Content uploaded by Selvi Arumugam
Author content
All content in this area was uploaded by Selvi Arumugam on Jun 11, 2019
Content may be subject to copyright.
IoT Enabled Smart Logistics Using Smart Contracts
Senthamiz Selvi Arumugam
Ericsson Research
Chennai, India
senthamiz.selvi@ericsson.com
Ramamurthy Badrinath
Ericsson Research
Bangalore, India
ramamurthy.badrinath@ericsson.co
m
Venkatesh Umashankar
Ericsson Research
Chennai, India
venkatesh.u@ericsson.com
Anusha Pradeep Mujumdar
Ericsson Research
Bangalore, India
anusha.pradeep.mujumdar@ericsso
n.com
Nanjangud C Narendra
Ericsson Research
Bangalore, India
nanjangud.narendra@ericsson.com
Jan Holler, Aitor Hernandez
Herranz
Ericsson Research
Stockholm, Sweden
{jan.holler,aitor.hernandez.herranz}
@ericsson.com
Abstract—Advancements in sensors and devices have
enabled Internet of Things (IoT) adoption in various sectors,
especially in domains looking to automate and increase their
real-time decision making capabilities to improve efficiencies.
Supply chain management in logistics is a perfect fit for
adoption of IoT, since it involves shipment of assets being
moved, tracked and housed by a number of machines, vehicles
and people each day. Smart Contracts are terms and
conditions parties can specify that assure trust in the
enforceability of the contract and provide visibility at every
step of a supply chain. IoT devices can write to a smart
contract as a product moves from the factory floor to the store
shelves, providing real-time visibility of an enterprises entire
supply chain. This paper proposes a smart logistics solution
encapsulating smart contracts, logistics planner and condition
monitoring of the assets in the Supply Chain Management
area. A prototype of the solution is implemented which
demonstrates accountability, traceability and liability for asset
handling across the supply chain by various parties involved in
a logistics scenario.
Keywords—Blockchain, smart contracts, IoT, AI planner,
edge analytics, machine learning
I. I
NTRODUCTION
Supply Chain Management (SCM) process involves
various activities to transport goods from point of origin to
the point of consumption. The process involves design,
planning, execution, control and monitoring of the Supply
Chain (SC) activities. The various parties involved across
the supply chain work in silos as they have their own
Enterprise Resource Planning (ERP) or manual systems to
carry out their functions. In addition each element of the
SCM follows different processes bridged by paper based
contracts or disparate digital systems with little or no
standards.
A typical process carried out with respect to logistics in
a supply chain involves a paper contract between the
purchaser and supplier detailing information like quantity of
goods, environment conditions, deadline and penalty
involved in case of violation. The supplier then prepares a
contract with the distribution network and the purchaser is
left to track and verify the shipment based on the contractual
obligations signed with the other parties with limited or no
feedback. There may also be a need for re-planning due to
unexpected route conditions or spoilage of goods in transit.
Dynamic adaptability to mitigate such situations is an
important requirement for the logistics process. If there is a
violation of terms and conditions the concerned parties need
to negotiate amongst themselves and come to a settlement.
This may lead to losses for one or both parties involved due
to lack of transparency and traceability. Options exist to
check the conditions in real-time however the information is
accessible for the party involved in the specific stage of the
logistics process and not for all. There is no unified way of
monitoring the entire process from a single viewpoint
leading to lack of transparency. Suppliers and purchasers do
not have visibility on the state of their assets until they reach
the destination.
The first three industrial revolutions came as a result of
the introduction of mechanization, electricity and IT.
Nowadays, the introduction of IoT and Services into the
industrial environment has triggered the fourth industrial
revolution with the vision of everything connected with
everything else [2]. Meanwhile the blockchain technology is
disrupting various verticals with its digitalized,
decentralized and public ledger features. The state of art
Logistic 4.0 [11] uses Cyber-Physical Systems (CPS)
monitors to control the physical processes, usually with
feedback loops where physical processes affect
computations and vice versa. CPS can identify or sense and
locate the item, and send the data to a computer which
collects and analyzes the relevant information. These
systems communicate with other systems or with humans
using the internet as a means of communication, so that data
can be shared in real time and coordinate the relevant
processes [3]. The demand for high-individualized products
and services is continually increasing. Therefore, supply
chain processes (inbound logistics and outbound logistics)
have to adapt to this changing environment, since due to the
increasing complexity, it cannot be handled with ordinary
planning and control practices [4].
Smart Logistics architecture explained in this paper
provides a holistic way of managing the logistics involved in
a supply chain with the help of IoT technologies, blockchain
based Smart Contracts, Machine Learning and Big data.
II. C
HALLENGES
The Food and Agriculture Organization (FAO) of the
United Nations published an article [5] that states that one
third of the food is spoiled before it reaches the end
customer. Lack of crucial flow of information amongst the
parties involved creates a supply and demand issue. A
contract may not cover all the aspects or scenarios that are
likely to happen in the logistics process leading to
241
978-1-5386-6965-5/18/$31.00 2018 IEEE
ambiguities and loss of resources and capital. Improper
planning leads to non-harmonization between channel
partners resulting in loss of goods, time and capital. All
these issues demand a need for an integrated approach to
solve the logistics problem in supply chain management. In
general, most experts ascribe a high importance to data and
analytics to the logistics sector [6]. This brings in vast
opportunities to improve performance and customer
satisfaction. Some of the fundamental challenges in the
logistics area are:
• Transparency: This is needed for the overall
optimization of SCM and effective use of resources
across the entire supply chain. Transparency is in
direct relation to trust which is of paramount
importance in this sector. Many experts [7] therefore
propose the use of Blockchain to provide this much-
needed trust.
• Traceability: This is needed in order to track the
movement of products across the supply chain. In
particular, in the area of food logistics, governments
are increasingly legislating for adopting traceability
systems to minimize food wastage [8]. Traceability
also becomes critical for consumers to know the facts
of the product origin and production methods of food
items.
• Accountability and Liability: This is defined
rendering necessary explanation to the buyer with
regard to the storage, transportation, insurance,
customs, inspection, supervision, packaging, value-
added procedures, stock management, order
management and similar services provided by the
third party logistics provider. Logistics accountability
is comparable to answerability; buyer expects service
at its best quality for the price paid. Logistics
accountability will also have a positive influence on
trust [9].
The challenges discussed above can be addressed by
employing a robust combination of blockchain based smart
contracts, logistics planner and condition monitoring of the
assets. Due to lack of space, the details of these technologies,
along with how they can address logistics related challenges,
are listed in Table I.
III.
S
MART
L
OGISTICS
S
OLUTION
Our proposed Smart Logistics solution implements the
end-to-end tasks starting from performance based supplier
recommendation, contract negotiation, logistics planning to
contract controlled asset monitoring and contract
fulfillment. We have built a holistic solution that combines
the power of the following building blocks which are
standalone and reusable micro services, see Fig. 1.
A. Brief Overview
A High Level Architecture with the core modules of the
Smart Logistics solution is depicted in Fig. 2.
1) Smart Contract: In 1996, Nick Szabo described a
smart contract as a set of promises, specified in digital form,
including protocols within which the parties perform on
these promises. [1]. The main objective of the smart contract
module is to digitize the paper based or manual procedure of
putting down the terms and conditions of a purchase
contract. The module identifies and recommends suitable
supplier(s) for the purchaser and enables the negotiation of
pricing and contract terms with the supplier. Once the
contract terms are accepted by both supplier(s) and the
purchaser, a purchase order is created. Quality and the
delivery time are considered as the key obligation terms in
the contract which is provided as an input to the Condition
Monitoring (CM) module. The smart contract module helps
to automate the entire smart contract life-cycle using a
blockchain based substrate encompassing contract
negotiation, partner selection, contract finalization, contract
enactment, contract monitoring, conflict/issue resolution
during contract enactment.
Fig. 1. The typical flow of events in the Smart Logistics Solution
Fig. 2. The high level architecture of the Smart Logistics Solution
2) Logistics Planner: A single work order may be
fulfilled by more than one supplier. Each supplier may be
located in a different location and can be fulfilling different
quantities. We assume a scenario where distributor network
consists of a warehouse and several trucks that are used to
move goods from supplier warehouse to the purchasers
delivery point. The logistics planner takes care of designing
the optimal plan for transporting the goods. The planner is
also invoked when there are events (unforeseen) that
jeopardize the plan, i.e., the steps in the plan cannot be
executed as intended, e.g., a truck breaking down and not
able to ship the products. This means that some parts of the
plan are affected and consequently the terms agreed to in the
contract may also be violated. It is important to be able to
detect this early and initiate a re-plan where necessary and
alert dependent parties where important.
3) Condition Monitoring: Constant monitoring of the
assets is required for successful contract fulfillment. This
component provides the functionality to collect all the data
from the sensors, controls and reconfigure the system that is
242
managing the assets. The distributed framework provides for
a possibility to distribute the intelligence acquired from the
data to appropriate location: namely edge (with sensors,
gateways like the trucks), fog (small clouds/servers deployed
on factories or warehouses) and cloud as big data centers
with unlimited capabilities. This framework harnesses the
power of edge analytics where latency is a stringent
requirement for maintaining assets integrity. Decisions are
carried out based on the existing conditions via a distributed
machine intelligence framework which tries to overcome the
limitations of network bandwidth and resource constrained
devices.
TABLE I. E
XISTING
C
HALLENGES
A
ND
P
OSSIBLE
S
OLUTIONS
Component
Challenges How can it help
Smart
Contracts
Complex and
Cumbersome
documents
Multiple
parties
Interactions
Accountability
and liability
Digitises Bills of Lading (Contracts).
Connects members of the supply
chain to a decentralized network and
allows them a direct exchange of
documents. Manages ownership of
documents on the blockchain to
eliminate disputes, forgeries and
unnecessary risks. Records the terms
of a trade. Automatically executes the
flow of money based upon signals
resulting from the flow of goods
thereby reducing processing costs.
Simplification of complex multi-
party systems delivery. Connects
banks, lenders, buyers, and suppliers
to streamline and automate
settlement, reduce fraud risk and
costs.
Use of an IoT sensor as an arbitrator
in the event of dispute, e.g., payment
of a penalty to the defaulting party.
Arbitration appeal from IoT device to
selected panel of arbiters. Decision
executed on the smart contract.
Enhanced tracing and verification to
reduce risk of fraud and theft.
Data provenance and reliability. [10]
IoT and
Condition
Monitoring
Transparency
Need for
integrity
Control
Need for
planning
optimization
of available
resources
Asset/inventory management and
remote monitoring
Predictive maintenance, operational
health monitoring and outage
management
Quality assurance and smart testing
Increased operational efficiency and
productivity
Increased customer control with easy
availability of realtime co-related
information
Data management and analytics
AI Planning
Ensuring a
viable plan
exists prior to
committing to
a contract
Unforeseen
events may
jeopardize a
contract or a
plan
Create initial plans that are optimized
and meet the specifications of a
contract
Decide the impact of an unforeseen
event on a plan and compute its
impact on some aspects of the
contract.
Compute alternative plans as
extensions to current plans in the
presence of some events that
invalidate an existing plan
B. Functional Architecture
The functional architecture of our solution, which
documents the interactions among its components, is
depicted in Fig. 3. Implementation Details and key features:
(Table II provides more details of these implementation
details and their practical realization.)
Smart Contract System (SCS): The SCS (see Fig. 4)
provides a recommended list of suppliers for a particular
item that needs to be procured (e.g. 1000 liters of Cream).
Once the list of suppliers has been selected, SCS sends
notifications to the corresponding suppliers and starts the
negotiation process between the purchaser and the supplier
regarding terms and conditions. After completing the
negotiation process a smart contract is established and a
purchase order is created. The SCS also periodically
receives updates from the planner and condition monitoring
modules that provide information about the current status of
the purchase order fulfillment. These updates are used by
the SCS to verify if the contract terms are followed as per
the agreement. In case of any violations, SCS triggers the
appropriate action defined in the contract, e.g., penalty or
even cancellation of the purchase order.
Key features:
• Cloud-based system modeling and development to
facilitate smart contract implementation
• The eContractAgent module focusses on user account
management, service management and smart contract
management.
• Once the contract is signed by all the parties, the
system deploys the smart contract in the contract
runtime environment implemented on Ethereum
Virtual Machine (EVM) to deploy, execute and
monitor the smart contract state machine.
• Distributed smart contract monitoring infrastructure
in the IoT cloud.
• Dynamic machine learning based approach for
efficient selection of a smart contract based on
transaction history from past smart contracts.
• Dynamic enforcement of smart contract terms and
conditions based on real-time data from distributed
sources
• Securing business transactions via techniques such as
distributed ledger and blockchain technologies
Logistics Planner: The Logistics Planner (LP) (see Fig. 5) is
responsible for designing and executing the optimal plan to
fulfill smart contract.
Key features:
• Planning: LP uses a standard AI planner that can plan
with timing considerations. The open source, domain
agnostic, OPTIC planner [18] using PDDL (Planning
Domain Definition Language) domain specification is
used as an input. However this can be replaced by
planner that take advantage of domain knowledge,
without needing to change other parts of the system.
243
• DN - Execution: The execution consists of a bunch of
trucks which move around containers. The controller
hands off plans to individual trucks (agents) and
monitors and reports on their progress. The executor
simulates the movement of the trucks like a
Distribution Network (DN) which interacts with the
controller. The data from the containers of the truck is
monitored by the Condition Monitoring module.
• The planner-executor system is controlled via a REST
API interface to a component called the Logistics
Planner (LP) controller. This provides a management
interface to start and stop the components as well as
initiate and monitor the progress of their execution
via HTTP interfaces.
• These components are loosely integrated (see Fig. 5),
re-targetable to any other use cases and in many cases
directly reusable.
Fig 3. Functional architecture: interaction between the components
Fig. 4. The internal architecture of the Smart Contract System.
Fig. 5. Internal architecture of the logistics planner
Condition monitoring using Distributed Machine
Intelligence framework: The goal of this component (see
Fig. 6) is to provide a reusable component that can be used to
gather sensor data and process it for analytical and actuation
purposes. The component provides the possibility to deploy
the intelligence and processing of data at multiple levels and
locations. It allows the data to be processed close to its
generation despite limitations such as computation, memory,
storage. For example, the functionality can be deployed at
edge devices, capillary gateways, fog nodes or cloud
datacenters. Aspects such as how is the asset being handled,
by whom, when and which highlight accountability and
liability is the key target of this module. Other applications
include:
• Prediction of individual best before end (BBE) dates
of items/goods based on contextual information.
(Predictive analysis)
• Decision for expediting delivery of goods thereby
reducing waste can be taken based on the sensor
readings.
• Analytics can be employed for calculating optimal
transportation routes based on the sensor readings in-
order to meet contract terms on time.
Fig. 6. Condition monitoring module interfaces
Key features:
• The condition monitoring module interacts with both
the LP and SCS through the CM-Decision Maker
(DM)
• The module facilitates a distributed application
environment developed using an actor based toolkit,
akka.io [14].
244
• A HTTP API is defined to interact with the system.
The CM-DM component provides a simplify API,
specific for the Smart Logistics.
• A cluster manager provides the control of all the
locations deployed (cloud, fog, edge, capillary
gateways,) and orchestrates the deployment of the
functionality. It provides a fault-tolerant and resilient
system.
• Multiple workers are deployed on remote locations,
and are used to deploy the distributed processors
depending on the needs of the application
• Communication components: means of interaction
between the devices and the processors.
• Local databases on locations are used to provide local
persistency of the data.
TABLE II.
S
UMMARY OF
T
HE
D
IFFERENT
R
EUSABLE
C
OMPONENTS
U
SED IN
T
HE
S
OLUTION
Component Technology stack Applications in this
solution
Smart contracts Cumulus –
Ethereum Virtual
Machine
Learning based
supplier selection
Binding users into
contracts
Logistics planner Optimization of plans
and schedules
Condition
monitoring
Akka
Python (packages:
pandas, numpy,
scikit-learn )
TensorFlow [17]
Condition monitoring
Asset management
(Predictive analytics,
anomalies detection,
forecasting, online
training etc.)
Data management
Low-latency control
IV. P
ROTOTYPE AND
R
ESULTS
Prototype of the proposed architecture that integrates the
core building blocks of the Smart Logistics Solution is
discussed. Consider the following use case to demonstrate
the capability of this solution. A purchaser needs to procure
1000 liters of Cream from supplier(s). The condition of the
asset (cream) in the contract expects utmost quality to be
maintained. The ideal temperature is required to be around 2
- 4°C and minimal vibrations by the truck to ensure the
cream quality. Penalties are set in the smart contract when
there is a breach in maintaining the temperature and an
allowable threshold of vibrations is crossed. These are the
conditions which the condition monitoring module needs to
report to the SCS and LP. The following sequence diagram
helps in depicting the flow of information across the
different components involved to fulfill the contract set by
the purchaser.
The purchaser places a procurement request with the smart
contract system (SCS) to place his purchase order (PO). The
SCS will recommend a set of suppliers along with their
ratings who can fulfill the PO. The SCS negotiates with the
suppliers and a contract is established with the PO being
acknowledged. The SCS interfaces with Logistics Planner
(LP) for planning and execution of the PO. The LP creates
the plan and executes on the Distribution Network (DN)
which is a set of trucks and warehouses. Based on the
agreed plan, execution begins where the assets (in this case
the Cream cartons) are being transported to reach the
destination. Meanwhile the asset is being monitored using
sensors such as temperature and vibration so as to maintain
the quality agreed in the smart contract. At any change in
the temperature or the vibration, an anomaly is sent to the
LP and SCS from the Condition Monitoring (CM) through
the Decision maker (DM) which monitors the truck
containers. Similarly, in cases when there are some issues
with respect to route conditions or deterioration of the
assets, the LP steps in to re-plan the trip to optimize the
resource utilization.
Fig. 7. An example sequence of events depicting a successful logistics
process of the Smart Logistics Solution
.
Based on the computation through the trip the SCS
expedites payments to the suppliers and also updates the
rating of the suppliers based on fulfillment or violation
counts agreed in the contract. This ensures accountability
and trust based on the facts and data.
V. R
ELATED
W
ORK
Toll Global Logistics deployed a RFID-based pallet
identification system to more efficiently track good and
shipments at its Singapore facility [12]. The company
estimates that this system will save more than 600 person-
days per year.
Transportation of dangerous goods is another potential
area for improvement. At present, the technology lacks
monitoring of dangerous goods’ condition inside the
container in transit which can act as a early warning and
hazardous situation prevention system. The application of
low-power sensor networks inside containers, as well as
automatic positioning in the cargo hold, can effectively
protect the security of maritime container logistics chain
[16]. Perishable goods, that are also temperature-sensitive
products, are a fundamental source of revenue for the cold
chain logistics enterprises; their management, however,
245
constitutes a severe challenge for cold chain logistics
enterprises [15].
Track and trace is the most common form of IoT in the
supply chain and a number of firms are seeing real rewards
from getting ahead with this technology. Decathlon a sports
retailer that owns 850 stores in 22 countries is a prime
example of this. [13]. There are lots of solutions out there
which focus only on optimizing using IoT technology or
Smart Contracts based on the needs but we provide a
solution that integrates all the aspects encompassing an end
to end solution for any logistics use case.
VI. C
ONCLUSIONS
This paper provides the implementation and key features
of the core components of the proposed Smart Logistics
solution. Its key value propositions are the following:
incorporation of technologies such as smart contracts,
distributed machine intelligence and IoT-based condition
monitoring to provide improved trust, traceability and
accountability; and an end-to-end solution spanning contract
negotiation and monitoring, optimal planning and asset
condition monitoring. Future work will focus on scaling up
our solution for larger and more complex logistics scenarios.
R
EFERENCES
[1] N. Szabo, “Formalizing and securing relationships on public networks”,
vol. 2, no. 9, 1997, Published Online at
http://firstmonday.org/ojs/index.php/fm/article/view/548.
[2] H. Kagermann, W. Wahlster and J. Helbig “Recommendations for
implementing the strategic initiative INDUSTRIE 4.0”, Final report of
Industrie 4.0 Working Group, 2013.
[3] M. Herman, T. Pentek, and B. Otto, “Design principles for Industry 4.0
Scenario”, In Proceedings of the 49th Hawaii International Conference
on System Sciences (HICSS), January, 2016, pp. 3928-3937.
[4] M. Premm, and S. Kirn, “A multiagent system perspective on industry
4.0 supply networks”, Proceedings of the 13th German Conference on
Multiagent System Technologies (MATES), 2015, pp 101-118.
[5] United Nations, “Food Loss and Food Waste”, Online at Food and
Agriculture Organization of the United Nations,
http://www.fao.org/foodloss-and-food-waste/en/.
[6] A. Tipping and P. Kauschke, “Shifting patterns: The future of the
logistics industry,”
Retrieved from https://www.pwc.com/sg/en/publications/assets/future-
of-the-logisticsindustry.pdf
[7] N. Hackius and M. Petersen, “Blockchain in logistics and supply chain:
trick or treat?” Proceedings of the Hamburg International Conference
of Logistics, Hamburg, October 2017.
[8] K. Kraisintu and T. Zhang, “The role of traceability in sustainable
supply chain management,” Master of Science Thesis, Chalmers
University of Technology, Gothenburg, Sweden, 2011.
[9] R. Kaynak and S. Bortecine, “Logistics service accountabilites and their
effects on service buyers trust,” Procedia Social and Behavioral
Sciences, vol. 111, pp 731-740, 2014.
[10] M. Warner, “Blockchain for shipping and logistics”, Presentations at
Hong Kong Maritime Week, November, 2017, Available online at
https://www.hkmw.hk/pdf/publication/blockchain-shipping-
andlogistics-presentation-sli.pdf
[11] L. D. Galindo, “The challenges of logistics 4.0 for the supply chain
management and the information technology”, Master Thesis,
Norweigien University of Science and Technology, 2016.
[12] “Advantages of RFID in transportation and logistics”, Motorola, 2014,
Available online at https://commenco.com/wpcontent/
uploads/2014/11/WhitePaper_Logistics_Motorola_RFID_Transportati
on_Logistics_White_Paper.pdf
[13] F. Roberts, “Delivering the goods: 8 examples of IoT transforming
supply chain”, in Internet of Business, 2016, Available online at
https://internetofbusiness.com/8-real-life-examples-iot-supplychain.
[14] “Akka”, Available online at https://akka.io
[15] W. Xu, Z. Zhang, D. Gong, and X. L. Guan, “Neural network model
for the risk prediction in cold chain logistics”, International Journal of
Multimedia and Ubiquitous Engineering , vol.9, no.8, 2014, pp.111-
124.
[16] Y. J. Zeng, S. W. Xu, X. Peng, and X. Q. Wen, “Shipping containers
of dangerous goods condition monitoring system based on wireless
sensor network”, Proceedings of 6th International Conference on
Networked Computing (INC), 2010, pp 1-3.
[17] “Tensor flow”, https://www.tensorflow.org
[18] “OPTIC”, https://nms.kcl.ac.uk/planning/software/optic.html
246
