Research On Optimal Energy Consumption Planning Of Terrain Sensitive Quadruped Robot

The structure of the field environment is complex,and the energy carried by the mobile robot in the process of single task is limited.Therefore,how to use the energy reasonably,so that the robot can complete as many tasks as possible under the support of limited energy has become a hot issue.In this paper,aiming at the energy constraints of quadruped robots,the research status of energy consumption optimization of robots at home and abroad is summarized and analyzed.On this basis,a multi-level grid terrain model is established based on the point cloud data,and the energy consumption model of quadruped robots is constructed based on the terrain factors,so as to provide the basis for the optimal path planning of energy consumption of quadruped robots.The specific contents are as follows:(1)The multi-level grid terrain model is established based on the point cloud data.In order to solve the problem of single structure and incomplete information of existing terrain model,a multi-level terrain model construction method is proposed.Firstly,three terrain factor information of grid height,slope and roughness are extracted from the point cloud.At the same time,the terrain information is obtained based on the GoogLeNet Inception_V3 network training,which is stored in different levels respectively.Then,combined with the mechanical characteristics of the quadruped robot,the trafficability of the grid is determined according to the trafficability evaluation function.The terrain modeling method provides the quadruped robot with complete information of trafficable area and terrain factor for energy consumption calculation.(2)The energy consumption model of quadruped robot is constructed by combining terrain factors.To solve the problem of complex mechanical structure and varied gait of quadruped robot,it is difficult to build energy consumption model.This paper proposes a method of building energy consumption model of quadruped robot by combining terrain factors.Firstly,the mathematical modeling of quadruped robot is carried out,including forward kinematics modeling,inverse kinematics modeling and dynamics modeling.Then,based on the webots simulation platform,the walking data of quadruped robot is obtained,and the energy consumption of quadruped robot is calculated under the condition of asynchronous height,step length and the same distance of walking under different slope conditions.The collected data are regressed in multiple linear ways.The results of the data basically conform to the normal distribution,and the model has good applicability.(3)A path planning algorithm for Quadruped Robot Based on optimal energy consumption.An improved A* algorithm is proposed to help quadruped robot reduce the energy consumption in the process of motion.Based on the previous work,the terrain model provides the required terrain factor information and terrain information,and the energy consumption model provides the calculation basis of robot energy consumption in different terrain.The energy consumption calculation method of quadruped robot is integrated into the node search function of A* algorithm,the objective function is changed,the penalty factor of repeated path is added,and the path with the least energy consumption is selected in the track cluster.The experimental results show that the algorithm can save about 11%~15% energy consumption.