Conference PaperPDF Available

A Sociometric Sensor Based on Proximity, Movement and Verbal Interaction Detection

AbstractThis work proposes the development of a sociometric
sensor capable of detecting proximity, movement and verbal
interaction between people. The sensor will allow measuring the
level of existing interrelations between the collaborators of an
organization, in order to determine and to measure the
organizational climate. The way to evaluate the current
organizational climate is subjective, through surveys that are not
accurate and that make the results are not adequate. The sensors
developed until now focus on proximity or movement, but not on
the 3 proposed variables: movement, proximity and speech. The
proposed portable device consists of a proximity detector using
RSSI (Received Signal Strength Indicator) based on the
NodeMCU WIFI ESP 8266 module. It also has a motion detector
that uses an accelerometer and finally a voice detection algorithm,
acquired via microphone, aimed at determining verbal interaction
between people. The validation of the equipment was done by
measuring the 3 variables in a sample of 50 people. The results
show a percentage of proximity detection of 90% up to a distance
of 3 meters between individuals, a percentage of success of 92.5%
in the detection of posture and physical activity and a percentage
of success of 87.5% in the detection of verbal interaction between
Keywords- sociometric sensor, RSSI, voice detection, motion
detection, proximity detection, organizational climate.
The organizational climate is defined as a collective
emotional state that is determined by the perceptions that
people have about the organizations in which they work and, in
turn, these perceptions depend on the interactions that occur
between the members of the organization and the activities that
take place there. A variable that evaluates the organizational
climate is the number of interactions that exist among the
members of the organization.
Globally, organizations are betting on organizational
climate assessment, as there is a relationship between
organizational climate and performance. For example, in Peru,
there are at least 27 human management consulting firms that
provide organizational climate measurement services.
However, none of these services have technological devices to
measure interactions in workspaces. Therefore, the imprecision
of current methodologies for the evaluation of the study of the
work environment, because they are based on the subjectivity
of the specialist at the time of taking surveys and focuses mainly
on observable experiences. Data collected in surveys and other
organizational climate assessment methods are subject to the
veracity of the responses and the collaboration of the person
being assessed. Therefore, the results obtained from the
psychological assessment of the organizational climate are not
absolutely accurate and may be biased by the subjective nature
of the data.[1].
A solution proposal has been made at the commercial level
of products such as Humanyze that was created by students of
MIT (Massachusetts Institute of Technology), which offers
companies to analyze social interaction data for organizational
climate, business optimization processes and work strategy.
However, this type of product does not exist in Latin America
and is not customized for the research area of verbal interactions
On the other hand, scientific literature has made proposals
such as those described below.
Daniel Olguín et al. present in [2] the design,
implementation of a device that measures and analyzes human
behavior. They propose the use of electronic modules capable
of measuring face-to-face interaction, conversation time,
physical distance from other people and motor activity in order
to measure individual and collective behavior. In addition, his
proposal was submitted to an experiment in the Marketing area
of a company for a period of one month, in which data was
collected for 2200 hours. Each employee was instructed to use
them during the workday and that they were subject to a
psychological study so that there was no violation of privacy.
The experiment yielded results of how employees on the same
floor related to other floors. On the other hand, this design has
a final weight of 110 grams, which is greater than the proposal
put forward.
On the other hand, Oscar Mozos et al. presents in [3] an
article that aims to detect stress-related behaviors, analyzing
data that are provided by a non-invasive sensor. In this case, a
portable sensor is used to record electrodermal activity (EDA)
or a photopletismograma (PPG), which are sensors of brain
activity in real time. In addition, a sociometric sensor is used to
measure social activity including movement of the person and
voice. It can be concluded, based on this article, that the
development of sociometric sensors is very useful for
psychologists because the data are measurable making
subsequent evaluations more reliable and above all of higher
quality. However, this sensor does not contain a proximity
measurement which does not allow us to know how sociable
one person is with another.
Trinh Do et al. present in [4] an analysis of the social
activities through bluetooth and infrared sensors of 50 people,
which used the sociometric sensor for 6 weeks. The article
makes 3 important contributions; first they create an algorithm
to manage in space-time context the social interactions with the
Bluetooth and IR sensor data. Second, an automatic method for
defining activities is introduced. And, finally, the article
Jorge Tuesta1, Demetrio Albornoz2, Guillermo Kemper3and Carlos A. Almenara4
1,2,3 Faculty of Engineering - School of Electronic Engineering
4Faculty of Health Sciences
Universidad Peruana de Ciencias Aplicadas, Lima, Peru,,,
A Sociometric Sensor Based on Proximity,
Movement and Verbal Interaction Detection
2019 International Conference on Information Systems and Computer Science (INCISCOS)
978-1-7281-5581-4/19/$31.00 ©2019 IEEE
DOI 10.1109/INCISCOS49368.2019.00042
Authorized licensed use limited to: Auckland University of Technology. Downloaded on May 23,2020 at 04:07:16 UTC from IEEE Xplore. Restrictions apply.
presents a case study of interactions in a real organization in
which they give real details of the quality of the sensors and that
precision can reach the recognition of activities. The
presentation of the information is rescued from this article by
means of interaction networks in people sensed by Bluetooth
and infrared sensors. The final product does not include a voice
activity register, which is necessary, since the most common
interaction between people is verbal.
Oren Lederman et al. present in [5] a prototype of badge that
records social interaction through the RSSI of Bluetooth and a
microphone to detect if the person interacts with speech. In
addition, they also developed software for phones where they
try to reduce the cost of hardware by making the sociometric
sensor over the phone. This prototype is very small and has
duration of 40 hours. However, it does not have an
accelerometer to capture the person's physical activity, nor an
SD module to store the information.
Jun Watanabe et al. presents in [6] an approach to the study
of the work environment with a sociometric sensor in a call-
center company, which demonstrated, through the data
recorded by the sociometric sensors, that the level of activity
during working hours is not correlated with the performance of
the work team. They also came to the conclusion that face-to-
face interaction leads to a high level of activities, much more
than teamwork. The authors made use of a sociometric sensor
and an IR (infrared) module to capture the signals of social
interaction and thus be able to make the analysis of
organizational climate. However, this proposal consists of two
separate modules, so this depends more on the user's use.
Based on the above, this paper proposes a sociometric
equipment weighing less than 100 grams, capable of jointly
detecting the level of proximity, verbal interaction and motor
activity. The acquired information is stored in a microSD
memory for later offline analysis. In addition, an important
contribution is the exchange of the Bluetooth module for a WIFI
module, which offers greater range and the possibility of easily
connecting to a base station.
The parts that make up the proposed sensor, as well as the
validation and results obtained are described in the following
The good organizational climate generates positive
consequences; the main ones are productivity, integration,
talent retention and a positive image of the company among
The study of the organizational climate provides the
specialist psychologist, information to detect the origin of
problems such as maladjustment, absenteeism, low productivity
and discomfort at work among others.
The sociometric sensor will acquire information on the
proximity between the individuals evaluated, the motor activity
in the work schedule, the verbal interaction of each one of them
and the time in which each variable is recorded. Therefore, the
sensor will provide the specialist with very reliable data to
improve his interpretation of the organizational climate and
take the corresponding corrective actions.
Fig. 1 shows a sociometric sensor acquiring social
interaction data in a work area. Note that each person must
previously authorize the use of the sensor (for privacy reasons)
in order to be able to install it preferably hanging from the neck
and attached to the chest.
Fig. 1. Final product working in the application scenario.
Fig. 2 shows the block diagram of the proposed sensor. It is
important to note that the central processor of the sociometric
sensor is a WIFI NodeMCU ESP8266 module that has a clock
frequency of 80/160  and a Tensilica Xtensa LX3
CPU. The CPU is 32-bit and ultra-low power [7]. This last one
was very taken into account for the choice of the processor since
the sociometric sensor is portable and therefore the
consumption of energy must be very controlled.
The details of the parts that make up the proposed sensor
will be described in the following sections.
Fig. 2. Block diagram of the proposed system.
Proximity data acquisition.
ESP8266 WIFI NodeMCU module enables easier collection
of proximity data and saves physical space because it includes
a WIFI device and a sufficiently fast processor on the
development board.
Proximity is analyzed by obtaining the RSSI that exists
between the WIFI modules that each sociometric sensor will
possess. The RSSI (Relative Received Signal Strength) is an
indicator of the perceived power in a receiver device [8]. The
RSSI unit is the decibel-milivatio () which is calculated
using the following expression:
Where is received power expressed in milliwatts ().
Theoretically the value of RSSI in  at a distance of
meters is calculated as follows [8]:
Where is the value of RSSI in  in the receiver at a
distance of one meter and the loss constant.
Authorized licensed use limited to: Auckland University of Technology. Downloaded on May 23,2020 at 04:07:16 UTC from IEEE Xplore. Restrictions apply.
In a real environment, the level of RSSI detected among
sociometric sensor is affected by obstacles in the work area. It
is for this reason that many times the loss constant must be
measured empirically.
Then the distance between the sensors can be calculated
using the following expression [9].
To map RSSI values to distance values between devices,
tests were performed by acquiring RSSI data every 10 seconds
so as not to saturate the memory. The acquisition was done at
different distances and the values were recorded as shown in
Table I. The distances used were 0.5m, 1.5m, 2.5m and 3.5m.
This is because an interaction is positive for a distance threshold
of 1.5m.
Table II shows the average values resulting from the RSSI
samples received for each distance evaluated. A graph of the
values in Table II is shown in Fig. 3.
0.5 m
1.5 m
2.5 m
3.5 m
Distance (m)
Acquisition of movement data.
The acquisition of motion data is obtained through the
MPU6050 module which is an IMU (Inertial Measurement
Units) with accelerometer and gyro that acquires the change of
speed in the 3 axes [10] in order to detect the degree of activity
possessed by the user of the sociometric sensor. The data are
acquired every 1 second with 16 bits of precision, so as not to
saturate the available memory space.
Fig. 3. RSSI vs Distance Graphic.
The flowchart of the motion data acquisition process is
shown in Fig. 4. The data is acquired by the central processor
of the development board and then stored in text format in SD
memory. In this case the information provided by RTC (Real-
Time Clock) and the accelerometer of the MPU6050 module is
Fig. 4. Flowchart for acquisition and recording of motion data.
Fig. 5 shows the data obtained from the accelerometer. It
can be seen that in column A is the time register and in column
B the accelerometer data. A movement is considered to have
existed when the values of the accelerometer are much higher
than those recorded at rest, as can be seen in rows 11 and 12.
Fig. 5. Information provided by the accelerometer.
Verbal Interaction Data Acquisition
For voice activity detection (VAD), the MAX9814 module
was used, which includes an omnidirectional microphone and
a low noise amplifier [11].
The microphone of the module was adapted to ensure
unidirectionality and avoid interference from signals that do not
0.5 1.5 2.5 3.5
DIstance (meters)
Authorized licensed use limited to: Auckland University of Technology. Downloaded on May 23,2020 at 04:07:16 UTC from IEEE Xplore. Restrictions apply.
correspond to voice signals emitted by the user carrying the
In the acquisition, a window of 50 ms was considered for
the taking of each sample sent by the voice module. These
samples were digitized by an ADC of 12 bits of precision. Then
a block of 18 samples is formed which is processed for the
decision of presence or no voice activity, which will be
reflected visually every second with the following words: VAD
(presence of voice activity) or NO VAD (no presence of voice
activity). Two metrics were used for this processing: The short
term average energy and zero cross rate of each of the 18
samples. Fig. 6 shows the flowchart for voice activity detection.
The calculation of the short term quadratic energy [11]
applied to the 18 samples per block is expressed as:
Where is the block index, () is the short term average
energy of the block , () is the acquired voice signal and is
the block size. In this case we have = 18.
The number of zero crosses refers to the number of times
voice samples change sign in a given time [12]:
()= ()−((−1))
Where 1() is the zero cross rate of the block and
() the function sign.
Fig. 6. Verbal Interaction Detection Algorithm Flowchart.
From experimental tests carried out in a working
environment, it can be verified that an energy value greater than
0.2 and the presence of at least one crossing by zero is a
sufficient condition to determine the presence of voice activity
(VAD). Whatever other conditions are present, it will be
considered as no voice activity (NO VAD).
A graph of the detection for VAD and NO VAD over time
is shown in Fig. 7.
Fig. 7. Speech Detection Differentiation Graph.
Data processing and recording.
The algorithms shown in the flow diagrams for proximity
detection, motor activity detection, verbal interaction detection
and data logging are executed using the CPU of the WIFI
NodeMCU module ESP8266.
Data logging is achieved using the DS3231 RTC module
and the microSD card adapter. The first module is used to
organize information with date and time in real time when the
data are acquired. The second module stores proximity,
movement and speech information in a text file that is later
visualized by the specialist psychologist.
For the validation of the sensor, a comparison was made of
real data and data sent by the proposed sensor for 40 cases of
possible interaction. The comparison was divided into 3 stages:
register of motor activity, register of voice activity and register
of proximity.
Motor activity detection: These results reflect an
individual's activity during working hours. That is to say the
time that remains seated, stopped or moving. The sensor in this
case detects whether or not there is motor activity. The
percentage of error and success obtained is indicated below:
- Number of cases: 40
- Number of hits: 37
- Number of errors: 3
% =  
   100
Where % is the motion detection error
With the values obtained in the tests of register of motor
activity it was obtained:
% = 3
40× 100 = 7.5 %
Authorized licensed use limited to: Auckland University of Technology. Downloaded on May 23,2020 at 04:07:16 UTC from IEEE Xplore. Restrictions apply.
Then the percentage of success in detecting movement was
92.5 %. The errors presented are mainly due to the sampling
times of the information sent by the sensor (1s), i.e. fast
movements are often not detected.
Voice activity recording: These results represent the verbal
activity of an individual with another individual. The
percentage of error and success in the detection of speech was
- Number of cases: 40
- Number of hits: 35
- Number of errors: 5
The value of the error percentage in voice detection %
is also obtained from (6):
 = 5
40× 100 = 12.5 %
The value of the percentage of success in speech detection
in this case was 87.5 %. As in the case of motion detection, the
adjustments of the sampling times of the sensor information
generate some detection errors.
Proximity recording: These results represent the proximity
of one individual to another. It was obtained the percentage of
error that the sensor has in the proximity metric.
- Number of cases: 20
- Number of hits: 18
- Number of errors: 2
The value of the percentage error in proximity detection
% is also obtained from (9):
 = 2
20× 100 = 10 %
The percentage of success in proximity detection in this case
is 90%. Proximity errors are mainly due to the presence of
obstacles that alter the signal reception level (RSSI) between
Among the 40 cases analyzed, 3 circumstances of data
acquisition for the sociometric sensor were established.
Case 1: Possible Interaction, Voice Recording and Close
Fig. 8 shows the data acquired by the sociometric sensor.
Fig. 8. Data in .txt file provided by the sociometric sensor for
case 1.
Case 2: No Interaction, no voice recording and proximity
greater than 1.5 meters:
Fig. 9 shows the data acquired for this case by the
sociometric sensor.
Fig. 9. Data in .txt file provided by the sociometric sensor for case 2.
Case 3: Possibility of the user speaking alone, voice logging
and no one near.
Fig. 10 shows the data acquired for this case by the
sociometric sensor.
Fig. 10. Data in .txt file provided by the sociometric sensor for case 3.
These three cases are the most common ones recorded in a
day-to-day assessment of the organizational climate. The main
contribution of this work is focused on the coupling of the 3
measurement variables for a more complete evaluation of the
organizational climate; in addition to providing a good success
rate for such evaluation.
In this work it was possible to develop a sociometric sensor
capable of providing additional reliable information for the
decision of the specialist psychologist.
The area of application of organizational climate evaluation
can affect, however, in the percentage of success of the
proximity detection.
Excessive noise in the evaluation application area could
affect the acquisition of voice activity data.
As a future improvement, other hardware modules that
improve the current success rate will be evaluated.
[1] "La Guía del Capital Humano 2017",, 2019.
[Accessed: 19- Oct- 2019].
[2] D. Olguín et al, “Sensible organizations: Technology and methodology
for automatically measuring organizational behavior,IEEE Transactions
Authorized licensed use limited to: Auckland University of Technology. Downloaded on May 23,2020 at 04:07:16 UTC from IEEE Xplore. Restrictions apply.
on Systems, Man, and Cybernetics, Part B (Cybernetics), vol.39, no. 1,
pp. 43-55, 2009.
[3] O. Mozos et al, Stress detection using wearable physiological and
sociometric sensors,” International journal of neural systems, vol. 27, no.
02, pp. 1650041, 2017.
[4] T. Do et al, Inferring social activities with mobile sensor networks,”
In Proceedings of the 15th ACM on International conference on
multimodal interaction, ACM,pp. 405-412, Dec 2013
[5] O. Lederman et al, Open badges: A low-cost toolkit for measuring team
communication and dynamics,” arXiv preprint arXiv, pp. 1710.01842,
[6] J. Watanabe et al, Resting time activeness determines team performance
in call centers,” In 2012 international conference on social informatics,
IEEE, pp. 26-31, Dec 2012.
[7] ESP8266 Datasheet, “ESP8266EX Datasheet Version 4.3” Esprssif.
Systems IOT Team. Datasheet, pp. 131, 2015.
[8] Suvankar, B., Nehe, Debajyoti. B., & Buddhadeb, S. (2017, December).
Estimate distance measurement using NodeMCU ESP8266 based on
RSSI technique. In 2017 IEEE Conference on Antenna Measurement &
Applications (CAMA), IEEE, pp. 170-173, Dec 2017.
[9] Erin-Ee-Lin, L. & Wan-Young, C. (2007, November). Enhanced RSSI-
Based Real-Time User Location Tracking System for Indoor and Outdoor
Environments. In International Conference on Convergence Information
Technology (ICCIT 2007), IEEE, pp. 1213-1218, Nov 2007.
[10] MPU6050 Datasheet. “MPU-6000 and MPU-6050 Register Map and
Descriptions Revision 4.2” InvenSense. Datasheet,pp. 1-46, 2013.
[11] MAX9814 Datasheet. “Microphone Amplifier with AGC and Low-Noise
Microphone Bias” Adafruit. Datasheet, pp. 1-13, 2016
[12] Lokhande, N. N., Nehe, N. S., & Vikhe, P. S. (2012, March). Voice
activity detection algorithm for speech recognition applications. In IJCA
Proceedings on International Conference in Computational Intelligence
(ICCIA2012), vol. iccia (No. 6, pp. 1-4).
Authorized licensed use limited to: Auckland University of Technology. Downloaded on May 23,2020 at 04:07:16 UTC from IEEE Xplore. Restrictions apply.
... Thomaz et al. [151], who classified food intake gestures using cameras and wrist-mounted commodity sensors. Occupancy estimation, counting and prediction [8], [163], [161] Infrared array Activity recognition and detection [87], [173], [106] Occupancy estimation [127] Object detection and tracking [48] Proximity Activity detection [152], [11] Proximity Detection of movement and verbal interaction [152] Recognizing eating activities [11] Pressure Human activity detection [126] Human posture detection [98] Occupancy estimation and prediction [26], [116] Sound Recognizing collaborative activities [164] Occupancy sensing and prediction [36], [116] Heat ...
... Thomaz et al. [151], who classified food intake gestures using cameras and wrist-mounted commodity sensors. Occupancy estimation, counting and prediction [8], [163], [161] Infrared array Activity recognition and detection [87], [173], [106] Occupancy estimation [127] Object detection and tracking [48] Proximity Activity detection [152], [11] Proximity Detection of movement and verbal interaction [152] Recognizing eating activities [11] Pressure Human activity detection [126] Human posture detection [98] Occupancy estimation and prediction [26], [116] Sound Recognizing collaborative activities [164] Occupancy sensing and prediction [36], [116] Heat ...
Full-text available
Increasingly, buildings are being fitted with sensors for the needs of different sectors, such as education, industry and business. Using Internet of Things (IoT) devices combined with analysis of data being generated by these devices, it is possible to infer a number of metrics, e.g. building occupancy and activities of occupants. The information thus gathered can be used to develop software applications to support energy management, occupant comfort, and space utilization. This survey explores the use of sensors in smart building environments, identifying different approaches to employ sensors in buildings. The most commonly used data-driven approaches for activity recognition in such buildings is also investigated, concluding by highlighting current research challenges and future research directions in this area.
... II. BACKGROUND A voice activity detector (VAD) is an algorithm that attempts to separate the audio stream into intervals where voice activity is present and where it is absent [5]. Commonly VAD methods used with sociometric badges use empirically found constant threshold to classify voice activity [2][3] [7][8][9]. Threshold values in these methods depend on the usage environment and it is hard to choose the threshold objectively since there aren't many tools to evaluate the correctness of the threshold value. The lower bound of the threshold values can be found by inspecting values in a period that is known to have no voice activity, but then it is up to the user to choose the threshold. ...
Sociometric badges are an emerging technology for study how teams interact in physical places. Audio data recorded by sociometric badges is often downsampled to not record discussions of the sociometric badges holders. To gain more information about interactions inside teams with sociometric badges a Voice Activity Detector (VAD) is deployed to measure verbal activity of the interaction. Detecting voice activity from downsampled audio data is challenging because down-sampling destroys information from the data. We developed a VAD using deep learning techniques that achieves only moderate accuracy in a low noise meeting setting and in across variable noise levels despite excellent validation performance. Experiences and lessons learned while developing the VAD are discussed.
Full-text available
Ubiquitous technology, big data, more efficient electronic health records, and predictive analytics are now at the core of smart healthcare systems supported by artificial intelligence. In the present narrative review, we focus on sensing technologies for the healthcare of Anorexia Nervosa (AN). We employed a framework inspired by the Interpersonal Neurobiology Theory (IPNB), which posits that human experience is characterized by a flow of energy and information both within us (within our whole body), and between us (in the connections we have with others and with nature). In line with this framework, we focused on sensors designed to evaluate bodily processes (body sensors such as implantable sensors, epidermal sensors, and wearable and portable sensors), human social interaction (sociometric sensors), and the physical environment (indoor and outdoor ambient sensors). There is a myriad of man-made sensors as well as nature-based sensors such as plants that can be used to design and deploy intelligent systems for human monitoring and healthcare. In conclusion, sensing technologies and intelligent systems can be employed for smarter healthcare of AN and help to relieve the burden of health professionals. However, there are technical, ethical, and environmental sustainability issues that must be considered prior to implementing these systems. A joint collaboration of professionals and other members of the society involved in the healthcare of individuals with AN can help in the development of these systems. The evolution of cyberphysical systems should also be considered in these collaborations.
Full-text available
Some of the recent developments in data science for worldwide disease control have involved research of large-scale feasibility and usefulness of digital contact tracing, user location tracking, and proximity detection on users’ mobile devices or wearables. A centralized solution relying on collecting and storing user traces and location information on a central server can provide more accurate and timely actions than a decentralized solution in combating viral outbreaks, such as COVID-19. However, centralized solutions are more prone to privacy breaches and privacy attacks by malevolent third parties than decentralized solutions, storing the information in a distributed manner among wireless networks. Thus, it is of timely relevance to identify and summarize the existing privacy-preserving solutions, focusing on decentralized methods, and analyzing them in the context of mobile device-based localization and tracking, contact tracing, and proximity detection. Wearables and other mobile Internet of Things devices are of particular interest in our study, as not only privacy, but also energy-efficiency, targets are becoming more and more critical to the end-users. This paper provides a comprehensive survey of user location-tracking, proximity-detection, and digital contact-tracing solutions in the literature from the past two decades, analyses their advantages and drawbacks concerning centralized and decentralized solutions, and presents the authors’ thoughts on future research directions in this timely research field.
Full-text available
We present Open Badges, an open-source framework an toolkit for measuring and shaping face-to-face social interactions using either custom hardware devices or smart phones, and real-time web-based visualizations. Open Badges is a modular system that allows researchers to monitor and collect interaction data from people engaged in real-life social settings. In this paper we describe the technical aspects of the Open Badges project and the motivation for its creation.
Conference Paper
Full-text available
While our daily activities usually involve interactions with others, the current methods on activity recognition do not often exploit the relationship between social interactions and human activity. This paper addresses the problem of interpreting social activity from human interactions captured by mobile sensing networks. Our first goal is to discover different social activities such as chatting with friends from interaction logs and then characterize them by the set of people involved, and the time and location of the occurring event. Our second goal is to perform automatic labeling of the discovered activities using predefined semantic labels such as coffee breaks, weekly meetings, or random discussions. Our analysis was conducted on a real-life interaction network sensed with Bluetooth and infrared sensors of about fifty subjects who carried sociometric badges over 6 weeks. We show that the proposed system reliably recognized coffee breaks with 99% accuracy, while weekly meetings were recognized with 88% accuracy.
Conference Paper
Full-text available
Improving team performance has long been a great concern of leaders and managers. Recent progress in wearable sensor technologies has given them a strong means of grasping their team conditions. Studies using such technologies have shown, for example, that the cohesion of a face-to-face network, as measured using wearable sensors, correlates with performance. However, causality between face-to-face communication and performance has remained unclear. We investigated, in a call center environment, the relationships between a team's activity levels while working and while resting and its performance. We found that the activity level while working does not correlate with team performance whereas that while resting does. Furthermore, we found that improving face-to-face communication leads to increased activity levels and to better team performance. Our results demonstrate that team performance can be improved by managing workplace activity levels.
Stress remains a significant social problem for individuals in modern societies. This paper presents a machine learning approach for the automatic detection of stress of people in a social situation by combining two sensor systems that capture physiological and social responses. We compare the performance using different classifiers including support vector machine, AdaBoost, and [Formula: see text]-nearest neighbor. Our experimental results show that by combining the measurements from both sensor systems, we could accurately discriminate between stressful and neutral situations during a controlled Trier social stress test (TSST). Moreover, this paper assesses the discriminative ability of each sensor modality individually and considers their suitability for real-time stress detection. Finally, we present an study of the most discriminative features for stress detection.
Conference Paper
Determining the beginning and the termination of speech in the presence of background noise is a complicated problem. This paper is concerned with labeling sections of speech samples based on whether they are silence, voiced or unvoiced speech. The labeling is done using calculations over the speech samples; zero crossing and short-term energy functions. The short-term energy and zero crossing rate of speech have been extensively used to detect the endpoints of an utterance.
Conference Paper
Existing researches on location tracking focus either entirely on indoor or entirely on outdoor by using different devices and techniques. Several solutions have been proposed to adopt a single location sensing technology that fits in both situations. This paper aims to track a user position in both indoor and outdoor environments by using a single wireless device with minimal tracking error. RSSI (received signal strength indication) technique together with enhancement algorithms is proposed to cater this solution. The proposed RSSI-based tracking technique is divided into two main phases, namely the calibration of RSSI coefficients (deterministic phase) and the distance along with position estimation of user location by iterative trilateration (probabilistic phase). A low complexity RSSI smoothing algorithm is implemented to minimize the dynamic fluctuation of radio signal received from each reference node when the target node is moving. Experiment measurements are carried out to analyze the sensitivity of RSSI. The results reveal the feasibility of these algorithms in designing a more accurate real-time position monitoring system.
We present the design, implementation, and deployment of a wearable computing platform for measuring and analyzing human behavior in organizational settings. We propose the use of wearable electronic badges capable of automatically measuring the amount of face-to-face interaction, conversational time, physical proximity to other people, and physical activity levels in order to capture individual and collective patterns of behavior. Our goal is to be able to understand how patterns of behavior shape individuals and organizations. By using on-body sensors in large groups of people for extended periods of time in naturalistic settings, we have been able to identify, measure, and quantify social interactions, group behavior, and organizational dynamics. We deployed this wearable computing platform in a group of 22 employees working in a real organization over a period of one month. Using these automatic measurements, we were able to predict employees' self-assessments of job satisfaction and their own perceptions of group interaction quality by combining data collected with our platform and e-mail communication data. In particular, the total amount of communication was predictive of both of these assessments, and betweenness in the social network exhibited a high negative correlation with group interaction satisfaction. We also found that physical proximity and e-mail exchange had a negative correlation of r = -0.55 (p 0.01), which has far-reaching implications for past and future research on social networks.
ESP8266EX Datasheet Version 4.3
  • Datasheet