Signal understanding and problem solving: A concurrent approach to soft real-time systems

In this paper we present the results of building and running a knowledge-based signal interpretation program on multi-pi-ocessors. A serial, knowledge-based program to interpret passive radar signals from multiple sites was re-designed and IE-implemented on a framework for concument problem solving. The primary objective for the use of multi-processors was to gain performance improvement. A number of experiments were performed to evaluate the pelfonnance gains under varying circumstances. Factors that motivate and constrain problem solving in parallel in general, and real-time signal interpretation in particular, are discussed.