Human hands are amazingly skilled at recognizing texture and handling objects with different shapes and sizes. Over the years, although remarkable progress has been made in robotic manipulation, achieving fine tactile feedback (e.g., contact force) and dexterous daily operation (e.g., adaptive grasping) remains a major challenge. One of the main reasons is that artificial hands lack skin-comparable tactile sensors that can detect subtle changes of both normal and shear forces (i.e., self-decoupled) and perceive stimuli with finer resolution than the average spacing between mechanoreceptors (i.e., super-resolved). In this thesis, we describe the design, fabrication, modeling, characterization, and applications of a soft tactile sensor with self-decoupling and super-resolution abilities.

The self-decoupling ability of the sensor is enabled by designing a sinusoidally magnetized flexible film (with the thickness ∼ 0.5 millimeters), whose deformation can be detected by a Hall sensor according to the change of magnetic flux densities under external forces. With such a sinusoidal magnetization pattern, the magnetic field under the film would have two self-decoupled components in terms of the magnetic strength 𝐵 and the magnetic ratio 𝑅_𝐵, by which the normal force and shear force can be decoupled respectively. Similar to the biological hyperacuity of humans, the artificial tactile super-resolution is a technique that leverages overlapping receptive fields of neighboring taxels to perceive stimuli details better than the sensor’s physical resolution. By establishing a theoretical super-resolution model, we have achieved a 15-fold improvement of localization accuracy (from 6 mm to 0.4 mm) as well as the ability to measure the force magnitude. Moreover, the localization accuracy can be improved to 0.1 mm (60-fold improvement) by employing a data-driven method.

This thesis also describes algorithm design for several important applications of the proposed sensor in robotics field, such as adaptive grasping, teleoperation, robotic palpation, surface recognition, and so on. Firstly, adaptive grasping of an egg and teleoperated needle threading are demonstrated by utilizing the force self-decoupling property and super-resolution ability of the sensor, respectively. Secondly, a Bayesian optimization-based algorithm was developed to localize and segment the hard inclusions (artificial tumor) in artificial tissue via autonomous robotic palpation with our tactile sensor, by which the tumor can be quickly localized within 30 iterations of the algorithm and precisely segmented from the surrounding soft tissue with a high sensitivity up to 0.999 and specificity up to 0.973. Lastly, a bi-directional LSTM model with attention mechanism was built to recognize different surface textures, by which a respective recognition accuracy of 97% for Braille characters and 99% for 60 types of fabrics is achieved.

In conclusion, this thesis reports a soft tactile sensor with both self-decoupling and super-resolution abilities, which are comparable to those of the human skin. This research provides new insight into tactile sensor design and could be beneficial to various applications in robotics field, such as autonomous robotic palpation, surface recognition, adaptive grasping, and human-robot interaction.