Genetic Algorithms in Autonomous Embedded Systems

The performance and usefulness of autonomous embedded systems (AES) can be enhanced by providing them with artificial intelligence (AI). Because embedded systems are generally constrained by mul- tiple factors (e.g., power consumption, processing speed, memory, etc.), fully-fledged AI implementations are not feasible for most AES designs. However, microprocessors targeted at embedded systems have improved to the point where it is possible to include certain AI methods in embedded designs. Genetic algorithms offer a modicum of AI that can successfully run on the newest generation of embed- ded processors, utilize minimal fixed storage, and are simple enough to integrate into an AES with beneficial results. This paper provides an argument for why genetic algorithms should be considered for autonomous embedded systems, and describes a method for imple- menting a genetic algorithm to control a small robot.

Gregg, Chris. "Genetic Algorithms in Autonomous Embedded Systems." University of Virginia Dept. of Computer Science Tech Report (2009).

This work has passed a peer-review process.