国产欧美精品一区二区,中文字幕专区在线亚洲,国产精品美女网站在线观看,艾秋果冻传媒2021精品,在线免费一区二区,久久久久久青草大香综合精品,日韩美aaa特级毛片,欧美成人精品午夜免费影视

Aiming at the problems of existing control algorithms, such as large trajectory control deviation and poor collision avoidance ability, the control process of hexapod robot is studied, and a control scheme based on visual inertial SLAM algorithm is proposed. The D-H model is studied, and the spatial motion process of the hexapod robot is analyzed. The HD camera is used to acquire the images of the job site and realize the coordinate conversion of the HD camera at the same time. The IMU unit is used to improve the stability of the SLAM algorithm model and carry out the calibration of errors and parameters. In the image feature set training, the CNN network model is used to improve the data training ability of SLAM algorithm model, and according to the maximum discount reward value after interacting with the field environment, the gait stability of the robot and the collision avoidance effect of the local area are improved. The experimental results show that the visual inertial SLAM algorithm can achieve the trajectory path optimization in the global scope, the path time is only 35.4s, and the number of collisions with obstacles is 1 in 10 collision avoidance tests, which is far better than other collision avoidance control algorithms. Therefore, the proposed algorithm can better meet the special needs of industry and engineering.