Robots can already have superhuman strength, speed, and precision, but what if they could be equipped with x-ray vision, too? I spoke to Ian Podkamien, Director of Business Development at Vayyar Imaging, about his company’s unique product: a radio frequency imaging sensor with potential robotics and machine vision applications.
“Our product has applications very relevant to safety and functionality for industry 4.0 environments,” said Podkamien. “The trend today is toward more and more ‘smart,’ autonomous robots. The limiting factor for how ‘smart’ robots can get is how much data they can collect on their environment.”
How it Works: Radio Frequency Imaging
Even if you aren’t familiar with this technology, you know the basic principle behind radar: The signal propagates in a cone from the emitter, then reflections return to the antenna.
Vayyar has expanded on the concept by combining a high frequency emitter with a large array of antennas in multiple possible configurations. On this sensor, the antenna array collects data in vertical and horizontal axes, as well as inherently in the depth axis, creating a real time, high frame-rate representation of everything happening in the sensor range. The sensor’s chip controls frequency and other parameters and manages the signals from many antennas.
These factors enable this sensor to create a high-accuracy image, amounting to a new type of camera which can see through solid objects, unaffected by lighting or visibility conditions. It also protects privacy, as the sensor collects no visible light. The sensor is low power, being supplied by USB. While the device is available in diverse form factors, most fall between the size of smartphone or a business card, says Podkamien.
Radio Frequency Imaging: Industrial Applications
As Ian said, the limit to what a robot can do autonomously is determined by the input data it can collect about its surroundings. Historically, the two main senses available to a robot have been touch, sensing applied forces and torques; and vision, via cameras, proximity sensors and 3D point imaging. The potential advantages to functionality and safety are many: from maintaining a safe distance between humans and robots in motion, to locating objects, to 3D scanning with access to interior surfaces not visible to laser or optical scanners.
Most sensors today are designed for specific tasks, such as bin picking. However, this new type of sensor trades high resolution visual imaging for a general purpose, widely applicable sensor with broader functionality.
Rather than having one sensor installed for each task, the Vayyar sensor can run several postprocessing applications on the same feed. This means one sensor can not only locate objects for multiple robots to pick at once, but can also monitor human traffic in the area and stop or slow robot motion accordingly. Equipped with this sensor, robots could see through gaylord boxes to identify quantity levels or into assembled parts to make basic quality checks.
Don’t forget, the sensor can do all this in the dark, and in harsh conditions such as splashing coolant, water or particulates. Because the radio waves penetrate through non-metallic objects, the sensor can be placed in a protective box and shielded from the production environment.
Podkamien has his own interesting idea for an industrial application: a wearable bracelet which maintains the safe perimeter around moving industrial robots. This could enable maintenance technicians to work safely in areas such as automotive welding lines, where access is restricted while the large, potentially dangerous machines work at high speed.
Vayyar in the Automotive Space
Speaking of the automotive sector, another exciting set of applications for the technology could be found in, and outside of, passenger vehicles. Said Podkamien: “In the automotive space, we are focusing on both the car interior and the car exterior product capabilities. On the interior vehicle, examples include counting the number of people in the car, tracking their positions, their movement, et cetera. For example, we can detect if a baby was locked in the car. It can alert if someone is not buckled, status after an accident or if the driver falls unconscious or asleep.”
Vehicle exterior applications could include detecting vehicles and obstacles, both while the car is in motion or for parking assist.
Radio Frequency Imaging May be More Flexible
How can the same sensor work on a wrist, making a simple alert, or in a highly complex moving environment running multiple postprocesses on the same data in real time? According to Podkamien, the output of the sensor is SPI, and the use case is flexible. The sensor could be connected to a microcontroller for the most basic functions, an industrial PLC or a high-powered PC for intense processing. The processing power needed depends on the analytics you want to do.
As you can tell, flexibility is the ‘name of the game’ for Vayyar. The sensor is not capable of things like precise scanning for quality and metrology, but over time, the company is looking to increase image quality to unlock possibilities in 3D scanning and other high-resolution tasks. “Flexibility is baked in, as well as specific capabilities available in custom solutions per the customer’s needs,” said Malcolm Berman, Vayyar’s Director of Marketing. The sensor is highly flexible, and the product also has versions in the consumer and medical spaces, with various products with capabilities such as showing the location of pipes, wiring or studs behind drywall, monitoring the safety and health of the elderly and vulnerable and detecting tumours in breast tissue.
What do you think? Can radio frequency imaging replace all other forms of computer vision? Let us know in the comments below, or tweet your thoughts to the author at @engcom_automate.
For more on Vayyar Imaging, check out their video: