The Shape Perception Method for Space Continuum Robot Based on Structure-Sensor Integration

Continuum robots are widely used in spacecraft interior detection and dangerous environment detection due to their multiple degrees of freedom and adaptability to narrow spaces. However, the shape perception problem of continuum robots seriously affects their closed-loop control and precise operation capabilities. This paper proposes a shape perception method based on a structural sensing integrated module, which helps to achieve high-precision shape perception of space continuum robots. First, the structural sensing integrated module is designed and its kinematics model is analyzed. Second, the data obtained by the sensor are processed by Kalman filtering, and a multiple keypoint error compensation method based on the BP neural network model is proposed. Next, a shape perception method based on Bezier curves is introduced, and a shape fitting accuracy evaluation index considering global information is designed. Then, an experiment is carried out on a shape perception test platform to verify the shape perception method of the continuum robot based on the BP neural network, and the experiments showed that its average relative distance error is less than 2.5%. Finally, slit crossing and docking tests are carried out on the space continuum robot, and the arrival accuracy is better than 3.5mm, which demonstrated the effectiveness of the proposed shape perception method and its potential application in actual space missions.