| After major disastrous accidents, such as the nuclear explosion accident at Fukushima, Japan, Tianjin hazardous explosion accident, developing a robot for disaster response have become a hot and difficult research field. It is also an important basic issue with important theoretical significance and practical application value. According to these special applications, American Defense Advanced Research Projects Agency (DARPA) hosted the world class advanced robotics challenge (DRC). The primary goal to develop a rescue robot system with the ability to use the rescue tools to implement rescue tasks in real disaster environment. And a Dual-arm robot system is the most likely to complete these tasks.The traditional dual-arm robot with 6-DOF arm is unable to avoid spatial singularity, obstacle avoidance, and so on. To solve these problems, taking the operation ability and strong movement flexibility of human arm as imitation object, the thesis has independent research and development a collaborative dual-arm robot prototype with redundant DOF. With the redundant degree of freedom, the dual-arm robot has the ability to avoid singularity, obstacle avoidance, avoid joint limits, and optimize joint torque under the premise that unchanged the end pose of dual-arms. What’s more, it also enhanced the operation capacity, load capacity and flexibility of the dual-arm robot.On this prototype platform, four key technical problems of the redundant inverse kinematic solution algorithm, movement and force planning and cooperative control were solved. The main research contents and innovation points were as follows:1) The kinematics model and algorithm research for dual arm robot with redundant degree of freedom have been done. First of all, the forward kinematic model for the whole dual arm system was established by using D-H method. Then, considering the redundant features, such as avoiding singularity, avoiding joint limits, a closed-form inverse kinematic algorithm for a 7 DOF redundant robot arm was proposed. The algorithm solved the problems which were difficult to determine the initial value and has cumulative error in the numerical algorithm. The efficiency of the kinematics models and the inverse kinematic algorithm have been verified by simulation and physical experiments.2) A universal cooperative operation motion planning method was proposed. Firstly, considering the motion and force constraints between dual arms, a whole cooperative operation process was decomposed into three modes, namely, dependent cooperative working mode, partly constrained cooperative working mode and full constrained cooperative working mode. Then, the relative position and orientation relationship among dual arm and object in the above modes were derived. And a unified symmetric motion planning method based on the position level was proposed. In this planning method, the motion of the two arms were planned from the objects. The planning method has the advantage that the dual arms have equal status. The motion trajectories of dual arms can be planned and obtained according to the operated objects trajectory at the same time. It also can effectively avoid the problem that the slave arm obtained information delayed in the master-slave motion planning method. In addition, it provided a theoretical analysis model for motion planning and force calculation for the dual-arm robot with redundant degree of freedom.3) The dynamic models of the dual arms and objects were derived and the dynamic optimization methods of internal force distribution were studied. Firstly, the dynamic models of the dual arms, objects and the combined dynamic models of both dual arms and objects were derived in the above three modes. These provided a theoretical analysis model for force planning. Then, the mapping relationships between the motion and force in joint space, operation space and grasping space were given. And considering the dynamic coupling problem in the third mode, three optimal schemes, minimum torque method, minimum norm method and minimum internal squeeze force method, were introduced and simulated. Through the analysis and simulation of the three methods, it can obtained that the weighted pseudo-inverse optimization method can realize zero squeeze force. Finally, a unified force planning method was proposed which provided a force model for controller design.4) The cooperative control and experiments have been done on a dual-arm robot. First of all, taking the relative position and force control as the controlled objects, two model based cooperative control schemes were developed to comply with different operating environment and tasks. One was a position cooperation control method based on kinematics model, the method was used to keep tracking position, relative position between the end of the dual arms. Another is a force cooperation control method based on dynamic model, the method was used to regulated both the target movement and force synchronization control. Secondly, a hardware and software platform of the control system was set up, the method to realize the concurrent execution the program of synchronous and asynchronous movements through a multi-channel cooperative control function module was given. And finally, according to the tasks in DRC, the motion planning and control schemes have been implemented in a dual-arm robot to perform DRC tasks, such as open door, remove debris blocking entryway and so on. Experiments verified the motion planning and control schemes were feasible.In conclusion, combining the methods of the theoretical derivation, simulation analysis and experimental validation, the problems of the redundant kinematics, motion and force planning, cooperative control have been comprehensive studied and experimental demonstrated on a dual-arm prototype. The thesis provided a full theory analysis for researching and developing redundant dual-arm robot arms. It also can greatly improve the cooperative operation ability of the dual arm robot under natural or man-made disasters and laid a solid foundation for it to be applied in human environments.
|
PDF Full Text Request