The Challenge of Autonomous Agents: Pitfalls and How to Avoid Them
In this chapter, the authors discuss some of the important pitfalls. They identifie on the basis of literature on robot design, many informal discussions, and their own experience. The authors show how the pitfalls can be avoided by discussing a number of approaches, which they have called distributed adaptive control. They also demonstrate how the approaches contribute toward solving some of the fundamental problems of artificial intelligence. Perhaps more pertinent to autonomous agent design is the hypothesis that there is a separate functional module for visual perception. In the traditional view the module enables the agent to recognize objects. An important step in designing a robot is, of course, the specification of its physical set-up, its body, its sensors, and effectors. The sensors and effectors rules that specify the robo's response to a particular sensory input pattern can be defined in terms of sense-act rules.