Autonomous Robots

, Volume 1, Issue 1, pp 69–92

A bayesian approach to real-time obstacle avoidance for a mobile robot

  • Huosheng Hu
  • Michael Brady

DOI: 10.1007/BF00735343

Cite this article as:
Hu, H. & Brady, M. Auton Robot (1994) 1: 69. doi:10.1007/BF00735343


Real-time obstacle avoidance is essential for the safe operation of mobile robots in a dynamically changing environment. This paper investigates how an industrial mobile robot can respond to unexpected static obstacles while following a path planned by a global path planner. The obstacle avoidance problem is formulated using decision theory to determine an optimal response based on inaccurate sensor data. The optimal decision rule minimises the Bayes risk by trading between a sidestep maneuver and backtracking to follow an alternative path. Real-time implementation is emphasised here as part of a framework for real world applications. It has been successfully implemented both in simulation and in reality using a mobile robot.


mobile robotsBayesian approachobstacle avoidancesensor model

Copyright information

© Kluwer Academic Publishers 1994

Authors and Affiliations

  • Huosheng Hu
    • 1
  • Michael Brady
    • 1
  1. 1.Department of Engineering ScienceUniversity of OxfordOxfordU.K.