In my last column, I talked about how and why the Tesla car killed two occupants in a Texas crash a couple of weeks ago. I alluded to some technology that could change the inside cabin game for vehicular data collection. To refresh, on April 14, 2021, the FCC issued an Order waiving its Section 15.255 technical and service rules for unlicensed operation in the 57 GHz to 71 GHz band to permit six equipment manufacturers to operate radar-based vehicle cabin monitors.
The companies include Vayyar Imaging, Valeo North America, Infineon Technologies Americas, IEE Sensing, and Brose North America. These players are looking to use in-vehicle radar to provide vehicular passenger safety and theft prevention applications. And, boy, howdy, could the Tesla have used this.
Such technology has the potential to do much more than just to prevent accidents. It can revolutionize in-cabin safety by using radar-based applications to prevent “hot car” tragedies, such as children and pets being left in vehicles succumbing to vehicular heatstroke.
This has significant implications for next-generation connected and automated vehicle technologies. It could also offer warnings if occupants are not properly restrained. Using a platform called occupant monitoring status (OMS), it can sense airbag deployment and alert first responders. It can determine if seat belts are properly tensioned, offer advanced eCall, monitor vital signs, gesture recognition, out-of-position detection, intruder detection, and more.
The technology is quite impressive. Using a single RFIC the platform can cover up to three complete seating rows and all footwells, as well as the open trunk. The 48-antenna MIMO array offers ultra-high-resolution, providing precision detection and classification data on all occupants as well as their posture and position. Talk about a vertical for 5G mmWave.
And it operates under Part 15 of the FCC rules and regulations meaning it is unlicensed. This is interesting since, historically, general mobile radar operations were prohibited because of the rule’s original intent. That was to “foster the potential of the 60 GHz band for allowing the development of short-range wireless radio systems with communications capabilities approaching those achievable only with coaxial and optical fiber cable.”
The FCC noted, however, that it was “prepared to allow the ‘narrow application of mobile radars for short-range interactive motion sensing’ but at reduced power levels.” In this case, they are open to allowing applications designed to protect life and property in this brave new world of driverless vehicles. The reason a waiver is required is because of the millimeter-wave radar sensors. These are designed to operate at higher power levels than allowed under existing rules.
Sounds like a God-send, does it not? However, as much as this would ratchet up vehicle safety by orders of magnitude, there is the darker side.
We are aware of the controversy around facial recognition. Now we are talking about total movement recognition.
There are many unanswered questions around this, the biggest being the one of what will be done with the captured data? While the players seeking to implement this are not really involved in data management, there is the issue of how captured data would be handled. Would vendors vet the data? Would a third party be involved? Would the data be shipped off to the government, particularly the NTSA or police departments? To me, this has all the makings of a virtual gold mine of data that mega data warehousing resellers like Google and Facebook would drool over to access.
And what about insurance companies, both vehicle and medical, getting ahold of the data? What if there is an accident? Does the data captured become part of a legal issue to determine who was at fault? What if someone is intoxicated and driving? Is there a legal or moral obligation to disable the vehicle or alert the authorities? What if the facial recognition recognizes a wanted fugitive?
Fast forwarding, if this were in effect at the top end, the Texas crash would not have happened. However, given the history of data mining, the players, and their flagrant disregard for privacy, I can see a real can of worms opening around this.
Do not get me wrong. I am 100 percent in favor of this. My reticence is around the data. To be fair, for this to work, optimally, will require the data to be sent to various entities based on its content. But the issues around privacy should be resolved long before this becomes implemented. Plus, there is also the issue of securing the data.
But this column is already long enough to not go there at this time. We all know what those issues are, anyway.
Ernest Worthman is executive editor of AWT magazine and a senior member of IEEE.