A Clamping Perception System Based on Multi-source Sensors for Amphibious Spherical Robots

Robots replacing divers to collect and search underwater is a widely concerned field at present. In order to enable the robot to better complete the grasping task underwater, building a grasp perception system suitable for the robot is helpful for the robot to complete the autonomous grasping task. This paper presents a grasp perception system based on an amphibious robot platform. This system mainly uses four kinds of sensors. They are IMU, pressure sensor, binocular camera, and ultrasonic distance sensor. In order to control the robot to maintain the balance posture in grasping, we combine the data of IMU and pressure sensor, so that the robot can get more posture information. The accurate relative position information between the robot and the target is obtained in real time by fusing the data of binocular camera and ultrasonic ranging. Underwater experiments showed that the robot had good perception ability.