Reading time ( words)
In today’s ever-increasing robotic world, future robots will need more human features to properly integrate comfortably into our society, especially if they are to have direct skin to skin contact with humans. With the added power of artificial intelligence (AI), robots will be able to increasingly perform more human-like functions. One example of improved sensory applications is the fingertips on a robotic hand. Through touch, the hand requires it to sense temperature, pressure, pain, and even air flow for proper human-like operation. Inventive applications of flex circuits and simple sensors integrated with robotic skin like silicon can emulate a human’s highly sensory fingertips.
To sense touch, we can use a resistive layer and connections to a microprocessor, which would be similar to a touchpad on your laptop. The resistor touchpad would allow the computer to know the position of the touch, as well as pressure sensing. The resistive layer could be printed on a flex layer with a skin-like latex, protective outer surface. Pressure would change the resistance, telling the AI smart robot to squeeze lighter or heavier. Alternatively, they can use capacitance, like the touch-key screen on your computer, to sense the position of touch on their fingers, as well as pressure.
To sense temperature, we can silk screen or implant resistive or chip temperature sensors on the flex layers connected to a microprocessor to measure finger temperature due to resistive or pulse output changes. A layer of thermally conductive skin material could transport the heat inward quickly. The microprocessor would compare the resistive reading to a look-up scaling chart and quickly generate an accurate temperature reading.
When the fingers of our robot touch something hot, our robot can assess if they are holding something too hot for the silicon skin. At that point, the robot can decide to suffer finger damage if needed to save a human life or protect property, unlike humans, who instinctively jump back quickly and drop the hot object. For example, if the robot is working in the lab, picks up a hot beaker of sulfuric acid, and senses that it is so hot it will damage its silicon skin layer, rather than drop the acid and possibly hurt humans or do damage to property, the robot can simply place the hot object back and walk over to the repair shop for new finger sensory pads.
To sense pressure, the flex circuit can have carbonized rubber pads attached to gold pads between the flex layers, or small integrated chip pressure sensors. As the finger pads press on an object, the resistance of the carbonized rubber will decrease, or the chip will send a signal regarding the pressure applied. As the softer carbonized rubber pressure sensors compress flat with increasingly more pressure, a series of harder, durometer rubber pads will continue to react to stronger pressures such as the difference between your robot picking up an egg or your bowling ball.
The AI computer would look at the object it needs to pick up, judge the size and weight, and adjust finger pressure accordingly. Known as tactile feedback, the robot can control pressure on their fingers, knowing which object they are to pick up, the fragile egg versus a hard, heavy bowling ball. The flex circuits can have capacitor sensors, which would offer an advantage over humans in that it could sense how close the fingertips are to an object by measuring capacitance change.
For pain or damage sensors, a series of fine 0.5-mil copper tracks on the outer layer flex circuit or embedded in the skin coating would open if the finger pad is severely damaged. The tracks would be covered with a thin, soft silicon or latex rubber to act as skin. The pain or damage tracks could be shaped similar to your fingerprints. The height of the copper or some other metal pain tracks would cause the silicon to follow the pain-sensing tracks and protrude with a similar shape, allowing for robot identification. Should your robot run amuck and rob an electronics parts store for a snack, it will leave its fingerprints all over the store.
To make the robot feel more human to us, and if they are required to touch you, the robot’s skin should be around 95°F. A series of screen-printed carbon resistors placed on one of the outer flex layers, with a little current and temperature feedback sensors, would heat the silicon skin slightly to the touch. The temperature sensors would allow the microprocessors to keep the skin heat constant and yet still measure fingertip contact temperature. Similarly, they could use Peltier cooling diodes or even cooling/heating coils embedded within the robot’s hands to keep the skin cool in hot climates and to disperse any heat generated by the electric motors.
The heat resistors can also measure air flow, just like your hands do, with a heated resistor and a non-heated resistor. Any air flow will cause a rise in current required to keep the temperature constant, and you can measure air flow as hair and skin do. There is not much room inside a human’s finger for too many tubes, wires, or sensors. The small space is needed for all the electric muscles and the joints. This is where flex circuits shine, as they can easily be made very small, thin, and very flexible as well as loaded with electronic sensors.
I predict the next generation of flex circuits may even be molded right into the robot’s silicon skin, increasing space available inside the fingers for little motors and skeleton structures. If you order the new SE 3000 version personal satisfaction robot, you will want as many tactile sensors as possible, as well as heated skin (just kidding about the SE 3000 robot). However, this column shows it is possible to use flex circuits for many other purposes.
This column originally appeared in the November 2020 issue of Design007 Magazine.