Keegan Garcia, Marketing Manager, Industrial mmWave Radar at Texas Instruments
Sensors, either independently, or on machines, act as eyes to see one’s environment — they aggregate information at scale and analyze and act on this data.
Millimeter-wave (mmWave) radar-based sensors are emerging as reliable workhorses in this space. Unlike other sensors based on optical, infrared, or ultrasonic technology, mmWave sensors use radio frequency (RF) so they can operate in challenging conditions such as in blinding sunlight, complete darkness, or in temperature or humidity extremes. Second, they don't invade privacy because they don't depend on visual images. mmWave radar sensors also deliver unique high-resolution sensing of range, velocity, and angle, and offer plug-and-play solutions in a variety of form factors. These advantages deliver endless possibilities for their use in industrial applications.
mmWave radar-based sensors operate in multiple RF bands. Before they can be adopted, it is important to know the use cases and regulations associated with each.
Typically radar sensors operate in the 24-, 60- and 77- GHz radio bands. International regulations affecting the use of each of these bands has been changing. Whereas the 24-GHz band was previously available, regulatory authorities in Europe and the United States are planning to sunset much of the available bandwidth starting in 2022. The 77-GHz band, common in the automotive industry, as seen in Figure 1 with in-cabin sensing child presence and drowsiness detection, is also limited in many countries for traffic infrastructure and level-sensing use cases. That leaves the 60-GHz band as the most open for industrial design applications.
Before being considered for industrial applications, sensors, mmWave radar-based or otherwise, must check off a long list of requirements. These include:
· High range resolution so you can better differentiate the objects you see
· Good velocity resolution so you can tell which way the object under consideration is moving or recognize signatures of motion
· Amenable form factors so they can fit into packaged sensor designs more easily
· Compliance with regional regulations about RF band use
Let's evaluate how mmWave radar sensors fit these requirements and explore the industrial applications which make these best use of these advantages.
The mmWave radar sensor transmits pulses of RF energy that essentially bounces off of objects and from the time-of-flight create a 3D “point cloud” of information. By analyzing the way in which each point in this cloud is placed, and comparing it against a predetermined “standard,” it's possible to detect and count things, including people and even recognize certain events. For example, the point cloud of information about a standing person will look different from one sitting on a chair or lying down. A software algorithm could then analyze this point cloud information at certain intervals, to determine not only when there's a difference in the number of people but also when there's a sudden change in the dimensions of any one person, i.e. whether that person has fallen.
You need better range resolution to be able to better differentiate between various points in that point cloud. This way, instead of seeing just one amorphous blob, you can identify each set of points as belonging to a separate person and track movement accordingly. Operating at the 60-GHz band with 4GHz of bandwidth can give a finer range resolution (3.75 cm) as opposed to 250MHz bandwidth at 24 GHz (60 cm). This means the 60-GHz band offers better differentiation and analysis of point cloud information and leads to more accurate decision-making.
Velocity resolution is necessary to track the direction and speed of moving objects, and even be used to separate certain motion signatures. By using this information, you can differentiate people – or moving autonomous robots - from static or moving clutter in the environment better. For example, in an outdoor intrusion sensor, the velocity signature could be used to separate swaying trees and bushes from a moving person. The general rule of thumb is that velocity resolution is directly proportional to the center frequency. So operating in the 60 GHz range improves the velocity resolution by about 2.5 times versus operating in the lower 24 GHz bandwidth.
Amenable form factors
While sensor selection is important, the total package in radar-based systems also includes an antenna. Among other factors, the antenna configuration affects how far the sensor can find objects and the field of view, i.e. how spread out the objects can afford to be.
Antenna design factors in the wavelength of the radar signal: longer wavelengths will require larger antenna arrays. By this logic, RF signals operating in a higher frequency range have a lower wavelength than those in the lower, 24 GHz range. So, operating in the 60 GHz range delivers an additional advantage: more compact (six times smaller) antenna arrays on printed circuit boards. That means mmWave radar-based sensors can fit in smaller form-factors, or in “cleaner” industrial designs.
The use of antenna-on-package (AoP) based radar sensors helps not only in accommodating the essential components of sensor systems in one compact bundle, it also reduces the dependence on RF expertise. Users can simply mix and match according to needs.
Compliance with regional regulations about RF band use
Even better, pin-to-pin and software-compatible modules allow development on multiple frequency bands simultaneously so deployment for each use case can re-use investment without the need to reinvent the wheel every time.
Given that different global regions have different regulations about RF band use for specific applications, creating a base package for all possible end uses, that can then be customized for RF bands, is much more efficient. Such universality also makes components easier to manufacture and deploy.
Industrial design applications
Now that we know the advantages of mmWave radar-based sensors, we can explore their applications in various fields, as seen in Figure 2:
Click image to enlarge
Figure 2: Examples of industrial applications using the 60- and 77-GHz frequency bands
mmWave radar-based sensors can work as eyes on robots on the manufacturing floor or delivering fulfillment functions in warehouses. Sensors on autonomous robots need a wide field of vision (FOV) so they can detect other robots or humans and avoid collisions. It's also important to determine the rate at which the other object is approaching to develop the response more precisely.
An AoP design typically provides a wide 130-degree view, which is helpful in a variety of industrial applications. Because of this wide angle view, there are fewer “dead zones,” (similar to a blind spot in a car) and an entire solution that covers full 360 degrees around the robot can be enabled with fewer sensors and therefore at lower costs. The smaller form factor of these mmWave radar-based sensors lend themselves to be mounted on a variety of industrial robots, including the more sleek collaborative robots (cobots) and smaller mechanical arms. AoP solutions also eliminate the need for expensive PCB substrates and specific RF expertise making them easier to adopt and integrate into comprehensive automation solutions.
The rise of smart buildings is driven by a desire to optimize operations based on the number of occupants on site.Utilities such as air conditioning and lighting can be programmed to dial up or down accordingly in order to better meet the building demand.
To get there, you need to first calculate the number of people in a building. mmWave-based radar sensors not only perform this function accurately, based on point cloud data, they also anonymize results. Product lifecycle management (PLM) functions of buildings only need be concerned with numbers, not the identity of the people in the building.
Similarly, mmWave radar sensors are useful in monitoring elderly residents who might be at risk for falls. Because the point cloud data can tell when an identified individual changes position, it can predict when the elderly person might have fallen. Because the sensor also measures the rate of change (velocity), it can also determine if the individual has simply gone to bed or actually fallen down. The radar sensor is especially valuable in these applications over other sensing technologies like cameras because of their nonintrusive form factor and adherence to privacy. The sensor also works in low-light conditions which is useful if the elderly resident is stumbling around in the dark.
These sensors can also be deployed on gates to help count the number of people entering through revolving doors or to see if any object is blocking the gate.
mm-Wave radar-based sensors deliver an array of advantages for automation and other sensing applications. Choosing the right bandwidth to operate in is also a cinch with AOP solutions and pin-for-pin packages. In many ways, we are just beginning to scratch the surface with respect to potential applications for this innovative technology.