一种自治水下机器人垂直面避碰规划方法

为提高自治水下机器人在垂直面跨越坡型、台阶型障碍的能力,提出了一种基于模糊控制的垂直面避碰规划方法。该方法以避碰声纳的输出作为输入,垂直面的深度调节量作为输出,直接建立了从障碍感知到避碰行为的映射。半物理仿真实验表明,该方法有效提高了某型自治水下机器人垂直面操纵性能,具有反应迅速、稳定性好、便于工程实现的优点。

This paper presents a vertical plane obstacle avoidance planning method based on fuzzy control for the advancement in the autonomous underwater vehicle’s capability of maneuvering over the gradual slope or step types of obstacles. The inputs of this method are outputs of the single beam obstacle avoidance sonars, and the output of it is depth decrement in the vertical plane. So the mapping is directly built from obstacle sensing to obstacle avoidance behavior. The semi-physical simulation tests indicate the validity of this method used by a certain autonomous underwater vehicle, and the virtues of rapid reaction, good stability, easy realization.

国家863高科技发展计划资助项目(2006AA04Z262)