基于人脑行为调控机理的移动机器人智能控制方法研究

项目名称： 基于人脑行为调控机理的移动机器人智能控制方法研究

项目编号： No.61304100

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 赵盼

作者单位： 中国科学院合肥物质科学研究院

项目金额： 23万元

中文摘要： 移动机器人在危险救援、侦察监测、无人作战等野外环境下需要具备高机动能力，以适应复杂地形和瞬息万变的攻防对抗需求。复杂未知地形和高速机动同时出现带来了模型的不确定性和控制稳定性问题，基于动力学模型的控制方法难以满足野外移动机器人的特殊控制要求。经验丰富的驾驶员能够在危险、复杂环境中随机应变，协调自如地高速驾驶车辆，为研究新的控制策略提供了借鉴。本项目拟采用仿人智能控制策略，通过分析人类驾驶员的驾驶行为模型，在对人脑系统行为中枢的信息调控机理进行抽象和模拟的基础上建立一种仿人分层递阶智能控制结构；设计基于小脑模型的神经网络实现对被控对象模型的非线性逼近；针对复杂环境和多任务的控制要求提出基于多中枢递阶协调机制以实现多模态控制及各模态的安全过渡。本项目的研究成果将解决移动机器人在不同轨迹、速度、地形和地质条件下轨迹跟踪的高机动性和高稳定性问题，为提高面向野外环境的移动机器人的控制性能提供新途径。

中文关键词： 移动机器人；仿人智能控制；小脑模型；；

英文摘要： Mobile robot needs to have a high manoeuvrable capability to adapt to the complex terrain and the changing needs of attack-defense countermeasure in the wild environment such as dangerous rescue, unmanned combat and monitoring. The complex unknown terrain and high-speed maneuverability at the same time caused problems on the uncertainty model and control stability, so that traditional control methods based on dynamic model were difficult to meet these special control requirements of mobile robot. The seasoned drivers could drive the car at high speed with the well coordinated and the ability to make immediate response in the dangerous and complex environment. This project choose a human-simulated intelligent control strategy. By analyzing of the human driver's driving behavior model, we set up a simplified and abstract model to simulate information regulation mechanism of the human brain behavior central. Then we propose an algorithm of hierarchical intelligent control to build a neural network base on the cerebellar model, and to establish a coordination mechanisms based on multiple central hierarchical model. The results of this project will solve the problems about the stability and maneuverability of mobile robot in different trajectory, speed, terrain and geological conditions. This research could provide a

英文关键词： mobile robot；human simulating intelligent control；cerebella model；；