Integrating behavioral, perceptual, and world knowledge in reactive navigation

  • Authors:
  • Ronald C. Arkin

  • Affiliations:
  • -

  • Venue:
  • Robotics and Autonomous Systems
  • Year:
  • 1990

Quantified Score

Hi-index 0.00

Visualization

Abstract

Reactive navigation based on task decomposition is an effective means for producing robust navigation in complex domains. By incorporating various forms of knowledge, this technique can be made considerably more flexible. Behavioral and perceptual strategies which are represented in a modular form and configured to meet the robot's mission and environment add considerable versatility. A priori world knowledge, when available, can be used to configure these strategies in an efficient form. Dynamically acquired world models can be used to circumvent certain pitfalls that representationless methods are subject to. The Autonomous Robot Architecture (AuRA) is the framework within which experiments in the application of knowledge to reactive control are conducted. Actual robot experiments and simulation studies demonstrate the flexibility and feasibility of this approach over a wide range of navigational domains.