Tactile Sensing in Robots

Image Credits: Willyam Bradberry/shutterstock.com

At the moment, robots cannot demonstrate behavioral emotions, and the design of humanoids that can represent human emotions can be complicated. Ideally, emotions in robots would improve their dialogue and performance in a social domain.

Synthesizing adaptive capabilities typically seen in humans is particularly challenging. For features like behavior navigation, emotional control, and personality, we are all still only scratching the surface of how autonomous robots could work.

On the other hand, physical feeling has been successfully implemented into robot systems involved in sensors, actuators, and software applications. These physical signals to allow a robot to feel an item and identify the object being touched.

Work carried out by researchers at the USC Viterbi School of Engineering at the start of the decade involved the construction of a robot that could identify a range of objects by feeling the texture of the material. The researchers built a robotic finger with tactile sensors embedded in the fingertip. By using specific algorithms designed to simulate human problem-solving strategies (e.g., algorithms that control exploratory movement), the robotic finger was able to distinguish different materials. For example, it was able to identify the difference between the texture of cotton and the texture of wool.

The tactile sensor used in this robotic fingertip contained a conductive fluid. During a simulation, the fingertip would slide over a surface to a different object or texture, which then triggered vibrations through the conductive fluid. These vibrations forced a change in the electrical current, which was then detected by electrodes in the fingertip. The research has demonstrated that functionality can define the use of a prosthetic hand by applying tactile sensors to this type of mechanical system. The video below describes the application of this BioTac tactile robotic sensor.

Tactile Sensors

There are several different tactile sensors, used for a range of applications, such as sensing normal pressure, skin deformation, and dynamic tactile sensing. They are one of the most common sensors used in robotics and include piezoelectric, piezoresistive, capacitive, and elastoresistive types.

The functional components of a standard tactile sensor include a micro-switch, which is a switch that is sensitive to a varying range of movement. It is an array of touch sensors that make up a larger sensor known as a tactile sensor.

An individual touch sensor will interpret the physical contact between the robotic finger and the textured surface. As soon as the robot has made contact with an item, a signal is transmitted to a controller.

Each touch sensor will be sensitive to more than one parameter. Each touch sensor will relay different information about movement, the texture of the surface, and the shape, size, and type of object. As illustrated in Figure 1, there usually are six touch sensors at either side of a tactile sensor. The touch sensor, mentioned here as a switch, is made up of a plunger, an LED, and a light-sensitive sensor.

During the contact of a tactile sensor with the surface of an object, the individual touch sensors detect different information about the size, shape, and texture of the object and transmit an electrical signal to a local controller. This controller then measures the real size, shape, and feel of the object.

Dynamic Tactile Sensing

Dynamic tactile sensing uses the functional principle of a tactile sensor mechanism to an intricate level; focusing more on the control over an object and being able to detect any slipping.

For a robotic hand to improve slip detection, the array resolution and scanning rate must be at its best to identify the movement of a grasped object before losing contact with it. One way to achieve this is by measuring the degree of vibration generated through the force of grasping an object, and this is one parameter that has been used to stimulate an electrical surge across conductive fluid to trigger electrodes and send information about object contact to a controller.

Piezoelectric polymer transducers are ideal for detecting vibration and work by sliding across the surface of an object before transmitting an electrical stimulus to the controller. Without tactile feedback, a robot will not be able to have control over the gripping force. Conversely, by detecting the slip of an object, a robot can re-establish its grip on an object.

Tactile feedback will be a crucial learning point for a robotic system; an unlikely concept without the use of intelligent sensor networks to help gather information about contact with an object. Tactile sensors in robots will help push forward human-robot interaction.

For example, currently, there is research being applied to developing the understanding of tactile feedback in surgical robots. Although the force sensing and actuation technology will need to be developed further before taking this to an in-vivo clinical setting.

Recent advances in tactile sensing have seen some alternative approaches to this problem. For example, BathTip only relies on a simple band and a cheap optical camera, like those found in most modern smartphones and laptop computers.

There are also recent moves towards open hardware approaches, such as those offered by the company TakkTile. This allows enthusiasts and hobbyists to play around at home with easily accessible and affordable materials, which in the past would have only been available to private research labs and institutions.

The future of tactile sensing will not only see it implemented into robot systems, but also into VR applications; where interaction with virtual worlds can provide feedback to a human user about simulated objects in their virtual environment, such as textures and feelings. Seattle-based startup HaptX is currently developing VR technologies with this in mind, such as their tactile feedback gloves which can also be used to operate robotic arms remotely.

Sources and Further Reading

• Rosen, J., Hannaford, B., Stava, R.M. (2011). Surgical Robotics: System Applications and Visions. New York: Springer Science and Business Media, LLC.
• Vallverdú, J., Casacuberta, D. (2009). Handbook of Research on Synthetic Emotions and Sociable Robotics: New Applications in Affective Computing and Artificial Intelligence. United States: Information Science Reference (an imprint of IGI Global).
• Siciliano, B., Khatib, O. (2008). Handbook of Robotics. Berlin Heidelberg, Germany: Springer Science and Business Media.
• University of South California Viterbi School of Engineering.
• Corradi, Tadeo, Peter Hall, and Pejman Iravani. ‘Tactile Features: Recognising Touch Sensations with a Novel and Inexpensive Tactile Sensor’. In Advances in Autonomous Robotics Systems, edited by Michael Mistry, Aleš Leonardis, Mark Witkowski, and Chris Melhuish, 163–72. Lecture Notes in Computer Science 8717. Springer International Publishing, 2014.
• VR Startup HaptX

This article was updated on the 26^th July, 2019.