基于新型多模态触觉传感器的机器人交互物体分类

多模态传感融合对于机器人探索外界环境十分重要，而现有的触觉传感器只能收集一种触觉模态信息，其收集到的多模态信息存在弱配对问题，为了解决此问题，研究了一种基于触觉和视觉融合的多模态触觉传感器。该传感器可以利用一个感知层同时收集2种异构触觉模态信息，弥补了传统触觉传感器的缺陷，同时可以利用收集到的多模态信息对不同物体的几何形状进行分类，在研究过程中，通过按压实验收集了圆形、正方形、长方形和三角形4种形状物体的触觉信息，再利用K最近邻(KNN)算法进行几何形状分类，实验结果证明了该传感器在区分不同物体的形状上具有良好的效果。

Robot Interactive Object Classification Based on Novel Multimodal Tactile Sensor

Multimodal sensing fusion is very important for robots to explore the external environment.However,the existing tactile sensors can only collect one kind of tactile modal information,and the multimodal information collected by them has the problem of weak pairing.In order to solve the problem,a multimodal tactile sensor based on tactile and visual fusion is studied.The sensor can use a perception layer to collect two heterogeneous tactile modal information at the same time,which makes up for the defects of traditional tactile sensors.At the same time,it can use the collected multimodal information to classify the geometric shapes of different objects.During the research process,the tactile information of circular,square,rectangular and triangular objects is collected through the pressing experiment,and then the K-nearest neighbor(KNN) algorithm is used to classify the geometry.The experimental results show that the sensor has a good effect in distinguishing the shapes of different objects.