基于混杂系统设计的网络视觉伺服系统轨迹跟踪控制研究

In recent years, the Networked Visual Servoing Systems(NVSS) has became the resarch focus of the international academic and engineering communities. This project plans to deal with NVSS three major difficulities of variable sampled delayed and uncertain system parameters, and makes use of a special class of hybrid systems which includs switchings and controlled impulses, called Piecewise Functionning Hybrid Systems (PFHS), to develop state observation and trajectories tracking control methods. Firstly, in view of uncertain continuous undelayed systems, the project tries to find effective ways to improve the performances of Classical Piecewise Continuous Controllers (CPCC) and the CPCC based Recursive Model Free Controller (RMFC). Reminding that the CPCC are derived by using the PFHS impulse effects. Secondly, by considering the variable sampled delayed and uncertain parameters NVSS, the project estimates the systems continuous undelayed state with the applications of PFHS, and combined with the approaches of linearization methods with sliding window, numerical interpolations methods, algebraic techniques and Luenberger state observation theory. Thirdly, the project deals with the NVSS output trajectory tracking control with former research results of state observer and CPCC/RMFC. Finally, the project builds, explores and validates the referred theoretical results on a dSPACE CNC technological based networked visual servoing System. These expected research results can be used to further enrich and perfect the NVSS theory, and provide new ideas and technical ways to resolve trajectory tracking control of industrial production systems, intelligent transportation systems and robot navigation systems under the networked communication conditions.

近年来，网络视觉伺服系统（NVSS）已成为国际学术及工程界的研究热点。本项目拟针对其可变采样、延时及参数不确定三大难点，提出基于包含切换与被控脉冲单元的一类特殊混杂系统，即间断运行混杂系统（PFHS）的状态观测与轨迹跟踪控制方法。首先，针对参数不确定的连续无延时系统，寻找有效方法改进基于PFHS的经典间断运行控制器及基于此的循环无模型轨迹跟踪器。其次，针对可变采样、延时且参数不确定的网络视觉伺服系统，基于PFHS结合滑动视窗线性化、数值插值、代数方法及Luenberger状态观测理论重构系统的连续无延时状态。再次，结合上两部分研究成果，进行输出轨迹跟踪控制研究。最后，本项目拟通过新一代dSPACE数控技术构造实物在环仿真平台，验证上述理论方法的有效性。预期成果将进一步丰富和完善NVSS理论，可为工业生产、智能交通与机器人导航系统在网络通信条件下的轨迹跟踪控制提供新的思路与技术途径。

近年来，网络视觉伺服系统（NVSS）已成为国际学术及工程界的研究热点。本项目针对其可变采样、延时及参数不确定三大难点，提出了基于包含切换与被控脉冲单元的一类特殊混杂系统，即间断运行混杂系统（PFHS）的状态观测与轨迹跟踪控制理论。首先，针对被控系统参数不确定的连续无延时系统，基于PFHS，发展了衍生的DPCC控制器以及循环迭代无模型轨迹跟踪控制器（RMFC），进一步引进L2最小化范数改进其结构与参数，发展了具有双环结构RMFC。并针对广义非线性系统，引入极局部模型降阶、时延估计等系统发展了新型无模型控制器。其次，针对可变采样、延时不确定的NVSS，对其采样时延输出的采样与时延三种不同关系，系统提出了一种基于PCHS，集成代数方法及Luenberger观测器的系统状态连续无时延的间断连续状态观测理论。再次，结合上两部分研究成果，提出了NVSS的输出轨迹跟踪控制策略。然后，针对具有采样、延时以及系统不确定的NVSS，引入事件触发机制，系统提出了基于事件触发的间断连续网络轨迹跟踪控制策略。最后，本研究通过新一代数控仿真手段(dSPACE-Matlab/Simuilink，FAST，LMS AmSIM, ADAMS等)验证了上述理论成果，并成功应用到相应的网络视觉系统、多源可再生能源优化系统、四旋翼飞行控制系统、11自由度外骨骼机器人系统、以及机器人伺服系统中。相应的研究成果将进一步丰富和完善NVSS理论，为工业生产、智能交通与机器人导航系统在网络通信条件下的轨迹跟踪控制提供新的思路与技术途径。