Development of Vision Based Tactile Sensor Rendering Distributed Contact Force /
Muhammad Raheel Qaisar
- 90p. Soft Copy 30cm
Tactile sensing in robotics plays a crucial role in enabling precise and safe object manipulation, particularly when dealing with fragile or delicate items that require accurate force control. Traditional tactile sensors primarily provide single-point force measurements, which limits their utility in complex manipulation tasks where distributed contact information is essential. As robotic systems advance toward more human-like dexterity, the demand for distributed tactile sensing has increased significantly. Various modalities have been developed for this purpose, including capacitive, resistive, piezoelectric, and vision-based tactile sensors (VBTSs). Among these, VBTSs have gained considerable attention due to their unique advantages, such as high spatial resolution, resistance to hysteresis, immunity to electromagnetic interference, and the capability to measure distributed forces accurately. Despite their promising attributes, the development of VBTSs still lacks a structured and generalized design methodology. This paper addresses that gap by proposing a comprehensive design framework for the development of a multimodal VBTS system. The proposed sensor architecture integrates visual and tactile stimuli using an elastic skin embedded with square fiducial markers, coupled with a depth camera. The core sensing principle relies on tracking the deformation of these markers under external contact, which allows for accurate estimation of distributed contact forces. The stiffness of the elastic skin was experimentally characterized, and this data was utilized to correlate marker displacement with applied forces through controlled indentation experiments. A prototype sensor was fabricated following the proposed framework, and experimental validation was conducted by performing object manipulation tasks. Results demonstrate that the sensor effectively estimates distributed contact forces and can handle fragile objects with precision. This study contributes a robust methodology for developing VBTSs and highlights their potential in advancing tactile capabilities in robotic systems.