infrastructure of data.
Sensors: The work of sensors is data collection from the
environment. It is cost-effective to use a mobile
crowdsensing system technique for collecting data because
it does not need specialized sensors installed everywhere,
which ultimately reduces procurement, installation, and
maintenance costs. The participants are equipped with the
necessary hardware, software, and knowledge of the
application to start gathering data as they conduct their daily
life. The heterogeneous sensing capabilities pose
fundamental challenges to MCS systems by affecting their
two main operations truth discovery and reward distribution.
The truth discovery refers to the process of aggregating and
analyzing the crowd-sensed data to estimate the ground truth
[28], while a reward distribution scheme is required to
reward participants according to their effort levels in truth
discovery process [13]. The task of sensing on the phones
can be triggered manually, automatically or based on the
current context.
FIGURE 2. Cloud-based mobile crowdsensing architecture
Mobile Devices: the task of mobile devices is to aggregate
and report the collected data to the cloud server or the group
owner. The subjective inputs (participants) use their mobile
devices to insert data into the system about their daily
locations and assessment of the phenomena. The group of
mobile device users who are participating in the sensing
campaign sense their surrounding environment in response
to a sensing request initiated by an agent, referred to as a task
publisher. Task publishers may represent machines or people
and be involved in a sensing task that is time-consuming
from the participant's perspective as well as consuming the
device's sensing, computing, and communication resources
[29].
Communication infrastructure: the communication
infrastructure is basically the data transport part tasked with
the transmission of the data from the participants to the
server system. The secure crowdsensing model considers the
privacy of user's data, restricts unauthorized access and
enhances the quality of service in the network.
Processing infrastructure: The processing infrastructure is
tasked with the data storage role mostly, but also does the
estimation, and inferring of the data. This layer transforms the
collected low-level, single-modality sensing data into the
expected intelligence through the application of the machine
learning, and logic-based inference techniques [17]. The
results of the processed sensed data may be displayed locally
on the carriers’ mobile phones or accessed by the larger public
through web-portals depending on the application needs.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
In MCS management there are several attributes that are
taken care of together with some operations to be performed.
MCS system must ensure that it guarantees security, privacy,
and trust to the service requester, participants, and mobile
users, when it is performing its role of user recruitment, task
creation and execution, and incentivizing. Ref.[30], outlines
the main common incentive mechanism frameworks with their
key elements as applicable in MCS which includes an auction,
lottery, bargaining game, contract, market-driven, and trust
and reputation. Of these incentivizing mechanisms the auction
incentive is the most widely used of them all. MCS platform
architecture is shown in Fig. 3.
Roles
User
recruitment
Task creation
Task
execution
Incentivizing
Features
Security
Privacy
Trust
Integrity
Reliability
.
.
Availability
MCS
Manage
ment
Participan
ts
Mobile users
Service
Requester Request
Response
Recruitme
nt & Task
allocation
Sensing
data
Incentives
FIGURE 3. The MCS management platform
IV. PRIVACY AND TRUST MANAGEMENT
The number of data users and participants involved in the
process of data collection is growing thanks to technology.
One of the key challenges posed is privacy preservation in
data mining which has emerged as an absolute prerequisite
for exchanging confidential information in terms of data
analysis, validation, and publishing. There is an existence of
ever-escalating internet phishing threats on the widespread
propagation of sensitive information over the web. Equally,
the dubious feelings and contentions mediated unwillingness
of various information providers towards the reliability
protection of data from disclosure often results in utter
rejection in data sharing or incorrect information sharing.
Furthermore, workers in MCS are also heterogeneous on
their privacy concerns, which makes privacy preservation
even more challenging. They mainly differ in (a) ratio of the
private locations along their paths, (b) extend they start to
treat their locations as disclosed, (c) compensation amount
asked for the partial disclosure of private locations [31].
Various privacy-preserving mechanisms have been put in
place to enhance the privacy preservation of the participants.
According to Ref. [32], ensuring privacy-preserving in MCS
system encourages mobile users to use MCS system
applications and participate in sensing and data collection.
Therefore, a properly designed framework for privacy
preservation must support the workers to flexibly adjust on
all three aspects, so as to provide most suitable participation
for every worker as this is a critical principle as MCS
essentially rely on these workers for data collection.
Another important issue in MCS campaigns is the
trustworthiness or reliability of the data collected. Trust is a
great issue as tasks are assigned anonymously and data is
collected from multiple locations or participants which can
be in one or another compromised leading to unreliability of
the collected data [33]. It is crucial to maintain a high level
of data trustworthiness which shows how much the data used
are trusted, authentic and protected from abuse so that
decision making should be based on precise, and certain data
[34]. A user who sends altered data can get his
trustworthiness degraded once detected since the MCS
system computes trustworthiness, and always it runs an
outlier detection algorithm to detect any form of data
degradation [35]. One of the factors that can lead to data
unreliability is the existence of uncertain factors like the
channel or surrounding noise and the difficulty of target-
sensing. At the same time, the participants can be deceived
by scammers or the participants themselves can be
participating in the MCS campaigns with malicious
intentions thus supplying falsified data since it is sometimes
difficult to identify them, especially when different task
actions are not liked due to privacy protection. As
participants can be participating in MCS campaign
anonymously making their identities undisclosed due to
privacy issues, the participants who provide falsified
information cannot be identified or eliminated more so in
opportunistic crowdsensing. Research has it that, if the same
task is assigned to multiple participants simultaneously it can
reduce the effect of malicious participation but on the hand
this consumes a lot of resources. This attribute of assigning
the same task to multiple users with zero rewards cannot
guarantee QoS of each task to be no less than a given
threshold [36]. This, therefore, poses a great challenge to
create a balance between user privacy, QoS and the
trustworthiness of the data or the source.
A.
USER RECRUITMENT IN MCS
The malicious attackers aim to destroy the functionality of
cooperative spectrum sensing so that the system cannot trust
the aggregated sensing results. This research considers four
types of spectrum sensing data falsification (SSDF) attacks
to test the resiliency of the proposed data aggregation
scheme: “always yes”, “always no”, “always false,” and
“always random.” Under the always yes attack scenario, the
malicious secondary users (SUs) always report the presence
of primary users (PUs) ignoring their real sensing results.
Under the always no attack scenario, the malicious SUs
always report the absence of PUs on the channel ignoring the
real detection results. Under the always false attack scenario,
the malicious SUs always report the opposite of their sensed
outcomes. Under the always random attack scenario, the
malicious SUs randomly generates a sensing result to report
to the system [37]. There is no mechanism to control the
behaviours of the SUs and this threatens the security of the
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
licensed users. The major behaviours of the attackers can
include but not limited to misbehaving, selfishness, cheating,
and malice. An attacker can misbehave and this is the
severest category of attacker behaviours as if a node
misbehaves and can apply any of the other categories,
decreasing the performance of spectrum sensing and sending
false information to prevent other nodes from utilizing the
spectrum [38]. The cooperative incentive mechanism is
applied in crowdsensing participant recruitment, to try and
model the cooperation between participants [39].
There arise two social dilemmas however when it comes
to pricing schemes in MCS where either the requester pays
rewards to participants before or after executing the task. If
payment is made before task execution, participants may pay
less effort in the sensing task, and if payment is made after
execution of the task, the requesters may lie about the quality
of sensing data in order to not to pay or to pay lee, rendering
the whole exercise of participant recruitment to be futile due
to false reporting schemes.
B.
PRIVACY MANAGEMENT
There are various technologies that have been researched on
and are in use for the application of privacy preservation in
MCS systems applications. These methods sometimes
require to be used to complement each other. Due to the
diverse qualities of users on different tasks, task allocation is
critical to all MCS systems, the efficiency of which depends
mostly on the participant location information to compute
the distances between tasks and workers. The longer the
distance between the target user to the target location of the
task, the greater the reward of completing the task, the
shorter the travel distance, and the more likely that the user
will accept the task and the fewer rewards the crowdsensing
server will pay [40]. However, the location information may
fall in the hands of an untrusted CS-server as well as the
incentives, concerns about privacy leakage and security
threat will discourage users from engaging in MCS. Thus,
location privacy preservation should be jointly taken into
account in MCS task allocation. Demands for the location
privacy preservation and the task completion rate when
developing an optimal task allocation method must be
considered in the design of the MCS systems. There are
various techniques that can be applied to ensure that the
privacy preservation is maintained.
Encryption: Encryption is the process of decoding
information thus rendering it meaningless to any
unauthorized users or systems. This technique can aggregate
the private data of Mobile Device Owner (MDOs) without
revealing MDOs individual data records [8]. Each
participant in the sensing campaign has to obtain an
encryption key to cipher his collected data and the encryption
key should be known to the sensing service buyer to decipher
the data. This mechanism requires a lot of computation
power and resource energy which makes it sometimes
unsuitable for most crowdsensing applications, particularly
when it is deployed on energy-constrained mobile devices.
This scheme of privacy preservation in MCS is required to
secure data fusion while guaranteeing traceability, as MCS
requires to balance privacy preservation against user
reliability. Thus the MCS system must ensure that the data is
always kept encrypted before they are transmitted to the
authorized entities in the system, and remains encrypted in
the system.
Anonymity: In the privacy preservation of the data and the
participants in the sensing campaign, anonymity is employed
in data collection and uploading of information. However, to
ensure that scrupulous and malicious participants do not take
advantage of this scheme, the Trust Authority (TA) can infer
the true identity of a given participant, given the anonymity
of the participants [18]. The pseudonym-based methods will
offer anonymity to the MDOs, but they also bring significant
cost and potential risk since a user may use a pseudonym for
a while and drop it and switch to a new one, and increases
processing time as the anonymity process gets more strict.
But for better privacy, it is prudent to employ a policy of
short-lived and frequently changed pseudonyms for better
privacy although the privacy can again be compromised if
any of the neighbors involved in the pseudonym exchange
are attackers. Sometimes it is hard to achieve anonymity
when both location and reputation are being incorporated
into the participant's query, and if full anonymity is provided
to the users, guaranteeing the trustworthiness of the reported
data is impossible.
Authentication: When participants are assigned tasks in a
sensing campaign, sometimes it is good to use a mechanism
of identifying the authenticity of the participants by ensuring
that the assigned task includes information about the task
description, the location of the task, the finish time of the
task, and the status of the task as well as the user identities
[41]. This method ensures that the participants do not submit
falsified information into the system which will compromise
the trustworthiness of the system, although it is seen as
compromising the privacy of the users on the other hand.
Therefore the MCS system is responsible for the access
control, by giving the necessary rights to authorized users
[32].
Non-repudiation: No participant should be able to change its
mind (e.g. deny or modify its data) once the data has been
submitted [42]. This implies that once a participant wishes to
take part in a sensing campaign, they are bound by the data
they submit although this is not the case for opportunistic
crowdsensing data collection.
Data aggregation: Data aggregation is a widely used
technique in wireless data networks. The data aggregation
algorithms are designed to gather the data and aggregate the
data to enhance the network's lifetime. This is a mechanism
where participants tend to distribute their collected data
among their neighbours, and Ref. [22] terms it as the process
of integrating data from multiple users into one message.
When a participant receives a request from the aggregation
server, each participant returns his data and the remaining
data of his neighbours, thus reducing the probability to
successfully attribute each sensor reading to its
corresponding mobile user. Performing crowd sensing with
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
the help of many individuals leads to the collection of a large
amount of data, necessitating data aggregation and
processing to converted data into high- level information
before being utilized by users and systems [22], which is
better for decision making than when it is from a single
source [17]. But an MCS service has to control the data
production process since mobile devices typically support
only limited filtering and aggregation mechanisms and often
deliver all raw readings to the cloud [1].
Confidentiality: various mechanisms should be put in place
to ensure that the confidentiality of the user’s information
and weights in crowdsensing systems is maintained. The
confidentiality of observed values or user’s sensitive
information collected by the cloud server (such as health
data, location, address, etc.) should be protected and
prevented from disclosure to other parts like to other users,
the cloud server, and any attacker [43]. For example,
aggregating health data, such as treatment outcomes, can
lead to better evaluation of new drugs or medical devices’
effects but may leak the privacy of participating patients.
Thus confidentiality is the technique of ensuring that the data
that is in storage or in transit is encrypted to conceal its
contents and only the data owners know the plaintext.
Verification and Validation: verification and validation refer
to the authentication and confirmation of the participant
identity in the MCS network before the participants can take
part in the sensing campaign. A few steps of verification,
checking and anonymization through the different
components can be employed to provide a higher level of
privacy to the participants [42].
C.
TRUST MANAGEMENT
The MCS system can be said to be trust-worthy if the user
feels safe to use, and also trusts to execute tasks without
secretly executing any harmful programs, and trust
management is one of the factors that affect performance and
lifetime of MCS [44]. The trust evaluation model calculates
the trust value based on the user’s communication behaviour.
The presence of malicious trustees in the system is notified
to the trustor in the MCS system. To enhance system
trustworthiness, it is critical for the trustor to recruit users
based on their personal features, e.g., mobility pattern and
reputation, although it leads to the privacy leakage of
participants [45]. The sensing data collected from the
surrounding areas are necessarily people-centric and related
to some aspects of mobile users and their social settings:
where they are and where they are going; what places they
are frequently visited and what they are seeing; how their
health status is and which activity they prefer to do. Social
event photos may expose the social relations, locations or
even political affiliations of mobile users.
The spatial data collected by the carried devices might
disclose mobile users’ trajectories. For example, Google
Maps collect the “anonymous” location information of
drivers for real-time traffic map generation but still exposes
the driving routes and trajectories of drivers. Further, the
more sensing tasks the mobile users are engaged in and the
richer data the users contribute to, the higher probability that
their sensitive information may be exposed. Therefore,
preserving the privacy of mobile users is the first-order
security concern in mobile crowdsensing. If no effective
privacy-preserving mechanism is on-shelf, it is of difficulty
to motivate mobile users to join in mobile crowdsensing
services.
There are three forms of trust in MCS: direct trust, indirect
trust, and comprehensive trust. Direct trust involves issues to
do with the knowledge obtained about the MCS system as
direct observation. The attributes of direct trust include
ability, integrity, availability, reliability, similarity, and
security. On the other hand, the indirect trust involves issues
revolving around the experience and the reputation gained
over time about the MCS system. The indicators used to
evaluate the direct trust of an MCS includes interactions, past
related experiences, and relationships. The experience is
constructed from the interactions between two entities, while
the reputation is constructed from all the experiences
towards an entity [33]. The comprehensive trust is the one
that incorporates the features of all the former two types of
trust.
According to Ref. [46], the size of the trust value respects
the performance of the node. The malicious node always
leads to declining trust value because of its bad
communication behavior, while the normal node is the
opposite. In this article, the sensor nodes monitor the
communication behavior of their neighbors to detect whether
there is any packet dropping or packet tampering. As shown
in Fig. 4, node evaluates the trust value of node, where
and are the common neighbours of the node
and.
k1
k2
km
x
y
Direct trust
Indirect trust
FIGURE 4: Trust components
Direct Trust: The research by Ganeriwal et.al [47], proposed
a trust evaluation model that utilizes the Beta distribution to
evaluate trust and proved that the trust values obey the Beta
distribution tendencies. The direct trust value  of node
to can be obtained by:
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
    
 (1)
where  denotes the number of cooperative interaction,
and  denotes the number of non-cooperative interactions
among the nodes and .
The original Beta-based trust evaluation model does not
consider the impact of external factors on the communication
interaction among the nodes, such as the packet loss caused
by network congestion. This problem can be solved by the
introduction of an abnormal attenuation factor to improve
the original model. The abnormal attenuation factor is the
probability of malicious attacks which is represented below:
  
 (2)
Where is the number of node non-cooperative
interaction caused by malicious attacks, and  is
the total number of node non-cooperative interactions. The
attenuation of the number of non-cooperative nodes detected
by node to can reduce the influence of external factors
on the trust value. The accuracy of trust evaluation is
improved compared with the original model and the formula
becomes:
    
 (3)
Indirect trust: to improve the value trust accuracy it is
necessary to obtain the indirect trust of the node from the
common and adjacent nodes between nodes and. The
expression for an indirect trust of neighbor node to node
is:

 (4)
To filter all false evaluations from malicious nodes, all
indirect trusts collected from adjacent nodes need to be
disposed to exclude the false evaluations which are above
the deviation threshold . The deviation degree of
indirect trust is:


 (5)
If the degree of deviation of indirect trust is greater the
, the indirect trust is dropped so that the false
evaluation of malicious node can be dealt with.
Comprehensive trust: In the process of evaluating trust,
apart from considering the direct trust, the indirect trust a
combination of these two forms of trust can be performed
so that comprehensive trust value of node to node can
be obtained as shown below:
   

 (6)
where is the weight of direct trust and in this research,  
. Therefore the direct, indirect, and comprehensive trust
in MCS can be employed depending on the scenario to be
evaluated that can guarantee the best results.
D.
CHALLENGES OF PRIVACY AND TRUST
MANAGEMENT IN MCS
There are many challenges when it comes to MCS
technology. One of the major challenges is how to balance
anonymity in safeguarding user privacy while maintaining
the reliability of the data and/the source. This is because
mobile crowdsensing has become a popular paradigm to
collaboratively collect sensing data from pervasive mobile
devices, and since the devices used for mobile crowdsensing
are owned and controlled by individuals with unpredictable
reliability, varied capabilities, and unknown intentions, data
collected with mobile crowdsensing may be untrustworthy
[9].
TABLE II
CHALLENGES, SOLUTIONS, AND OPPORTUNITIES OF PRIVACY AND TRUST MANAGEMENT IN MOBILE CROWDSENSING
Challenges
Opportunities
Source
Incentivizing human user,
Quality of sensed data,
Reliability of sensed data,
Participant selection,
Trust level discrimination,
Security and privacy,
Big data processing,
Heterogeneous vendors of sensors,
Incompatible design and types of
sensors,
Requirements of coverage quality,
Participant recruitment,
Hybrid networking,
Satisfying the requirements of incentive
mechanism,
Varied user grouping,
Cross community sensing and mining,
Requirements of data timeliness,
Energy consumption of mobile sensing
devices.
Prediction,
Real-time data delivery,
Efficient data collection,
Enables visualization,
Spatial sampling diversity,
Influence maximization,
Improved anywhere anytime user
engagement,
Bridge of the gap between space and
physical space,
Cross-space data mining.
[48][4]
[11][14]
[16][17]
[18][20]
[49][26]
[50][51]
[29] [52]
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
Credibility improvement of the data supplied by the MCS
system is another challenge. This is because MCS systems
are subject to collusion attacks where a group of malicious
users can collaboratively send fake information to mislead
the system [29]. Defending the data credibility requires
strong defense mechanisms to curtail the collusion of
participants. Ref. [39] proposed a two-phase group-buying
based auction mechanism for recruiting workers in MCS,
which makes it hard for participants to know each other and
maybe even supply wrong sensing information.
The concept of employing users and devices to collect data
from the real world poses significant social and
technological and economic challenges, solutions, and
opportunities. From the social point of view, if users of the
MCS systems are not motivated, they can provide unreliable
data which will not be meaningful. Various methods of
motivation can be applied like incentivizing participants, and
also ensuring their privacy. Some technological challenges
which can emerge include issues like compatibility of
hardware.
TABLE II presents a summary of the various MCS
challenges, solutions, and opportunities. As depicted in the
table there are a number of challenges since MCS technology
in a new and emerging area still that requires to be
researched.
V. SIMULATION RESULTS AND DISCUSSIONS
In this research, the new approach was proposed for SiTBaM
used for evaluating trust in MCS paradigm. The SiTBaM
simulation setup parameters are as shown in table III. From
the table, the quality of data values was set at 0 to 1, which
implies that 0 is least QoD, while 1 is highest QoD. The
number of tasks was varied from 100 to 1600 in multiples of
400, and their corresponding damage level versus the
percentage of malicious users recorded as demonstrated in
the simulation results. TABLE III
SIMULATION SETUP FOR SiTBaM
Parameter
Value
Number of user ranges
100-10000
Duration of time slot
60 sec
Task duration
1800 sec
Number of task ranges
100-1600
Damage level range
0-50
% of malicious users
0-100
Quality of data values
0-1
Interactions
500
Cooperative threshold
0.6
Uncooperative threshold
0.3
The assumptions of the proposed model in relation to trust
evaluation are:
1. The higher the number of tasks the lower the
damage level
2. The higher the number of malicious users the higher
the damage level.
3. The higher the number of quality users the higher
the probability density.
The Experience Model is normalized in the range [0,1] and
it specifies three trends: Development, Loss, and Decay. The
development trend implies that the cooperative interaction
experience is in the increase, the loss trend indicates that the
cooperative interaction experience is decreasing, while the
decay trend indicates that the experience in the cooperative
interaction is either increasing or decreasing as there exists
neutral or no interaction that is taking place. The decay trend
is the worst as it is unpredictable in regard to the future trend.
In the experience model, the experience between any two
users can be established and updated by the use of an
aggregation model on any virtual interactions. The
reputation of each user can be calculated based on all the
experiences between all users, and a value of trust
relationship is also calculated by aggregating the experience
and the reputation. Therefore to find the trustworthiness of
the system users, the user experience and reputation are very
critical. Due to cooperative interactions the experience
increases and uncooperative interactions cause the
experience to decrease. Consequently if not interactions
occur at all, the experience decays. Therefore the
determining factors for the decrease, increase, or decay of
experience include the intensity of interactions, interaction
scores, and current experience value of the MCS system, as
shown in Fig. 5.
FIGURE 5: Experience model with development, decay, and loss trends
From Fig. 6, a series of interactions was randomly created
where the number of interactions (n) was set at n=500, and
cooperative threshold and the uncooperative threshold was
set at 0.6, and 0.3 respectively. The experience model was
generated as shown with the development, decay, and loss
trends.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
FIGURE 6 Quality of services (QoS) vs, number of requested services
From Fig. 6, the QoS improves significantly when the
number of requested services increases for trust-based
schemes, average-QoD-based schemes, and 3-degree
polynomial regression schemes. However, the QoS remains
the same when the random selection scheme is employed.
All this however from the figure indicates and confirms that
the QoS is higher when the Trust-based scheme is employed.
FIGURE 7. Percentages of Malicious users vs. QoS Score
From Fig. 7, when the percentage of malicious users
increases the quality of service (QoS) decreases in that order.
This means that the presence of malicious users affects the
performance of MCS system QoS, and the measure has to be
put in place to ensure that this existence of malicious users is
reduced to minimal margin to yield better results.
(a)
(b)
(c)
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
(d)
FIGURE 8. (a)-(d) Comparison of Damage levels in Trust-based, vs.
PrevBest-QoD-based, vs. Average-QoD-based Schemes
From Fig. 8 (a)-(d) the damage levels of data values versus
the varying the number of malicious users were compared in
three different schemes Trust-based scheme, Average-QoD-
based scheme, and the PrevBest-QoD-based Schemes. It was
observed that the trust-based scheme provides the least
damage level to the obtained data even after the number of
malicious users is increased. Actually as the number of
malicious users is increased the damage level decreases but
on a marginal value. The PrevBest-QoD-based schemes
record the highest damage level even as the number of
malicious users increases.
FIGURE 9. Comparison of Schemes with varying task numbers
From Fig. 9, the damage level is high when the number of
tasks is lower based on all the three schemes of trust.
Therefore when the number of the tasks is higher the better
the results in terms of the damage caused. And from this,
there is a big marginal gap between the damage levels when
the PrevBest-QoD-based scheme is used compared to the
rest of the schemes.
(a)
(b)
(c)
FIGURE 10. (a) (c) The User Models of different MCS systems
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
From Fig. 10 (a) the QoD range is between the intervals (0,1)
and the highest quality users produced the highest quality-
of-data (QoD) scores in most sensing tasks with the highest
distribution QoD value is 0.95, with a probability density
function (PDF) value of 0.4. The low-quality users produced
the lowest and below-average QoD scores by recording the
highest score of 0.78 with the PDF value of at 0.15. The
malicious users produced the above-average QoD scores of
0.28 and PDF of 0.3.
From Fig. 10 (b) the QoD interval ranges are (0 to 1) and the
highest quality user-produced QoD value of 0.95, at a PDF
value of 8. The low-quality user produced the highest QoD
value of 0.78 with PDF value of 3. The malicious user
accounted for highest QoD value of 0.28, with PDF of 5. This
malicious user recorded higher value than the low-quality
user.
In Fig. 10 (c) the QoD interval remain (0 to 1) and the highest
quality user records the highest QoD value of 0.76 with the
PDF of 3.7. The low-quality user records an above-average
QoD of 0.56 and a PDF of 3.4. However again the malicious
user records highest QoD values than both the high-quality
user and the low-quality user of 0.9, and PDF of 4.7. These
cases in Fig. 10 (a) to (c) demonstrate that the QoD can vary
from system to system and the PDF values recorded vary for
different users (highest quality, lowest quality, and malicious
users). This depicts that the system trust levels are equally
different across the various models.
VII. CONCLUSION
The evolution of mobile devices has led to the vast evolution
of mobile crowdsensing technology, where mobile devices
are used to sense, collect, and transmit information
seamlessly. In this survey, we discussed the overview of
MCS, schemes of MCS, and the challenges, opportunities,
and solutions of MCS. Since the MCS is applicable in almost
all of life scenarios, the applications of MCS in terms of its
importance was discussed. The MCS architecture was
discussed where the MCS framework and the architecture
are highlighted in broad. Since a large number of participants
take part in sensing campaigns, their privacy is of utmost
importance and hence the MCS systems privacy-
preservation and trust management were discussed. This is
very crucial because even if there are incentives, mobile
device users can shy off from participating in sensing
campaigns if their privacy is not guaranteed. The simulations
were based on Trust-based scheme, and it was compared
with other schemes. The results obtained indicate that trust-
based scheme offers the best results when compared to its
counterparts which were discussed as follows.
1. The damage level is lower under the trust-based
scheme even in the event of an increase in the number
of tasks or number of malicious users.
2. The QoS value is high even when the presence of
malicious users tends to increase in number.
3. The QoD scores produced depends on the security
level of the MCS system. The QoD score can be high
for high-quality users, or malicious user, but always
the QoD score is average for the case of low-quality
users.
In summary, the performance of trust-based mechanisms in
privacy and trust management in MCS systems is high, and
more algorithms should be developed to enhance the level of
trust in the MCS application.
ACKNOWLEDGMENT
This work is supported by the National Research Fund
(NRF) Kenya, Multidisciplinary Research grant 2016/2017.
CONFLICT OF INTEREST
The authors declare no conflict of interest for the publication
of this article.
REFERENCES
[1] M. Marjanovic, A. Antonic, and I. P. Zarko, “Edge computing
architecture for mobile crowdsensing,” IEEE Access, vol. 6, pp.
1066210674, 2018.
[2] J. Feng, T. Li, Y. Zhai, S. Lv, and F. Zhao, “Ensuring Honest
Data Collection Against Collusive CSDF Attack With Binary-
Minmaxs Clustering Analysis in Mobile Crowd Sensing,” IEEE
Access, vol. 7, pp. 124491124501, 2019.
[3] C. Fiandrino et al., “CrowdSenSim : a Simulation Platform for
Mobile Crowdsensing in Realistic Urban Environments,” IEEE
Access, vol. 5, pp. 34903503, 2017.
[4] F. Wang, L. Hu, R. Sun, J. Hu, and K. Zhao, “SRMCS: A
semantic-aware recommendation framework for mobile
crowdsensing,” Inf. Sci. (Ny)., vol. 433434, pp. 333345, 2018.
[5] S. Yang, J. Bian, L. Wang, Y. Fu, H. Xiong, and H. Zhu,
“EdgeSense : Edge-Mediated Spatial-Temporal Crowdsensing,”
IEEE Access, vol. 6, pp. 9512295131, 2019.
[6] C. Jiang, L. Gao, L. Duan, and J. Huang, “Scalable Mobile
Crowdsensing via Peer-to-Peer Data Sharing,” IEEE Trans.
Mob. Comput., vol. 17, no. 4, pp. 898912, 2018.
[7] M. Pouryazdan, C. Fiandrino, B. Kantarci, T. Soyata, D.
Kliazovich, and P. Bouvry, “Intelligent Gaming for Mobile
Crowdsensing Participants to Acquire Trustworthy Big Data in
the Internet of Things,” IEEE Access, vol. 5, pp. 2220922223,
2017.
[8] X. Zhang, L. Liang, C. Luo, and L. Cheng, “Frameworks for
Privacy-Preserving Mobile Crowdsensing Incentive
mechanisms,” IEEE Trans. Mob. Comput., vol. 17, no. 8, pp.
18511864, 2018.
[9] T. Zhou, Z. Cai, K. Wu, Y. Chen, and M. Xu, “FIDC: A
framework for improving data credibility in mobile
crowdsensing,” Comput. Networks, vol. 120, pp. 157169, 2017.
[10] M. Arafeh, M. El Barachi, A. Mourad, and F. Belqasmi, “A
Blockchain-based Architecture for the Detection of Fake
Sensing in Mobile Crowdsensing,” in 2019 4th International
Conference on Smart and Sustainable Technologies (SpliTech),
2019, pp. 16.
[11] M. Li, Y. Gao, M. Wang, C. Guo, and X. Tan, “Multi-objective
Optimization for Multi-task Allocation in Mobile Crowd
Sensing,” Procedia Comput. Sci., vol. 155, pp. 360368, 2019.
[12] A. Capponi, C. Fiandrino, B. Kantarci, L. Foschini, D.
Kliazovich, and P. Bouvry, “A Survey on Mobile Crowdsensing
Systems: Challenges, Solutions and Opportunities,” IEEE
Commun. Surv. Tutorials, vol. 21, no. 3, pp. 24192465, 2019.
[13] H. Cai, Y. Zhu, Z. Feng, H. Zhu, J. Yu, and J. Cao, “Truthful
incentive mechanisms for mobile crowdsensing with dynamic
smartphones,” Comput. Networks, vol. 141, pp. 116, 2018.
[14] C. M. Angelopoulos, O. Evangelatos, and S. Nikoletseas, “A
User-enabled Testbed Architecture with Mobile Crowdsensing
Support for Smart , Green Buildings,” in IEEE ICC 2015 SAC -
Internet of Things, 2015, pp. 573578.
[15] Z. Peng, X. Gui, J. An, T. Wu, and R. Gui, “Multi-Task oriented
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
data diffusion and transmission paradigm in crowdsensing based
on city public traffic,” Comput. Networks, vol. 156, pp. 4151,
2019.
[16] S. Bradai, S. Khemakhem, and M. Jmaiel, “Real-time and
energy-aware opportunistic mobile crowdsensing framework
based on people’s connectivity habits,Comput. Networks, vol.
142, pp. 179193, 2018.
[17] B. Guo, Z. Yu, X. Zhou, and D. Zhang, “From participatory
sensing to Mobile Crowd Sensing,” in 2014 IEEE International
Conference on Pervasive Computing and Communication
Workshops (PERCOM WORKSHOPS), 2014, pp. 593598.
[18] G. Sun, S. Sun, J. Sun, H. Yu, X. Du, and M. Guizani, “Security
and privacy preservation in fog-based crowdsensing on the
internet of vehicles,” J. Netw. Comput. Appl., vol. 134, pp. 89
99, 2019.
[19] E. Wang, Y. Yang, and K. Lou, “User selection utilizing data
properties in mobile crowdsensing,” Inf. Sci. (Ny)., vol. 490, pp.
210226, 2019.
[20] H. U. I. Li, D. A. N. Liao, G. Sun, D. U. Xu, and Z. Han, “Two-
Stage Privacy-Preserving Mechanism for a Crowdsensing-Based
VSN,” IEEE Access, vol. 6, pp. 4068240694, 2018.
[21] J. I. A. Xu, W. Bao, H. Gu, L. Xu, and G. Jiang, “Improving
Both Quantity and Quality : Incentive Mechanism for Social
Mobile Crowdsensing Architecture,” IEEE Access, vol. 6, pp.
4499245003, 2018.
[22] H. Vahdat-Nejad, E. Asani, Z. Mahmoodian, and M. H.
Mohseni, “Context-aware computing for mobile crowdsensing:
A survey,” Futur. Gener. Comput. Syst., vol. 99, pp. 321332,
2019.
[23] M. Marjanović, L. Skorin-Kapov, K. Pripužić, A. Antonić, and I.
Podnar Žarko, “Energy-aware and quality-driven sensor
management for green mobile crowdsensing,” J. Netw. Comput.
Appl., vol. 59, pp. 95108, 2016.
[24] M. Mehdi, G. Mühlmeier, K. Agrawal, R. Pryss, M. Reichert,
and F. J. Hauck, “Referenceable mobile crowdsensing
architecture: A healthcare use case,” Procedia Comput. Sci., vol.
134, pp. 445451, 2018.
[25] J. Bou Abdo and J. Demerjian, “Evaluation of mobile cloud
architectures,” Pervasive Mob. Comput., vol. 39, pp. 284303,
2017.
[26] A. Suliman, H. Otrok, R. Mizouni, S. Singh, and A. Ouali, “A
greedy-proof incentive-compatible mechanism for group
recruitment in mobile crowdsensing,” Futur. Gener. Comput.
Syst., vol. 101, pp. 11581167, 2019.
[27] X. Li and D. W. Goldberg, “Toward a mobile crowdsensing
system for road surface assessment,” Comput. Environ. Urban
Syst., vol. 69, pp. 5162, 2018.
[28] F. Shi, Z. Qin, D. Wu, and J. A. McCann, “Effective truth
discovery and fair reward distribution for mobile crowdsensing,”
Pervasive Mob. Comput., vol. 51, pp. 88103, 2018.
[29] R. Azzam, R. Mizouni, H. Otrok, A. Ouali, and S. Singh, “GRS:
A Group-Based Recruitment System for Mobile Crowd
Sensing,” J. Netw. Comput. Appl., vol. 72, pp. 3850, 2016.
[30] T. Luo, S. S. Kanhere, J. Huang, S. K. Das, and F. Wu,
“Sustainable Incentives for Mobile Crowdsensing : Auctions ,
Lotteries , and Trust and Reputation Systems,” IEEE
Communications Magazine, no. March, IEEE, pp. 6874, 2017.
[31] G. Luo, K. Yan, X. Zheng, L. Tian, and Z. Cai, “Preserving
adjustable path privacy for task acquisition in Mobile
Crowdsensing Systems,” Inf. Sci. (Ny)., p.
https://doi.org/10.1016/j.ins.2018.12.013, 2018.
[32] N. P. Owoh and M. Mahinderjit Singh, “Security analysis of
mobile crowdsensing applications,” Appl. Comput. Informatics,
vol. doi.org/10, 2018.
[33] N. B. Truong, G. M. Lee, T. W. Um, and M. MacKay, “Trust
Evaluation Mechanism for User Recruitment in Mobile
Crowdsensing in the Internet of Things,” IEEE Trans. Inf.
Forensics Secur., vol. 14, no. 10, pp. 27052719, 2019.
[34] M. Yu, H. Lin, and J. Hu, “A data trustworthiness enhanced
reputation mechanism for mobile crowdsensing,” in 2017 IEEE
International Conference on Internet of Things (iThings), IEEE
Green Computing (GreenCom) and Communications, IEEE
Cyber, Physical and Social Computing (CPSCom), IEEE Smart
Data (SmartData) 2017, 2018, vol. 2, pp. 743747.
[35] B. Kantarci, S. Member, and H. T. Mouftah, “Mobility-aware
Trustworthy Crowdsourcing in Cloud-Centric Internet of
Things.”
[36] T. Hu, M. Xiao, C. Hu, G. Gao, and B. Wang, “A QoS-sensitive
task assignment algorithm for mobile crowdsensing,” Pervasive
Mob. Comput., vol. 41, pp. 333342, 2017.
[37] J. Wang, I. R. Chen, J. J. P. Tsai, and D. C. Wang, “Trust-based
mechanism design for cooperative spectrum sensing in cognitive
radio networks,” Comput. Commun., vol. 116, pp. 90100, 2018.
[38] M. Khasawneh and A. Agarwal, “A Collaborative Approach
towards Securing Spectrum Sensing in Cognitive Radio
Networks,” Procedia Comput. Sci., vol. 94, pp. 302309, 2016.
[39] T. Liu, Y. Zhu, and L. Huang, “TGBA: A two-phase group-
buying based auction mechanism for recruiting workers in
mobile crowdsensing,” Comput. Networks, vol. 149, pp. 5675,
2019.
[40] H. Shen, G. Bai, Y. Hu, and T. Wang, “P2TA: Privacy-
preserving task allocation for edge computing enhanced mobile
crowdsensing,” J. Syst. Archit., vol. 97, pp. 130141, 2019.
[41] M. Yang, T. Zhu, K. Liang, W. Zhou, and R. H. Deng, “A
blockchain-based location privacy-preserving crowdsensing
system,” Futur. Gener. Comput. Syst., vol. 94, pp. 408418,
2019.
[42] T. Dimitriou and I. Krontiris, “Privacy-respecting auctions and
rewarding mechanisms in mobile crowdsensing applications,” J.
Netw. Comput. Appl., vol. 100, pp. 2434, 2017.
[43] G. Xu, H. Li, C. Tan, D. Liu, Y. Dai, and K. Yang, “Achieving
efficient and privacy-preserving truth discovery in crowdsensing
systems,” Comput. Secur., vol. 69, pp. 114126, 2017.
[44] K. Z. Ghafoor, L. Kong, A. S. Sadiq, Z. Doukha, and F. M.
Shareef, “Trust-aware Routing Protocol for Mobile
Crowdsensing Environments,” in 2018 IEEE Conference on
Computer Communications Workshops (INFOCOM WKSHPS):
2018 IEEE Infocom MiseNet Workshop, 2018, pp. 8287.
[45] J. Ni, K. Zhang, Q. Xia, X. Lin, and X. S. Shen, “Enabling
Strong Privacy Preservation and Accurate Task Allocation for
Mobile Crowdsensing,” J. LAtex Cl. Files, vol. 14, no. 8, 2015.
[46] T. Yang, X. Xiangyang, L. Peng, L. Tonghui, and P. Leina, “A
secure routing of wireless sensor networks based on trust
evaluation model,” Procedia Comput. Sci., vol. 131, pp. 1156
1163, 2018.
[47] S. M. B. Ganeriwal S, Balzano L K, “Reputation-based
framework for high integrity sensor networks.,” J ACM Trans.
Sens. Networks, vol. 4, no. 3, pp. 132146, 2008.
[48] K. Yi, R. Du, L. Liu, Q. Chen, and K. Gao, “Fast participant
recruitment algorithm for large-scale Vehicle-based Mobile
Crowd Sensing,” Pervasive Mob. Comput., vol. 38, pp. 188199,
2017.
[49] M. Abououf, R. Mizouni, S. Singh, H. Otrok, and A. Ouali,
“Multi-worker multi-task selection framework in mobile
crowdsourcing,” J. Netw. Comput. Appl., vol. 130, no. January,
pp. 5262, 2019.
[50] K. Agrawal et al., “Towards Incentive Management Mechanisms
in the Context of Crowdsensing Technologies based on
TrackYourTinnitus Insights,” Procedia Comput. Sci., vol. 134,
pp. 145152, 2018.
[51] M. E. Gendy, A. Al-Kabbany, and E. F. Badran, “Maximizing
Clearance Rate of Reputation-aware Auctions in Mobile
Crowdsensing,” in 2019 16th IEEE Annual Consumer
Communications and Networking Conference (CCNC ), 2019,
pp. 56.
[52] F. Wang, W. Jiang, G. Wang, and S. Guo, “Influence
maximization by leveraging the crowdsensing data in
information diffusion network,” J. Netw. Comput. Appl., vol.
136, pp. 1121, 2019.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.2968797, IEEE Access
VOLUME XX, 2017 9
DOROTHY MWONGELI KALUI is currently
a Ph.D. student at University of Science and
Technology Beijing, China. She is a Lecturer in
the Department of computer science, School of
Computing and Informatics Meru University,
Kenya. She holds MSc. Information Systems
from The University of Nairobi, Kenya. Her
current research interests include Spatial
Databases, Mobile Data Privacy, the IoT security
and Application of Information Technology in
organizations especially in Financial Institutions.
DEZHENG ZHANG, Director of Beijing Key
Laboratory of Knowledge Engineering for Materials
Science, University of Science & Technology
Beijing, China. Director/Professional lead for
knowledge engineering in specific domains,
especially in Traditional Chinese Medicine and
Materials Science. Doctor supervisor, main
research directions cover data mining and knowledge
discovery, ontology-based knowledge base
construction and intelligent information processing.
Promoting, developing and supporting research and teaching programs
in University of Science & Technology Beijing, China
GEOFFREY MUCHIRI MUKETHIA is an
Associate professor and dean school of computing
and Technology, Muranga University of
Technology. He received his BSc. in Information
Science from Moi University in 1995, his MSc.in
computer science from Periyar University in 2004,
and his Ph.D. in Software Engineering from
Universiti Putra Malaysia in 2011. He has wide
experience in teaching and supervision of
postgraduate students. His research areas include software and business
process metrics, software quality control, Component-based software
engineering, data privacy and technology adoption.
JARED OKOYO ONSOMU received a B.S.
degree in Computer Science from The University
of Nairobi, Kenya in 2009 and the M.S degree in
Computer Science from The University of
Nairobi, Kenya in 2014. His research interests are
in the area of Artificial intelligence and machine
learning.
Article
Based on the Internet of Behavior (IoB), mobile crowd-sensing (MCS) utilizes the Internet of Things (IoT) to recruit users by analyzing behavioral patterns. MCS is widely used in numerous large-scale and complex monitoring services, but it cannot provide stable and high-quality services due to nondeterministic user mobility and behaviors , which has a vital impact on recruiting high-quality users. In this paper, a stochastic semi-algebraic hybrid system (SSAHS) model is constructed to characterize the user mobility and behaviors of the MCS systems. Based on the definition of probabilistic path and task execution rate, a nondeterministic evaluation mechanism is proposed to measure the nondeterministic user mobility and behaviors, and to give the probability of the user completing the MCS task under the specified time bound and space conditions. The greater the probability is , the higher the quality of the user. Furthermore, a user recruitment scheme based on a nondeterministic evaluation mechanism (NUR) is developed. The NUR uses historical user data to predict user mobility and behaviors ; therefore, high-quality users are recruited to quickly upload reliable sensing data. We conduct simulation experiments based on a real-world user traces dataset , Geolife. The results show that compared with competing recruitment strategies, NUR achieves a higher QoS for the same MCS sensing tasks.
Conference Paper
Full-text available
With the emergence of mobile crowdsensing (MCS) we now have the possibility of leveraging the sensing capabilities of mobile devices to collect information and intelligence about cities and events. Despite the promise that MCS brings, this new concept opens the door to a multitude to security and privacy threats and attacks. Indeed, the human involvement in the crowdsensing process and the openness of this process to any participant, render the task of securing MCS environments a very challenging task. In this work, we propose a Blockchain based hybrid architecture for the detection and prevention of fake sensing activities in MCS. Our architecture leverages the capabilities of the Blockchain network and introduces a new role to the MCS architecture to ensure the validation of the collected information. Combining both data quality metrics along with behavioral analysis based participants' reliability scoring, our solution is able to detect variations in behavior and quality of contributions. The proposed solution was tested with real life data collected from 200 mobile users, over the span of 2 years, and the results obtained are very promising.
Article
Full-text available
Mobile crowd sensing has attracted the attention of many researchers with the development of mobile phones. Many studies focus on the optimization of single-task oriented allocation, and most of them optimize only one objective. We consider multi-objective optimization for multi-task allocation in mobile crowd sensing with a limited participants pool. We formulate the problem for temporal-spatial coverage tasks. Two kinds of methods called constrain method and heuristic algorithm method are used to solve the multi-objective optimization problem. For the constrain method, we propose two new greedy algorithms: GMaxEOQU and GMinEOIP. GMaxEOQU pursues the max overall quality utility with budget constraint while GMinEOIP minimize the overall incentive payment with overall quality utility constraint. For the heuristic algorithm method, we propose a Pareto-optimal particle swarm optimization algorithm (PPSO) to search a set of Pareto-optimal solutions for the platform. Extensive experiments validate the performance of our algorithms.
Article
Full-text available
Mobile crowd sensing (MCS) is considered as a powerful paradigm which takes advantage of the pervasive sensorembedded smartphones to collect data. However, MCS assumes all workers always are trusted, and thus offering opportunities for malicious workers to conduct the crowd sensing data falsification (CSDF) attack. To suppress such threat, recent efforts have been made to trust mechanism. Currently, some malicious workers can collude with each other to form a collusive clique, and thus not only increasing the power of CSDF attack but also avoiding the detection of trust mechanism. To ensure honest data collection in MCS, we must fight against such collusive CSDF attack. Noting that the duality of sensing data, we propose a defense scheme called BMCA from the design idea of binary-minmaxs clustering analysis to suppress collusive CSDF attack. In the BMCA scheme, the logic AND operation corresponding to the type of “1” and “0” historical sensing data is used to measure the similarity between any two workers. Based on this, we find the feature that collusive CSDF attackers usually hold high trust value and a low variance in their similarity vector. To detect collusive CSDF attackers, the min and max variance analysis is introduced to design a new binary-minmaxs clustering algorithm. Moreover, the BMCA scheme can perfect trust evaluation to prevent the trust value growth of collusive CSDF attackers. Simulation results show that the BMCA scheme can enhance the accuracy of trust evaluation, and thus successfully reducing the power of collusive CSDF attack against data collection in MCS.
Article
Full-text available
As crowd sensing gains more popularity, the issue of recruiting the best workers for a fair cost becomes more relevant. In this paper, we tackle the problem of workers' greediness, where workers overprice their data to improve their profi�t, by presenting an incentive mechanism for group recruitment crowd sensing systems. The proposed incentive mechanism, which consists of a selection and a payment mechanism, considers not only the Quality of Information (QoI) of the worker and his contribution to the overall QoI of the group, but also the cost of his contribution compared with other workers. This technique avoids the selection of greedy members since their costs will be high compared to the QoI they offer, hence forcing them to align with the other members by advertising their true cost. On the other hand, the proposed payment mechanism rewards workers based on the improvement they offer to the QoI and the cost of the group. We prove that the proposed incentive mechanism is strategy-proof and demonstrate its effectiveness by adapting it to an existing group recruitment framework, and comparing it with the original model through experiments with real datasets.
Article
Full-text available
Mobile crowdsensing (MCS) has gained significant attention in recent years and has become an appealing paradigm for urban sensing. For data collection, MCS systems rely on contribution from mobile devices of a large number of participants or a crowd. Smartphones, tablets, and wearable devices are deployed widely and already equipped with a rich set of sensors, making them an excellent source of information. Mobility and intelligence of humans guarantee higher coverage and better context awareness if compared to traditional sensor networks. At the same time, individuals may be reluctant to share data for privacy concerns. For this reason, MCS frameworks are specifically designed to include incentive mechanisms and address privacy concerns. Despite the growing interest in the research community, MCS solutions need a deeper investigation and categorization on many aspects that span from sensing and communication to system management and data storage. In this paper, we take the research on MCS a step further by presenting a survey on existing works in the domain and propose a detailed taxonomy to shed light on the current landscape and classify applications, methodologies and architectures. Our objective is not only to analyze and consolidate past research but also to outline potential future research directions and synergies with other research areas.
Article
Full-text available
Mobile Crowd-Sensing (MCS) has appeared as a prospective solution for large-scale data collection, leveraging built-in sensors and social applications in mobile devices that enables a variety of Internet of Things (IoT) services. However, the human involvement in MCS results in a high possibility for unintentionally contributing corrupted and falsified data or intentionally spreading disinformation for malevolent purposes, consequently undermining IoT services. Therefore, recruiting trustworthy contributors plays a crucial role in collecting high-quality data and providing better quality of services while minimizing the vulnerabilities and risks to MCS systems. In this article, a novel trust model called Experience-Reputation (E-R) is proposed for evaluating trust relationships between any two mobile device users in a MCS platform. To enable the E-R model, virtual interactions among the users are manipulated by considering an assessment of the quality of contributed data from such users. Based on these interactions, two indicators of trust called Experience and Reputation are calculated accordingly. By incorporating the Experience and Reputation trust indicators (TIs), trust relationships between the users are established, evaluated and maintained. Based on these trust relationships, a novel trust-based recruitment scheme is carried out for selecting the most trustworthy MCS users to contribute to data sensing tasks. In order to evaluate the performance and effectiveness of the proposed trust-based mechanism as well as the E-R trust model, we deploy several recruitment schemes in a MCS testbed which consists of both normal and malicious users. The results highlight the strength of the trust-based scheme as it delivers better quality for MCS services while being able to detect malicious users. We believe that the trust-based user recruitment offers an effective capability for selecting trustworthy users for various MCS systems and, importantly, the proposed mechanism is practical to deploy in the real world.
Article
Mobile crowdsensing engages a crowd of individuals to use their mobile devices to cooperatively collect data about social events and phenomena for special interest customers. It can reduce the cost on sensor deployment and improve data quality with human intelligence. To enhance data trustworthiness, it is critical for service provider to recruit reliable mobile users based on their personal features, e.g., mobility pattern and reputation, but unfortunately it could lead to privacy leakage of mobile users. It becomes challenging to resolve the contradiction between user privacy and task allocation in mobile crowdsensing. To address the issue, we propose SPOON, a strong privacy-preserving mobile crowdsensing scheme supporting accurate task allocation from geographic information and credit points of mobile users. Specifically, service providers can recruit mobile users based on their locations, and select proper sensing reports according to their trust levels but without invading user privacy. By utilizing proxy re-encryption and BBS+ signature, sensing tasks are protected and reports are anonymized to prevent privacy leakage. In addition, a privacy-preserving credit management mechanism is introduced to achieve decentralized trust management and secure credit proof for mobile users. Finally, we show the security properties of SPOON and demonstrate its efficiency on computation and communication.
Article
The information sensed by a mobile device is just data, while the aggregation of information sensed by thousands of devices is knowledge. This is the basic idea of Mobile CowdSensing (MCS). In MCS, the user selection problem has always been a key issue, where the task publisher attempts to recruit a suitable set of users to cooperatively sense the events or data at some particular time points and places. Hence, the temporal and spatial constraint becomes the main criterion to select the users. However, whether a user prefers to accomplish the sensing data not only depends on the temporal and spatial relationship, but also lies on the data property. In other words, a user is likely to take the sensing task if it is interested in the data to be sensed. In this paper, we propose a user Selection utilizing data properties in mobile crowdsensing (SPM), where a triple-layer structure considering not only the temporal and spatial probability, but also the data's property is formulated, and task finishing probabilities are calculated in both the intentional and unintentional situations. Then, a greedy algorithm is proposed to select a suitable set of users for finishing the sensing tasks. We conduct extensive simulations based on three widely-used real-world traces: geolife, roma/taxi, epfl and a synthetic data set. The results show that, compared with other user selection strategies, SPM finishes the largest number of sensing tasks.
Article
The algorithm of influence maximization aims at detecting the top-k influential users (seed set) in the network, which has been proved that finding an optimal solution is NP hard. To address this challenge, finding the trade-off between the effectiveness and efficiency may be a more realistic approach. How to accurately calculate the influence probability is a fundamental and open problem in influence maximization. The existing researches mainly adopted the pair-wise parameters to denote the influence spread probability. These approaches suffer severe over-representing and overfitting problems, and thus perform poorly for the influence maximization problem. In this paper, we calculate the influence probability by learning low-dimensional vectors (i.e., influence vector and susceptibility vector) based on the crowdsensing data in the information diffusion network. With much fewer parameters and opposed to the pair-wise manner, our approach can overcome the overfitting problem, and provide a foundation for solving the problem effectively. Moreover, we propose the DiffusionDiscount algorithm based on the novel method of influence probability calculation and heuristic pruning approach, which can achieve high time efficiency. The experimental results show that our algorithm outperforms other five typical algorithms over the real-world datasets, and can be more practical in large-scale data sets.