Research On Control System Of Humanoid Robot Based On ROS

With the progress of the robot industry,there are more and more scenarios in which robots are used to replace human beings.Humanoid robot gradually entered people’s vision.This paper selects the type of robot control system based on the self-designed humanoid robot in the laboratory,studies the trajectory planning,collision test of the robot,and tests the control system.Firstly paper summarizes the previous work on laboratory humanoid robots,and determines the control object and expected target of the control system under the existing work progress.And paper analyzes the functional requirements and real-time performance of the humanoid robot control system.On this basis,the control system is designed and layered,the software and hardware platforms are selected according to the results,the communication scheme based on CAN bus and TCP protocol is determined.Finally,the fuzzy control of stepper motor is studied.In order to reduce the pressure on the hardware platform and realize motion control,the tasks are distributed on multiple MCUs,and a distributed control system based on CAN bus is used.The STM32F4 series chip is used as the core platform of the motion control node.First carry out the transplantation of Free RTOS real-time operating system and LWIP communication protocol stack.CAN communication,NETCONN based communication and RS485 communication modules are realized.According to the function of each node,the task flow is designed based on Free RTOS.The kinematics analysis of humanoid robot is carried out,and Moveit is used as the trajectory planner in this paper.Moveit configuration of ROS robot operating system is completed based on 3D model.Trac-IK is used as the inverse resolver and AABB bounding box is used for collision detection.Then the obstacle avoidance planning based on random sampling is studied and discussed,the simulation test is completed.Through the research of trajectory optimization based on polynomial interpolation,continuous trajectory curve is generated by interpolation operation of trajectory points,and using five times the paths smooth spline interpolation method.Realize control-driven development based on action and topic communication mechanisms,connect motion controllers with trajectory planners,the human-computer interaction interface is designed based on QTCreator.Finally,the hardware connection of control system based on ROS is completed,and the joint test of functions of all levels is carried out.Stepper motor control error in 0.005 rad,servo motor control error in 0.001 rad,servo control accuracy in 0.009 rad.