Optimization of Tactile Information Grabbing Points Based on Proximity Algorithm (KNN)

In a complex and uncertain environment, there may be problems such as excessive noise when distinguishing object information through visual sensors, which affects the success or failure of the grasping task. These problems can be avoided by implementing grasping tasks through tactile sensors. Therefore, how the robot can accurately grasp unknown objects through its tactile sensing system has become the research goal of researchers. This article takes the robot’s tactile sensor to recognize object information as the research object, and tests the success rate of object grabbing under different k values based on the proximity algorithm, and then optimizes the tactile sensor performance through filtering and denoising to compare the grabbing success rate before and after optimization, and find that filtering after denoising, it can effectively improve the success rate of grabbing non-rotating bodies such as rectangular parallelepipeds and triangular prisms.