Emergent Control and Planning in an Autonomous Vehicle

We use a connectionist network trained with reinforcement to control both an autonomous robot vehicle and a simulated robot. We show that given appropriate sensory data and architectural structure, a network can learn to control the robot for a simple navigation problem. We then investigate a more complex goal-based problem and examine the plan-like behavior that emerges. Autonomous agents An autonomous agent can be abstractly defined as a mapping from a sequence of sensory inputs to an appropriate action in response to these percepts. Such an agent is autonomous to the extent that its behavior is determined by its immediate inputs and past experience, rather than by its built-in control (Russell & Wefald, 1991). We are interested in investigating the cognitive capabilities of autonomous agents. We believe that cognitive behavior can emerge from the reactive, situated activity of autonomous agents. Some consensus exists about how to design autonomous agents. There should be a relativ..