The classic image of the robot in a factory is a vast arm stretching to perform basic welding, lifting and spraying on bare sheets of metal. But this picture is rapidly becoming outdated. That generation of robotic arms tended to be static, performed only a few fixed functions, and was segregated from human operators and other machines. It had scant awareness of its surroundings and little ability to react dynamically to changes in the environment.
As part of the Industry 4.0 trend, automation equipment, including robots, is becoming more mobile and versatile. It is often connected to cloud computing systems, and is required to be aware of, react to and even learn from changes in its operating environment. This includes the ability to operate alongside humans without endangering them.
The digital element of the design of these robots will be radically different, drawing on increased computing resources and sophisticated software. The analogue and sensor functionality will undergo a similar scale of transformation.
Two fields of sensor operation will be of crucial importance to next generation robots: motion and presence sensing.
In a robot, each segment and joint requires multiple motion sensors to measure position in one, two or three dimensions and to enable precise feedback to the commutation scheme of brushless DC (BLDC) motors.
Where older robots typically consisted of a single arm, new types for the smart factory are mobile, moving on ‘legs’ or consisting of a ‘torso’ on a wheeled pedestal. Many also have ‘heads’ that move and rotate on a ‘neck’. They require more position or motion sensors, often in more compact, mobile designs. The resolvers used in older robot types gave precise position information, but were bulky and expensive.
Magnetic position sensors
New robotic equipment is more likely to use magnetic position sensors in joints and motors. Magnetic sensor ICs, fabricated with standard CMOS technology, are small and operate with a cheap, paired magnet, to minimise sensor system cost. They can provide precise position information – the more precise the data, the higher the torque produced by the motor, the smoother its operation and the more accurate the movement of the robot arm.
The AS5047P rotary position sensor from AMS, for example, provides 14-bit core resolution and may be used with high‑speed motors running at up to 28,000 rpm.
Magnetic position sensors are also suitable for operation in harsh surroundings, such as a magnetically noisy environment like a factory.
Presence sensing, or spatial sensing, plays a critical role in enabling the smart factory robot to safely and rapidly perform tasks. Vision sensing will enable them to ‘see’ and identify objects around them, to navigate and avoid obstacles.
Conventionally, industrial robots use basic 2D visible light cameras because they are reliable and cheap. The latest image sensors for such cameras increase the level of detail that they can capture. For example, the CMV50000 image sensor IC from AMS, contains 47.5Mpixels and captures 30frames/s at full resolution.
Increasingly, however, robots will combine visible light imaging with infrared (IR) light detection and ranging (lidar) systems. Lidar technology, which is also being developed for use in autonomous vehicles, can provide detailed large-scale depth maps of the surroundings. The implementation of precise lidar sensing will depend on the integration of advanced IR vertical cavity surface emitting lasers (VCSELs) and high speed IR photo detectors.
New-generation robots are also likely to use ultrasonic sensors, which are also used in cars as parking sensors. They have a limited range of up to 2m, and a maximum field of view of just 30°. They are relatively cheap and accurate at close range, but susceptible to temperature and pressure changes, and to interference from other ultrasonic sensors nearby. Presence sensing systems in new robots will require a mix of technologies, including vision and ultrasonic sensors.
Safer operation of smarter robots
Innovative sensor technologies will enable new, smart factory robots to operate more independently and more safely. If the flood of data from sensor systems such as lidar is processed by new, more sophisticated computing and machine learning systems, there is also the prospect of robots working collaboratively with humans in a range of manufacturing environments to enhance the productivity of people and of the factory as a whole. oMark Donovan is a senior product marketing manager for AMS AG’s position sensors business line