Research On Modular Joint And Inverse Kinematics Solution Of Tendon-driven Robot Arm

Due to the increasingly severe trend of population aging and the continuous decline in fertility,the elderly care problem is facing enormous challenges in China.The accelerating trend of population aging has brought about two major social issues: labor force shortage and insufficient living security for the elderly.The development and application of robots are the most promising way to solve a series of social needs such as labor force shortage and living security for the elderly.The traditional 7-DOF(seven degrees of freedom)humanoid robotic arm is usually developed from industrial robotic arms,it always has some problems such as heavy weight,low load-to-weight ratio,and poor flexibility.Moreover,the robotic arm configuration is different from the human arm configuration which makes the movement behavior of the robotic arm difficult to predict and is not conducive to rapid integration into human living environments.The closed solution of the inverse kinematics of the traditional7-DOF manipulator with end-joint pose separation can be obtained quickly and accurately based on the analytical method.However,the end-joint of the robotic arm with human arm configuration is coupled,and the inverse kinematics solution of the robotic arm is approximated by constant iteration of the algorithm based on numerical method,which makes the results in low efficiency and insufficient accuracy of the inverse kinematics solution of the robotic arm with this configuration.Therefore,it is of great significance and social benefit to study a 7-DOF humanoid robotic arm with anthropomorphic,lightweight,and high load-bearing capacity.A tendon-driven robotic arm based on multi-motor coupling is improved to meet the development requirements of a new service manipulator.The robotic arm with 7-DOF(degrees of freedom)adopts a modular joint design method,which has the characteristics of a lightweight,compact structure,high output ratio,and a certain flexibility.The kinematic equations of the humanoid robotic arm are established,also the algorithms for accurately and quickly solving the robotic arm workspace and workspace volume are studied.Based on the analytical method,the inverse kinematics closed-form solution and global arm angle optimization are researched.Meanwhile,a control system for the tendon-driven robotic arm is built.The main research content of this thesis includes the following aspects:(1)A tendon-driven robotic arm based on multi-motor coupling is improved to address the problems for different configurations from the human arm and the low load-to-weight ratio of traditional 7-DOF humanoid robotic arms.The robotic arm is divided into three modules:shoulder joint module,elbow joint module,and wrist joint module.The rope transmission characteristics of a single motor driving a single joint are analyzed,based on this,the torque mapping relationship and angle mapping relationship between the motor and joint are derived.The motor layout and wire winding method of two-motor coupling drive two-joint and three-motor coupling drive three-joint models are studied separately.The motor layout and wire winding method of every joint module of the robotic arm are determined.Moreover,the torque mapping relationship and angle mapping relationship between multiple motors and multiple joints in each module are derived.Therefore the decoupling of each joint in the joint module is realized.The correctness of the mapping relationships in each joint module is verified by building an experimental platform for the tendon-driven robotic arm,thereby the correctness of the robotic arm design is verified.(2)A Density-reducing Monte Carlo method is proposed to address the problems of low precision and waste of encrypted point clouds for solving manipulator workspace in the Monte Carlo method and the improved Monte Carlo method.Based on the characteristic of uneven distribution of random points in the Monte Carlo method,the inner and boundary regions of the workspace are clear by uniformly densifying the initial workspace of the robotic arm.Only the boundary region is encrypted by adopting the extended joint and the cyclic encryption of random points,which ensures the accuracy of the workspace and reduces the density of the random point clouds in the workspace.In addition,the influence of the initial point cloud parameter,each axial segmentation voxel parameter,accuracy threshold parameter,extended joint parameter,and cycle number parameter on the accuracy of the workspace are studied.Finally,the effectiveness of the Density-reducing Monte Carlo method is verified by simulation analysis.(3)In order to improve the accuracy and efficiency of the traditional volume method for solving the workspace volume of robotic arms,a full index-solving method is proposed.Based on the voxel grid method,the workspace is discretized into multiple voxels,and the voxels are labeled by a unified full index labeling rule.Based on the error analysis of the workspace volume,the internal and external boundary voxel judgment methods are proposed to obtain the boundary voxels with errors.The boundary voxels are cyclically refined by an exponential refinement method.Finally,the effectiveness and versatility of the full index-solving method are verified by simulation analysis.(4)In order to solve the problem that the closed solution of inverse kinematics of tendon-driven robotic arm with end-joint pose(position and orientation)coupling and base-joint pose separation cannot be obtained by the traditional analytical method,an integrated inverse kinematics method is proposed.The establishment methods of positive and reverse order coordinate systems of the robotic arm are studied,and the D-H parameter mapping relationship between the two coordinate systems is derived.The inverse solution model for the reverse order robotic arm is established by analytical methods,and the closed-form solution for the reverse order robotic arm is obtained.The inverse kinematics closed-form solution for the positive order robotic arm is indirectly obtained based on the D-H parameter mapping relationship.Moreover,the mapping relationship between joint limit and joint arm angle is also derived,the range of each joint’s arm angle is solved by this mapping relationship.And the global arm angle range of the robotic arm is obtained by the intersection operation of these ranges.In addition,the global arm angle optimization algorithm is proposed to address the optimization problem of global arm angle,and the relationship between each optimization factor in the global arm angle optimization algorithm and the global arm angle as well as the optimal combination of optimization factors are studied.Finally,the effectiveness of the integrated inverse kinematics method and the global arm angle optimization method is verified by simulation analysis.(5)A tendon-driven robotic arm control system is built based on the ROS control system.The control system is divided into the upper computer,lower computer,and robotic arm in hardware.On the software,the decision-making layer of the robotic arm control system is preliminarily built according to the Move It framework.Based on the simulation communication framework of Move It and Gazebo,the construction of the decision-making layer of the robotic arm control system is improved,and the real-time communication between Move It and Gazebo is realized.The correctness of the decision-making of the control system is verified by simulation analysis.According to the communication mechanism between ROS and the hardware system,the application layer interface program of the upper computer is written for function modules.And the communication between the upper computer and the lower computer is realized through TCP/IP protocol.The hardware-oriented driver layer program is written in the lower computer,and the communication between the lower computer and the motors of the robotic arm adopts the RS485 interface.Then,the feasibility of the tendon-driven robotic arm control system is verified by experiments.Finally,based on the control system,the end repeatability positioning accuracy of the tendon-driven robot arm and its various joint modules is tested through experiments.To sum up,the research work and related results of the thesis are summarized,and the innovation of the thesis is introduced.Furthermore,future research on the tendon-driven robotic arm is prospectively introduced according to the existing research results.