类人双臂移动机器人整体性运动规划及基于图像和深度信息融合的目标抓取机理研究

As the realization of expanding the fixed- and single-manipulators, the high redundancy humanoid dual-arm mobile Robot service acquires greater workspace and better smoothness, more flexibility in the process of the target tracing and the mobile operating as the more redundancy. The humanoid dual-arm mobile service robot is a system of high redundancy with nonlinearity, multivariate and uncertainty. To realize the non-cooperative target tracing.Firstly, a high redundancy humanoid dual-arm mobile service robot is built; Secondly, a whole cooperation control algorithm-whole Resolved Motion Rate Control algorithm (wRMRC) which decomposes the high redundancy humanoid dual-arm mobile service robot end-effectors into the movement of the moving platform and each 3D manipulator joint harmoniously through the inverse kinematics equation is applied to the motion planning of the robot; Thirdly, the target recognition and the grasping mechanism based on the image and the depth information fusion is studied on the robot; Fourthly, an improved A* algorithm based on the enhanced probability map in the 3D C space is adopted the path planning of the robot. At last, the 3D head pursuit technology is applied to the target tracing of the high redundancy humanoid dual-arm mobile robot. Under the condition that the parameters selection and the initial conditions limitation are satisfied with head pursuit technology, the locomotion performance and the configuration of the high redundancy humanoid dual-arm mobile robot are analyzed. At last we will test on the robots in the lab. One of them functions as the humanoid dual-arm tracing mobile robot and another as the non-cooperative target robot. At the same time we adopt the binocular stereovision to observe the information of the non-cooperative target and test the effectiveness of the wRMRC algorithm and the 3D head pursuit method. By the comparison between the single- and humanoid dual-arm mobile robots based on this algorithm, the humanoid dual-arm mobile service robot possesses the more redundancy and is able to track and operate the different objects at the same time. Moreover, the redundant humanoid dual-arm mobile robot has the characteristics of better smoothness, more flexibility, greater operational space and the harmonious cooperation between two arms and the differential driving platform during the whole mobile operational process. That lays the foundation for practical application of the high redundancy mobile humanoid dual-arm mobile service robot.

高冗余类人双臂移动服务机器人由于实现了对固定机械臂和单机械臂的扩展，使得机器人拥有更大的操作空间，并同时具有移动、对目标进行跟踪和操作的功能。类人双臂移动服务机器人是一个高冗余和非线性不确定性系统，本研究为了实现对非合作目标的快速跟踪和抓取操作，首先构建一个高冗余类人双臂移动服务机器人系统；采用整体性协作控制算法（wRMRC）进行运动规划；采用基于图像和深度多传感信息融合的目标识别和抓取机理研究；三维C空间增强概率地图改进型A*算法进行路径规划；并将三维前置追踪技术应用在该机器人系统的追踪非合作目标上，在研究满足前置追踪的参数选择和初始条件限制条件下，分析其对高冗余类人双臂移动服务机器人的运动性能影响及配置要求。本申请将在高冗余类人双臂移动服务机器人上进行算法验证：一台作为追踪平台，一台作为非合作目标平台。采用视觉系统和激光轮廓传感器对非合作目标信息进行观测，用实验验证算法的有效性。

类人双臂移动服务机器人是在非结构环境下为人类提供必要服务的多种高技术的集成体。类人双臂之间以及与移动平台之间存在着高度的协同关系，双臂协调操作是提高机器人系统操作能力、负载能力、可靠性及扩展操作空间的有效途径。类人双臂无碰撞路径规划方法可以使机器人类人双臂运动过程中互不相碰，扩大了机器人工作空间，可应用于家庭移动类人服务机器人和与机械手臂生产实践相关的领域，具有很广阔的研究和实际应用价值。.基于多模态信息融合的物体识别和抓取规划一直是图像处理、机器视觉、深度信息处理、机器学习、人工智能等领域的研究焦点。本课题以基于激光传感器的深度信息和视觉传感器的图像信息进行融合进行运动目标检测用于对运动目标的跟踪抓取操作，在移动服务机器人领域有着重要的意义与应用价值。.机器人的运动规划经过多年发展，形成了多种不同的规划方法。考虑到机器人抓取的实时性要求，如何提高高自由度机器人的运动规划效率仍然是一个需要解决的问题，如何确定移动平台的目标位置以使得该位置处多机械臂能更好地协调完成抓取操作工作尤为关键。.本课题构建了一个高冗余类人双臂移动服务机器人作业系统，针对非结构外部环境和运动目标，对类人双臂移动服务机器人的模型构建、整体性运动规划、基于多传感信息融合目标识别和抓取、路径寻优、目标前置跟踪以及目标抓取操作等关键技术进行了研究，实现了高冗余类人双臂移动服务机器人对运动目标的跟踪和抓取操作任务。服务机器人工作环境是一个非结构化的复杂环境，环境、机器人本体及类人双臂末端执行器都处在一个动态的过程中，因此研究高冗余移动类人双臂机器人对运动目标的跟踪和抓取操作具有很重要的实际意义。整体性运动规划和路径寻优以及跟踪操作是智能移动服务机器人的几个关键技术问题，本课题结合具体的高冗余类人双臂移动服务机器人对这四个关键技术进行了研究，为类人双臂服务机器人走向产业化建立一定理论和实践基础。