The Michigan Engineer News Center

New research program aims to make better “sense” of the world

Applications of this research range from soil sensors which allow for increased understanding of global climate change to futuristic sensory skins which can monitor the integrity of an object.| Medium Read

Pictured above: Achilleas Anastasopoulos, Mahta Moghaddam, Sandeep S. PradhanDemos
Teneketzis (PI)

Imagine dozens or even hundreds of miniature sensors informing us about climate conditions, water quality, bodily health, structural safety, and air pollution, and also tracking movement, monitoring plant operations, and even detecting forest fires. This is our future, and in some cases is happening now, but how can we make these tiny sensors work together in perfect harmony?

A new 5-year $2.5M research program funded by the National Science Foundation, called “Controlled Sensing, and Distributed Signal Processing and Decision Making in Networked Systems,” aims to address fundamental issues that arise in networked systems so that they can operate with maximum efficiency. This is especially critical as the individual sensing devices continue to be scaled down to millimeter size and even smaller.

Networked systems are comprised of a family of interconnected sensor nodes that sense data about the environment, communicate with each other, and communicate with “decision centers.”

With applications that include soil sensors to increase our understanding of global climate change, millimeter-scale networked sensors to monitor air and water quality, structural integrity, and perform surveillance, as well as futuristic “sensory skins” to monitor the properties of the item itself or its surroundings, improving the communication within these networked systems is a high priority, yet involves significant challenges.

Individual sensor nodes must communicate with each other as well as with more central decision centers, and this activity involves sensing, signal processing, communication, and decision making functions. At the same time, there are often extreme limits on energy, data storage, and computational capabilities within each node.

“We will be using basic ideas from stochastic optimization, distributed computation, and probability theory to develop novel methodologies that address the underlying challenges,” stated Prof. Demos Teneketzis, PI for the project. Teneketzis is an expert in stochastic control, especially as it is applied to multi-agent decision and control, wireless communication systems, and information theory.

Prof. Teneketzis intends to demonstrate the methodologies developed in the project in related research that involves smart environmental sensor webs, which estimate soil moisture using in situ sensors. The data derived from the sensors will be evaluated against satellite-derived measurements of the same soil samples. There are currently three sites with the sensors already in place or planned for installation: Michigan (Matthaei Botanical Gardens in Ann Arbor), Oklahoma, and California. The soil moisture project is led by Prof. Mahta Moghaddam, and includes Teneketzis as co-PI.

In addition to the applications already named, the research findings of this project are expected to be useful to NSF’s National Ecological Observatory Network (NEON) program, to NASA’s earth science program, including Soil Moisture Active Passive (SMAP) mission, and to the U.S. National Oceanic and Atmospheric Administration (NOAA) monitoring program.

Other Michigan investigators on the project are Profs. Achilleas AnastasopoulosMahta Moghaddam, and Sandeep S. Pradhan. Co-PI on the project is Prof. Venugopal Veeravalli of the University of Illinois at Urbana-Champaign. Veeravalli is Director of the Illinois Center for Wireless Systems (ICWS). He will be working with Profs. Tamer Basar and Angelia Nedic at UIUC.

Student uses a device to measure bacterial levels in urine samples

Blue Sky: Up to $10M toward research so bold, some of it just might fail

Inspired by startup funding models, Michigan Engineering reinvents its internal R&D grant structure. | Medium Read