Research On Robot Attitude Control System Based On Human Motion Capture

Humanoid robot technology is an interdisciplinary subject covering many disciplines such as bionics,engineering mechanics,sensor technology,embedded technology and control technology.The control of attitude is one of the hotspots of humanoid robot research at home and abroad.Compared with multi-legged crawling robots,humanoid robots have outstanding advantages in replacing humans in high-risk environments.However,the method of robot control is usually to input a fixed sequence of gestures,and embed the algorithm in the process of performing the action,so that the robot has a certain degree of self-regulation ability.When the robot faces a complex environment and the work target has high randomness,it works.The efficiency is greatly reduced.Therefore,this paper proposes a robot attitude control system based on human motion capture.After acquiring the human body attitude,it controls the servo output through wireless transmission,drives the robot to copy the human body motion,and improves the mobility and controllability of the humanoid robot.In order to control the robot's recurrence of human body motion,this paper builds a humanoid robot prototype,designs the overall architecture of the system,studies the human body attitude calculation algorithm research and human-robot attitude mapping relationship,and converts the attitude information of human body into actuator driving signal of robot.The actuator drive signal finally completes the reproduction of the human body posture by the humanoid robot prototype,and the overall performance of the system is verified by experiments.This article studies from the following aspects:1.This paper builds a small humanoid robot prototype and builds a simplified model of the robot space structure.The prototype has 19 degrees of freedom,and the degree of freedom of each joint steering gear is determined according to the overall structure of the prototype and human motion characteristics.2.This paper designs the hardware circuit and software flow of the robot attitude control system based on human motion capture.The system is divided into human motion capture end,robot attitude control end and upper computer end.The human motion capture end collects the human body data through 11 attitude acquisition nodes,and then performs the attitude calculation on the data processing node of the DSP as the main control chip.The attitude receiving node of the robot attitude control terminal receives the human body attitude data and performs human-robot attitude mapping,and the steering gear control node drives the steering gear to output the corresponding posture.The upper computer can display the data of the posture changes of the human body in real time,and the wireless data transmission module is used to transmit the human body posture information between the terminals in the system.3.In this paper,the human body attitude calculation algorithm is studied.The attitude acquisition nodes are initialized and calibrated at the startup time to determine the initial position deviation angle.The measured attitude angle and the initial position deviation angle are calculated to obtain the actual joint attitude change angle.4.This paper studies the human-robot attitude mapping relationship,analyses the human body structure and the degree of freedom allocation of the robot prototype,determines the corresponding relationship between the rotation direction of human joints and the rotation direction of the robot servo,derives the human-robot attitude mapping relationship,and obtains the output parameters of each joint servo of the robot prototype when the motion changes with respect to the human body.5.In this paper,the motion control experiment of the robot attitude control system based on human motion capture is carried out.The experimental results prove the feasibility and practicability of the human motion capture end,the robot control end and the upper computer end designed in this paper.