基于包容结构的移动机器人混合控制结构研究

Hybrid architecture of mobile robot based on subsumption architecture

体系结构是机器人系统研究的一个重要内容,针对传统的模式存在对真实环境的适应能力差、处理应激避障行为和直接规划行为过程复杂等一系列问题,提出一种以包容式框架结构为基础,将行为规划与行为输出相结合的混合式控制结构.行为规划层以“注意力”调整策略为决策方法,通过不同时刻对各层行为注意力不同,决定由哪一层控制机器人,较完整地保留了子行为的决策意图,得到更合理的行为综合结果.仿真试验采用了包括漫游、避障、行为管理在内的6层控制结构,在不同情况下对机器人运动轨迹进行了模拟,试验结果证明提出的混合结构是简洁有效的.

Control architecture is an important investigative content in mobile robots. After studying the subsumption architecture, a new hybrid architecture is introduced, which is based on the subsumption architecture and integrates the behaviour planning and behaviour output. The process of behaviour planning uses the concept of attention to coordinate the behaviours. It decides the control layer through the degree of attention which can hold the intention of sub-behaviour perfectly. The simulation uses 6 layers of control architecture including wander,avoiding obstacle, behaviour planning,etc. The results show this architecture is simple and effective in some different situations.