Class03

Consider a network of sensors, each of which has limited sensing resources, which is tasked with collecting noisy classification information on objects. Many modern avionics systems include multiple sensors as well as individual sensors capable of focusing on different objects with different modes. In order to achieve an accurate classification of all objects of interest, it is important to coordinate the allocation and scheduling of the different sensors and sensor modes across objects. The adaptive sensor management problem consists of selecting and scheduling the sensor modes which are applied to objects of interest, taking into account the collected past information into the selection of future sensing actions.

This talk presents a model for a class of adaptive sensor management problems involving the goal of classifying a known number of objects with unknown type, given a finite amount of sensor resources, where the sensor performance parameters are time-invariant, so that the performance parameters associated with a sensor observing an object with a given mode do not depend on the time that sensing activity occurs. The problem is restated as a partially observed stochastic control problem with resource constraints. While exact solution of this model is computationally difficult, we develop an approximation that provides a provable, computable lower bound on the achievable performance of a network of resource-limited sensors. Furthermore, this approximation is used as part of a receding-horizon control policy which achieves performance close to the lower bound. The talk compares the performance of the proposed algorithms to that of alternative algorithms proposed in the literature that seek to maximize information gain on scenarios involving 100 objects.

David Castañon received his Ph.D. in Applied Mathematics at Massachusetts Institute of Technology(1976), and his B.S. in Electrical Engineering at Tulane University (1971). He is an Associate Editor of Computational Optimization and Applications. He is also a member of the IEEE Control Systems Society, Board of Governors, the Institute of Electrical and Electronics Engineers and the Society for Industrial and Applied Mathematics Previously, Dr. Castañon was Senior Scientist at Alphatech, Inc as a and was also a Research Scientist at Massachusetts Institute of Technology, Engineering, Electrical Engineering, Laboratory for Information and Decision Systems.

Host:
Student Host: