| 基于双模式模型预测控制算法的多智能体编队控制 |
| 投稿时间：2015-08-05 修订日期：2016-05-13 点此下载全文 |
| 引用本文：张颖,王明兴.基于双模式模型预测控制算法的多智能体编队控制[J].上海海事大学学报,2016,37(4):82-86. |
| 摘要点击次数: 972 |
| 全文下载次数: 455 |
|
|
| 基金项目:国家自然科学基金（61273068，61673259）；上海市科学技术委员会国际交流合作项目（15220721800） |
|
| 中文摘要:在对多智能体的编队控制上，当输入和系统状态受到约束时，模型预测控制算法比传统的输入输出反馈线性化控制算法具有显著的优势，但传统的模型预测控制算法需要在线优化控制，从而导致巨大的在线负担.为减小这种在线负担，提出一种双模式模型预测控制算法.该算法使用模型预测控制器对控制变量进行在线优化，使得未来某时刻的系统状态进入终端约束集内；此时将系统状态作为输入输出反馈线性化控制器的输入，将系统状态驱动到稳定值；在目标函数中加入避碰函数来有效避免邻近多智能体间的碰撞.仿真结果表明，当输入和状态受到约束时，双模式模型预测控制算法在对多智能体编队控制上比仅使用输入输出反馈线性化控制算法具有明显的优势. |
| 中文关键词:多智能体 编队控制 双模式 模型预测控制 反馈 避碰 |
| |
| Multi-agent formation control based on dual-mode model predictive control algorithm |
|
|
| Abstract:The Model Predictive Control (MPC) algorithm has significant advantages over the traditional input output feedback linearization control algorithm for multi agent formation control when considering the constraints of inputs and system states. But the traditional MPC algorithm requires online optimization, which leads to significant online burden. To reduce the online burden, a dual mode MPC algorithm is proposed, where the model predictive controller is used for online optimization of control variables, making a system state at a certain future moment enter into the terminal constraint set. At this moment, the system state is regarded as the input of the input output feedback linearization controller, and then the system state is driven to a stable value. The collision avoidance function is added into the objective function to effectively avoid the collisions between multi agents. The simulation result shows that the dual mode MPC algorithm has distinct advantage than only using the input output feedback linearization control algorithm in the formation control of multi agents when the inputs and system states are restrained. |
| keywords:multi-agent formation control dual mode model predictive control feedback collision avoidance |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
