Tactile sensors are based on a range of different technologies they are utilized in. They’re broadly defined as data acquisition devices designed to sense different properties via direct contact with the object. Robotic applications have expanded their horizon in agriculture, healthcare, and autonomous systems in various unstructured environments where tactile sensors are highly beneficial. Their deployment helps measure, detect, and convert information, mainly acquired by physical contact with the objects.
Although tactile sensing has expanded exponentially in recent years, they are still underdeveloped compared to vision sensing. Researchers predict that this underdevelopment is due to the complexity of the sense of touch.
Applications of robotic tactile sensors
In 1980, robotics was initially determined as the study of perception and action and their intelligent interaction with the environment. Today, touch sensors’ integration has played a significant role in the robust development of flexible and adaptable robots capable of exploring different environmental sectors and safely interacting with humans.
Various robotic platforms use tactile sensors in their developments, like fingerprints, hands, forearms, and torso. This has helped study embodied cognition, perception, recognition, learning, and interaction.
Robotic fingerprints built with tactile sensors are generally equipped with piezoelectric elements, which recognize objects’ properties like shape, texture, etc. Studying and understanding these properties are used to determine experimental procedures like pushing, squeezing, sliding, and tapping of different objects. The use of pressure and force sensors has also played a significant role in object detection and manipulation. For example, the use of tactile sensors in prosthetic hands helps in mimicking the natural motion of a human hand.
Object and shape exploration
Different perception and control approaches are used to determine the object and its shape in various tactile robotic platforms. For instance, the Puma robot is integrated with planar tactile sensors, which help extract object edges and their orientation based on the geometrical moments. A similar use of geometrical moments is used in the KUKA arm with planar tactile sensors, which can recognize the shape and texture of various objects in contact with its fingertips.
Rolling and enclosing exploration
Based on the kurtosis observed on the tactile fingerprint of a five-finger robotic hand, the rolling and enclosing features can be explored. For example, the iCub humanoid robot using tactile sensor information from hands and fingertips achieved higher accuracy in the enclosure procedure for an object recognition task. In recent years, the robotic hand developed by Barrett Hand Inc., which is covered by tactile sensors in its fingertips and palm, has been used to understand dexterous manipulations and study the perception of touch.
There are a plethora of benefits provided by the tactile or touch sensors to develop robots that mimic human tasks. A range of technologies has also been developed around these sensors, each with its advantages and limitations.