Most motion planning tasks are heavily constrained with respect to many different independent and dependent variables. The classical planning algorithms for complex motions generally require a very detailed mathematical analysis that cannot include all real-world dynamic effects. This paper examines using a scout cluster architecture to plan motions based only on an estimate of the feedforward dynamic response of the system. The technology is equally well suited to both linear and nonlinear system control. The theoretical basis of the scout cluster architecture has been previously discussed by the author (1999).

In this paper, the scout cluster is introduced as a processing architecture capable of developing rapid solutions to difficult cognitive problems. Most cognitive problems have a non-deterministic aspect to them, making the discovery of an efficient algorithm for their solution difficult or impossible. Scouting performs rapid parallel searches in hardware in an uncoordinated manner that makes it almost infinitely scalable. A solutions generated by the architecture are "accidental" or "ingenious", since they are not constrained by a preconceived plan for what solutions must look like. Several cognitive problems are examined and simulated on a scout cluster, and favorable results are obtained.

The authors describe a background study, course developments, and concepts for a curriculum in electronics test engineering. Their conclusion is that appropriate academic programs in electronics test engineering should be developed at the B.S., M.S., and Ph.D levels. The identification of the common background and education, as well as the implementation, of these programs requires strong university/industry cooperation and support. As exemplified by the present study, this process can occur at the local level. However, it is suggested that a broader input and a greater visibility are warranted, and it is recommenced that an organization such as the Test Technology Technical Committee coordinate the development of model curricula for 'meeting the tests of time'.>

Graph theoretic properties of the PLA folding problem which not only give insight into the various folding problems but also provide efficient algorithms for solving the problems are presented. This work is based on the transformation of the PLA into graphs where cliques (completely connected subgraphs) in the graphs correspond to PLA folding sets. A simple heuristic technique known as the greedy algorithm can often identify near-maximum cliques in polynomial time. Experimental data show that this technique is extremely effective when applied to the graphs which generally arise in folding. Variations of the general folding problem such as bipartite folding and constrained folding are addressed.>