Multimodal Sensor Networks (MuSe)

Wireless sensor networks (WSNs) consist of battery-powered devices that sense the surrounding environment, eventually process the sensed data, and then relay them over a multi-hop wireless network to a base station or sink for usage by human operators or other applications. As such WSNs have seen applications in a wide array of scenarios: from wildlife monitoring in natural environments to road traffic analysis in urban areas.

In the past 15 years, researchers in academia, incl. fellow CReWMaN, and the industry have provided solutions to many problems in traditional WSNs, including area coverage and connectivity, communication protocols at the physical, link, network, and transport layers, and processing in the form of data fusion and estimation with special attention to energy efficiency.

Advances in hardware technology have enabled the integration of a wide array of sensors, more powerful microcontrollers, and different radios not only in sensor nodes, but also in smartphone and other electronic devices. We are now working out models, solutions, and analysis thereof to operate these multimodal WSNs (MWSNs) in a secure and energy efficient fashion with guaranteed quality of information for the end users. The research problems that we are addressing include the following ones:

  • Fusion of complex data streams such as audio and video streams originating from video sensor networks (VSNs) and smartphones.
  • Decision making based on the large, heterogeneous, dynamic, live data streams, and contextual information.
  • Energy efficient operations of sensors, including duty cycling, lifetime maximization in treebased MWSNs, and alternative sensing and communication paradigms

Projects

People

Faculty

Students

  • Mayank Raj
  • Pratool Bharti
  • Francesco Restuccia

Collaborators

  • Dr. Vipul Gupta, Oracle Labs