- Decision Boards
- Research Clusters
- Joint Seminars
- Testbed & Simulation
- Summer School
- Visitor Programme
- In the News
- News Archive
Two COOJA plugins and manuals have been published to integrate the TWIST testbed in COOJA and to take checkpoints and perform rollbacks both in TWIST and COOJA.
The 4th International Workshop on Networks of Cooperating Objects for Smart Cities 2013 (CONET/UBICITEC 2013), colocated with CPSWeek 2013, accepts submissions until January 28th, 2013.
The 19th CONET newsletter has been published. You can read on Virtual Organizations for Multi-Model Based Embedded Systems and on the UvA Bird Tracking System.
Recognizing Emotions using Wireless Sensor Networks
- Cluster leader: TI
- Core partners: NUIG, TUB, UCY
- Associated members: UNIBO, UNICAL
Body Sensor Networks (BSNs) enable continuous measurement of physiological parameters, such as heart rate, muscular tension, skin conductivity, breathing rate and volume, during the daily life of a user. When combined with contextual information extracted from the environment through Wireless Sensor Networks (WSNs), these parameters could be used to infer emotions, mood, depression, and levels of stress and anxiety. The inferred psychological state can be used as enabler for novel applications in areas such as marketing, e.g. to evaluate customer appreciation of products/events and to deliver advertisements when users are most receptive, in mobile communication, e.g. to transmit feelings/mood/emotions measured directly on the person, in health care, e.g. to monitor depression and cognitive disorders, and in safety, e.g. to notify situations of danger or detect levels of attention. Using BSNs for emotion, mood, stress or depression recognition is an open research problem and requires development of novel signal processing techniques to interpret and fuse the data collected by multiple sensors.
General objective of the REWSN cluster is to investigate the application of cooperative techniques between Body Sensor Networks (BSNs) and Wireless Sensor Networks (WSNs) to recognize human emotions and other metal conditions. For reaching this goal research efforts must focus on several aspects such as signal processing algorithms, communication protocols, sensor integration and packaging as well ass applications’ definition and final prototyping. Moreover, all the above mentioned components require the definition and implementation of a common software platform to cooperatively manage all the components.
Cooperating patterns are analysed in this cluster at different levels:
- Cooperation among devices on the same BSN for distributed the signal processing load (SPINE approach)
- Cooperation among different BSNs for detecting social relevant interactions
- Cooperation among other WSNs to detect person’s interaction with the environment
Research topics in this cluster are mainly focused on
- Signal processing algorithms definition and implementation
- Distributed and collaborative signal processing approaches
- Communication protocols study and deployment
- Sensor Integration and packaging
- Data fusion algorithms
- Software platforms
- Applications’ scenarios
- Reference architecture definition
An Open Source software framework for distributed signal processing in Body Sensor Networks (http://spine.tilab.com)
Telecom Italia is leading an Open Source project called SPINE (Signal Processing In Node Environment) for distributed signal processing in wireless sensor networks. The main idea behind this project is that resource constrained platforms, such as sensor nodes, need to share computation power and functionalities therefore need to cooperatively work to get the best results. With this in mind, since last year we developed as software platform that enable the distributed approach for signal processing intensive applications. In the last version of the framework, we added support for emotion sensors (ECG, EIP and HRV) as well as support for Shimmer nodes that has been used in the handshake recognition prototype. SPINE has been adopted as common software platform for REWSN application developments. Meanwhile REWSN partners are working on the next generation of the SPINE framework, SPINE2, that will provide a platform independent core able to run on different sensor nodes’ platforms and a task oriented architecture to improve nodes collaboration for distributed signal processing applications.
Detection of Socially Relevant Interactions
SPINE framework has been used as software platform for an application prototype showing handshake detection. A cooperative approach between different BSNs for handshake detection has been studied and implemented in SPINE.
Detection of other social relevant interaction will be studied in the next year as REWSN research activity.
Emotion Extraction from heartbeat signals using Heart Rate Variability Analysis methods
As previously said, a heart rate variability (HRV) sensor has been added as supported sensor in SPINE1.3. Its signals have been collected and studied to get information about the person’s stress level.
This analysis will be enhanced with other emotions during next year activities.
Multi-Purpose Dynamic Personalization Filters based on Cognitive and Affective Human Factors
Cluster partners are also actively working in combining signals from biosensors with current anxiety levels as well as trait anxiety and emotional regulation in order to show correlational behaviouristic patterns. Identification mechanisms correlating real time signals taken with affective factors and dynamically reconstruct/adapt content based on rules are created. As a result, they provide content adaptation concepts applied to user emotional states and social activities.
Integration of wireless sensing and visual information for operator head-tracking
Part of the REWSN activities during the first two years has also been the development of a head- tracking device that integrates accelerometers and gyroscopes for accurate orientation estimation, uses wireless transmissions and audio and also integrates image-based estimation. The system has been validated through a wide range of tests and experiments and will be further exploit in emotion recognition scenarios as part of the next activities.
The REWSN cluster has presented ongoing work as demo prototypes in two international events.
The handshake recognition system based on the SPINE framework has been presented during the 7th European Conference on Wireless Sensor Networks (EWSN) which was held on February 2010 in Coimbra, Portugal and an activity monitoring application showing SPINE1.3 functionalities has been shown during the CONET/NEWCOM++ Workshop on "Cooperating Objects and Wireless Sensor Networks" held in Bologna, Italy in May 2010. Furthermore, during that workshop Telecom Italia has been invited to give a talk about “Wireless Sensor Applications for Smart Grids and BAN”.
Then, SPINE (Signal Processing In Node Environment) has been updated and version 1.3 has been released in Open Source in February 2010. SPINE Open Source community is growing and several international universities, such as University of California at Berkeley, University of Texas at Dallas, University of Bologna and others, are actively involved in the project.
The cluster relevant work has resulted in several publications and common papers.