Vayyar Imaging recently unveiled its second-generation radio-frequency 3D sensor, a "system-on-a-chip" (SOC) that provides high-resolution images for a wide variety of machine vision applications, including robotics, transportation, smart buildings, security, and retail. Engineering.com got an insider's view of the sensor's operation and we're excited to share it with our readers!
Radio Frequency Imaging
RF imaging works by emitting a radio wave and detecting its reflection. Based on the amount of time it takes to pick up the "echo," the receiver calculates the distance to an object's surface. Multiple images, closely spaced, are then used to assemble a 3D view of an object.
Imaging sensors can be used for obstacle detection and collision avoidance systems in human-driven and autonomous vehicles. Smart facilities employ RF sensing to determine room occupancy, helping a building automation system control lighting and HVAC, as well as providing security. Retailers can analyze a customer's every move and use the information to optimize product location on shelves and to modify store layouts. Since the system uses RF instead of cameras, individual privacy is maintained.
A Look Inside
To obtain an accurate 3D view, Vayyar's sensor uses 72 transceivers that emit RF signals of various frequencies - low frequencies (around 3 GHz) to penetrate materials and high frequencies (up to 81 GHz) to detect slight movements and provide angular resolution. Since low frequencies pass through materials, the sensor can "see" through a wall, creating a 3D image of what's behind it. It's not really "X-ray vision," as X-ray wavelengths are about four million times shorter than radio waves, but RF does the job effectively and with a side benefit: nobody nearby needs to wear a lead-filled apron.
Frequency sweeping is used to determine the composition of different materials based on the attenuation of the reflected waves. Each material absorbs and reflects various frequencies in a unique way, much like an object that reflects blue light is perceived as "being blue." By measuring the reflected waves at each frequency, the sensor can identify the composition of substances, for example, to determine the percentage of fat and protein in milk. (I wonder if it can help me reverse-engineer the ingredients in my favorite commercial beers.)
Vayyar representatives wouldn't share the company's trade secrets, but they gave us a pretty good insight into the chip's functionality. RF imaging systems take multiple measurements and perform a probability analysis to determine which reading is the most accurate. To take a measurement, each transceiver sends radio waves in a unique sweeping pattern of frequencies. As the reflected waves are received one at a time, the chip's integrated digital signal processor (DSP) uses a sophisticated algorithm to calculate distances and compile a 3D image based on hundreds of individual measurements. The sensor is precise enough that it can detect a person's position and movements (standing, sitting, falling), as well as breathing patterns, using a combination of Doppler shift and positional changes, which makes it useful in applications such as elderly care, baby monitoring, and sleep apnea detection, with no need for wearable electronics.
The transceivers are built into the chip itself and connected to the 72 antennas mounted on the printed circuit board. Vayyar uses half-wave, omnidirectional antennas, which give a wide field of view. High frequencies are picked up by antennas as small as 2 mm, while 20 mm antennas are tuned to the low frequencies.
For more about the chip's applications, check out Engineering.com's article about Vayyar's first generation sensors and watch the company's video:
Images and video courtesy of Vayyar Imaging
______________________________________________________
Follow Dr. Tom Lombardo on Twitter, LinkedIn, Google+, and Facebook.