Remote Motion Control System Research On Bipedal Robot

The study of bipedal robots has been carried out since the bionic robot appeared synchronously.With the development of embedded chips and the emergence of more efficient controller,researchers can control bipedal robot more flexible.Compared with other types of robots,bipedal robots have the advantages of Flexible motion and strong applicability,which can move in a variety of complex terrain,and across the obstacles.It can help or replace humans to complete a lot of work,such as carrying objects,disaster relief,high-altitude operations,etc.As the movement of human is a complex and sophisticated process,therefore,controlling the robot to achieve stable walking is still an unsolved problem,which is the biggest threshold that bound bipedal robot out of the laboratory.Based on the current research on bipedal robot,this paper proposes a bipedal robot control system combined with remote control platform and Motion-sensing technology.The system combines somatosensory technology,simulation technology,remote measurement and control technology and intelligent control technology.The controller's data acquisition,the calculation of the robot motion trajectory,the optimization of the attitude data,the foot force measurement and the steering gear control are realized.The system also separates the console from the controlled robot by the internet remote control system,and realizes the visual field feedback of the robot by using the virtual reality technology.In this paper,the study of bipedal robot mainly includes the following contents: Using Microsoft's Kinect sensor to capture human pose data;Using NI's CompactRIO and accompanying software platform LabVIEW as a bipedal robot controller to achieve remote control;Using MATLAB to realize the kinematic model of the robot,the trajectory calculation and the stable data preprocessing;Using the pressure sensor and SCM to achieve the actual ZMP measurement and transmission;Comparing actual ZMP with the expected data to adjust the unstable attitude and achieve closed-loop control;Aiming at the real-time control requirement of this system,this paper also proposes to use RBF neural network to deal with the data,to promote the system real-time,offer the robot self-learning ability and improve the stability of the robot;Using SolidWorks to draw the robot 3D model and use Lab VIEW to achieve motion matching;Using the virtual reality technology and binocular camera to transmit the robot vision to the controller.Through the construction of the hardware circuit and software debugging,the system can achieve the movement of robot stabilized and balanced.The improvement of the algorithm and the application of the intelligent algorithm greatly shorten the processing time of the robot data and realize the stable synchronization between the robot and the controller at constant speed.The design of the control system meets the design requirements and the system can run stably.