Understanding Multi-Instrument Measurement Systems.

Peter Bajcsy

Understanding Complex Systems Symposium, May 17-20, University of Illinois at Urbana-Champaign 2004.

The presentation addresses the issues related to theoretical modeling and experimental understanding of multi-instrument measurement systems that deal with multi-dimensional multi-variate data and include sensor networks, wireless communication, data acquisition, fusion, analysis, and modeling and visualization components. Research and development of such systems requires an interdisciplinary approach bridging areas of Sensor Technology, Image and Signal Processing, and Physics- and Statistics-Based Modeling of Sensor Data.

We present a range of multi-instrument problems and solutions which fall into the following broad categories: (1) image pre-processing, (2) real-time detection and recognition, (3) multi-spectral 3D scene modeling based on electromagnetic and statistical prediction models, (4) data fusion and (5) hazard aware spaces. Examples from each broad category will demonstrate the modeling and experimental complexity, such as, real-time systems for machine vision and robot control applications, near-real time multi-spectral scene modeling phenomenology, and experimental frameworks for fusing multi-dimensional multi-variate and multi-sensor data using uncertainty analysis.