Industrial Robot


Viewpoint 



Force sensing
A young child can explore and learn and compensate for unknown dynamics by 
prodding, pushing, touching, grasping and feeling. Force sensing and software 
research could soon allow artificial mechanisms to do the same.
Force sensing has its roots in strain gauges, piezoelectrics, Wheatstone 
bridges, automation, robotics, grippers and virtual reality.
That force sensing research has now become commonplace and has expanded from 
those roots to include so much more: video games, athletic equipment, drug 
delivery, braking, computer aided design, autonomous machines such as steerable 
needles and sensing on flexible printed circuits and at the atomic-scale.
The haptic system in humans includes sensing of muscle and tendon states as well 
as tactile sensing of skin deformation. New artificial haptic interfaces attempt 
to mimic those systems by detecting the motion of a human operator without 
impeding that motion, and to feed-back forces from a teleoperated robot or 
virtual reality environment.
Robots typically lack the ability to operate in unstructured and unknown 
environments but force sensing and compliance can help. That work could lead to 
robots that work alongside us, providing assistance to a worker on a production 
line or helping around the house. To achieve this, robots will need to perform 
their tasks within our human world. But our environments are complicated, 
changing, uncontrolled and difficult to detect reliably and then understand.
New generations of force sensors can help to overcome many of these 
difficulties. Indeed, force sensing is already making surgical robots a more 
common sight in hospital operating theatres and many procedures requiring 
accuracy and precision are moving to surgical robots. These surgical systems can 
be more dexterous (and have better vision), but when they are mixed with 
realistic haptic feedback then they may outperform the best human surgeons.
An aim in all this and other force feedback systems is that the operator or 
controlling computer is provided with high fidelity forces and torques. That 
requires compact devices with conflicting capabilities such as low apparent mass 
and inertia and, therefore, low friction (especially static friction) but while 
having structural stiffness and a wide range to provide an even feel to a human 
operator or understandable feedback to a computer.
To achieve this, new robots are being created for force- controlled applications 
with direct-drive brushless motors and crank-rod mechanisms to provide 
low-friction joints without gearboxes. That makes it possible to control forces 
without any extra force sensor as the motors can become the transducer by 
monitoring and then regulating the currents to the motors.
The nano-mechanical properties of materials has become an area of increasing 
importance and at the atomic-scale, structure and reactivity of single crystal 
semiconductor surfaces are being investigated along with the interaction forces 
between molecules. From this work, force sensors for chemicals, viruses, and 
bacteria are being developed along with intelligent surfaces. These can use 
micro-fabricated transducers to detect minute forces.
Ultra-thin, flexible printed circuit force sensors can already be constructed 
with layers of substrate with conductive material and layers of 
pressure-sensitive ink.
As all this moves out from the research laboratories and into the real world, 
then some of the research can be expected to move on to addressing the high 
bandwidth and natural learning required to integrate visual sensors with 
compliant force sensors; all in real time. Those systems will also have to 
provide robust precision without having to rely on complicated perceptual 
modeling and explicit planning. And of course both those sensor systems may only 
have low level signals to work with in a noisy and changing environment.
The author
David Sanders – Systems Engineering Research Group, University of Portsmouth, 
Portsmouth, UK