Development of a Miniature Photometric Vision-Based Tactile Sensor

This study presents the design of a miniature Vision-based Tactile Sensor (VTS) for compact integration with robotic grippers. VTSs have become a significant research area in the past decade known for their high-resolution capabilities. However, a common challenge faced by the VTS is their large size, which significantly limits the application scenarios. To solve this problem, we in this paper formulate the sensor structure design as an optimization problem constrained by the limitations of the camera and acrylic board and develop an optimally small sensor structure by solving the problem. Meanwhile, we use white LED lights paired with multi-color fluorescent pigment paintings for illumination. The pigment paintings were a more efficient and compact replacement for the traditional multi-color LED lights. The optimization, in conjunction with the pigment paintings, help reduce the sensor size while maintaining the high-resolution characteristics of VTSs. The sensor has a low cost and can be easily integrated into existing robotic systems. For verification, we installed two such sensors on a 2-finger parallel robotic gripper to perform various grasping tasks. The results show that the sensors could satisfactorily reconstruct the depth map of the contact surface. We also compared the proposed design with variations that have different lighting, camera, or mirror setups. The comparison results highlighted the importance of the various components and the effectiveness of the optimization.