Research On Lattice Distortion Actuating Self-reconfigurable Robotic System And Its Key Technologies

Self-reconfigurable robots have an immense application foreground in the fields such as danger zone reconnaissance,disaster relief,space exploration,duo to the diversity of their structure and function,so that they become a hot topic in robotics research.However,only the combination of multiple modules can reflect the advantage of self-reconfigurable robots.The larger the system is,The more diverse its structural and function are.Self-reconfigurable robots have two basic types,including chain type robots and lattice type robots.Relative to chain type robots,lattice type robots are easier to compose large-scale systems and can best reflect the advantages of self-reconfigurable robots.However,the traditional lattice type systems are static structures and can only move through an inefficient water like motion,and they cannot achieve whole motion without reconfigurating constantly.To solve the problem,this thesis proposes a lattice distortion actuating self-reconfigurable robotic system,which can achieve the whole motion through the local lattice distortion.Then,the principle of lattice distortion actuation,the unit realized technique of modules,configuration optimization and motion control strategy are systematic studied in the thesis.1.For the problem that lattice type robots cannot achieve the gaits of chain type robots,a method that can drive the robot to complete the whole movement through local lattice distortion is proposed in this thesis.First,based on the topology structure features of self-reconfigurable robots and lattice forms of crystals,using the cube as the static form of the distortable lattice is determined.Then,based on the mechanical model of octopus arm,the cubic lattice is made as a parallel mechanism and its degree of freedom is analyzed by screw set method.Finally,a lattice distortion rule is proposed and its kinematics characteristics are analyzed.The distortable lattice is proven to be a “PRP” motion chain and be able to work as a rotation joint.This study provides a mechanism base for the whole movement of the lattice distortion actuating robot.2.Based on the characteristics of modules of self-reconfigurable robots and the symmetry index of crystal unit cell,and the new requirements of achieving the lattice distortion driving,the unit achieved principle for the module of the lattice distortion actuating self-reconfigurable robot is proposed.And then,based on this principle,the module structure of LDSBot is designed.Finally,self-reconfiguration actions and lattice distortion actions are analyzed.LDSBot system is proven to be capability of fulfilling tasks through whole movements,based on the work space analysis of the mixed mechanism composed of distortable lattices.3.Aiming at the problem that self-reconfigurable robots have too many redundancy modules in their configuration to achieve whole motion,a task-oriented configuration optimization method of the lattice distortion actuating self-reconfigurable robot is presented.The optimization is divided into two step.First,based on the mathematic description and the given task,the mechanism synthesis is implemented by using an intelligent algorithm and a configuration with the least motion joints is obtained.Then,the topological optimization based on the strength strategy is implemented by using the KNN method to classify and select the modules,and a configuration with the least modules is obtained.So that the mass of robot are reduced because the redundancy modles are reduced through the optimization.4.Aiming at the problem that too many modules and loops in the configuration make the control too complex,inspired by the neural control system of octopus,a bionics two-layer control strategy for lattice distortion actuating self-reconfigurable robots is proposed.First,based on the structure of LDSBot modules and the adjacency matrix in graph theory,a new configuration description model is proposed.It can only describe the position of module,but also describe the pose of module.This is the foundation for the modules with high symmetrical structure to recognize their role.Then,the master module in the high layer sends massages.A parent-child relationship is formed between slave modules in the low layer based on the sequence of they receiving massages,and a spanning tree is generated,thereby,the problem of too many loops in configuration is solved.Finally,every module acts based on the task command,the role dynamically determined by its own connection state and the action rule in its knowledge data,which let the robot generate whole motion.5.Aiming at the problem that common dynamic simulation softwares cannot complete the dynamic simulations of large-scale self-reconfigurable robots,based on the dynamic engines Vortex and Physx and render engine OSG,a virtual reality dynamic simulation platform—TRIME(Three-dimension Robot In Mechanics Enviroment)is developed.The efficiency of configuration optimization and motion control algorithm are tested and verified through the simulations of large-scale robot system.Similarly,some LDSBot modules are manufactured,and the feasibility of self-reconfiguration motion by modules and distortion motion by lattices are validated.