Commercializing tactile sensors to improve robot dexterity

Tactile Sensors Highly application dependent

At a fundamental level, different applications require different sensing capabilities—temporal and spatial resolution, sensing range and sensitivity, and even the dimensionality of sensing.

Tactile perception of the torso and arms of a humanoid robot may be needed to detect collisions and possibly classify human intentions through physical interactions. For example, gently pushing on the back of a humanoid may indicate that we want it to move forward. This can be achieved with low spatial resolution and one-dimensional tactile sensing such as pressure sensing.

On the other hand, the palms and fingers of humanoid robots require tactile perception to aid manipulation. An ideal tactile sensor for robotic dexterity is modular, has scalable size (for customizing spatial resolution), is highly responsive and has customizable sampling frequency (for customizing temporal resolution) and sensing range, and is multi-dimensional.

Of course, there are trade-offs between some of these features. High spatial resolution, high temporal resolution, and high dimensionality cannot all be achieved simultaneously. And, if they can be, it will have a significant impact on the downstream processing required to use the resulting data. This is one aspect of integration complexity that should also be considered when determining usefulness, along with other aspects of integration complexity that need to be considered including cabling, communications, and power.

Reliability is a must, but robustness may depend on cost and replaceability

Reliability can mean a few different things. I think it means that repeatedly applying the same stimulus to a sensor will produce the same output. Sensor drift, typically caused by temperature (and humidity) changes, electromagnetic interference, and other electrical disturbances (e.g. capacitance), can reduce the reliability of the sensor, thereby reducing confidence in the data and the overall usefulness of the sensor.

Robustness is somewhat related to reliability, as the sensor must be able to withstand repeated stimulation. The number of stimulations the sensor is expected to withstand will depend on the application and the cost and effort associated with replacing part or all of the sensor.

For example, in grocery e-commerce fulfillment, a tactile sensor on a bin picking robot might experience 1,000 cycles per hour. If the sensor wears out on a daily/monthly/yearly basis, then replacing the sensor wear part might be acceptable if the price is low.

Consider affordability as the value of the problem the sensor solves

Affordability of a sensor is an interesting metric that must be considered along with the usefulness of the sensor and the value of the problem the sensor solves. The more useful a sensor is, the more people will be willing to pay for it. The affordability of a useful sensor for a robot destined to build a scientific outpost on Mars will be very different than the affordability of a sensor for a robot that ends up sorting packages in a postal facility.

The right sensor for the right purpose at the right price

In robotics, tactile sensing for the sake of sensing is the wrong approach—it drives up cost, power consumption, and processing requirements, and may provide little return. Tactile sensing must be targeted and planned.

Robotics companies should ask themselves the following: What is the purpose of sensing, and what sensors can be used to achieve this purpose?

The same sensor may be used for multiple purposes in multiple areas of the robot, but if not, then the right sensor must be used in the right place for the right purpose.