Opportunistic, collective, crowd-sourced activity and context recognition in Complex Sensor Networks

Thanks to technological advances, sensors are now embedded in objects, in our environments, and even in our clothing. They are available in ever larger quantities, and, in a few years, sensing will be pervasive.

This transformation from formerly sensor-poor environments into sensor-rich environments changes the paradigm for activity recognition. Rather than thinking about which sensors to deploy for a given recognition task, the question is now how to best make use of available resources.

Thus, we investigate:

• Frameworks and programming models for the efficient distributed execution of context recognition algorithms
• Distributed and cooperative algorithms for context recognition
• New applications arising from sensing possible on a large scale.

Running EU-funded and CH-funded project briefs

Smart-DAYS: Smart Distributed daily living Activity-recognition systems

Activity recognition in daily life is characterized by an open-ended environment where the set of relevant activities may not be known at design-time.

In Smart-days we devise activity recognition running on smartphones which are able to learn and expand their repertoire of known activities at run-time in the most autonomous manner. We achive this by combining:

• Unsupervised hierarchical clustering and semi-supervised learning to support high-level activity recognition
  
• Crowdsourcing to exploit the knowledge acquired by other users’s devices and web databases to bootstrap activity recognition.

OPPORTUNITY: Opportunistic activity recognition

Methodologies are missing to design context-aware systems that

• work over long periods of time despite changes in sensing infrastructure (sensor failures, degradation);
• provide the freedom to users to change wearable device placement;
• can be deployed without user-specific training.

This limits the real-world deployment of AmI systems.

We develop opportunistic systems that recognize complex activities/contexts despite the absence of static assumptions about sensor availability and characteristics.

SOCIONICAL: Novel applications enabled by large-scale sensing

Nowadays almost every person has a cellphone equipped with sensors. Sensing on such a massive scales opens up avenues for new applications.

Understanding human collective behavior is a challenging multidisciplinary problem with a large number of applications: assistance in emergency and disaster scenarios, urban space planning, as well as smart-traffic management.

Motivated by these possibilities, we investigate how to recognize human collective behaviors from sensors currently included in mobile phones, as well as additional modalities available in our research platforms. We also consider the new possibilities to provide smart assistance based on human collective behavior sensing, e.g. in an emergency situation.

Past Projects

SENSEI: Integrating the Physical with the Digital World of the Network of the Future

We focus on Body Area Networks which include small sensors worn on the body and found in the immediate environment of the user. These networks typically have a highly dynamic topology of heterogeneous network nodes, and include mobile devices such as cell phones or PDAs as well as resource constrained sensor motes with limited battery power and processing capabilities.

The main challenges lie in discovering devices in the environment, organizing distributed processing, and designing adaptable context recognition algorithms.

e-SENSE: Capturing ambient intelligence for B3G networks using Wireless Sensor Networks

Titan: Distributed processing environment for Wireless Sensor Networks

Group Members

• Dr Daniel Roggen
• Alberto Calatroni
• Long-Van Nguyen-Dinh
• Mirco Rossi
• Claudia Villalonga
• Martin Wirz
• Zack Zhu

Former members

• Dr Clemens Lombriser
• Dr Kilian Förster