Protecting cyclists from the rear is obviously a very serious problem and not particularly easy to solve, which is why nobody's done it yet!
So I actually think the reason's no one has done a dynamic system like this based on vision before is mainly because from a technology standpoint, this wasn't feasible until November 14th, 2018. Before that you'd have to use 2x NVIDIA Tx2, one running disparity depth, and on running object detection. And that's $1,400 just for boards, 20W of continuous power, and quite heavy/big.
So I think that's why we haven't seen something like this. Now with low-cost, low-power, and low-volume solutions (like here), making something like this is possible for the first time in human history.
And to be clear, two companies that I'm aware of have made and launched successful products to help keep the cyclist safe from behind:
So they're value proposition is they used correlated risk to have dynamic strobing. So they strobe brighter in intersections and other places they've found to be higher risk to the cyclist (intersections, roundabouts, etc.), based on GPS build into the light. They also report data back to the cloud to help make cycling safer. We plan to do this, but also with video data and actual risk (based on the vision element we have).
I actually tried to get ahold of them about this idea, and potential collaboration, but I'm guessing timezone there was an issue - and I was just some dude sending a LinkedIN request at that point.
So See.Seense I think is a great product. It's more economical, it uses correlated risk-based-on-position, to help reduce the probability that riders will be hit.
It however doesn't have the capability that CommuteGuardian will have, which is to assess realtime risk, and warn you and the driver. So it's kind of a different capability set - but anything that helps to save bikers' lives, we like - so we definitely endorse that product as a lower-cost alternative to help reduce your risk as a biker, and to help make our roads safer for bikers. It's also nice that you could say mount one on your backpack, helmet, and seat post.
So this product is for touring riders exclusively, whereas CommuteGuardian is for bike commuters (hence the name), and See.Sense is likely applicable to both markets.
So the Garmin solution, uses radar to kind of 'wake up' touring riders, who say may be riding for long stretches w/out having to share the road. They're riding in remote, rural areas, most often. It alerts them when there is a car behind them that they'll need to share the road with.
This seems to be a great solution for touring riders, and the folks I've talked to about it, love it. It meets their needs perfectly. And it's intentionally targeted exactly at their use-case, not at commuting (as in a commuting scenario, such an LED display will simply always be on).
It's interesting as it does do this dynamic strobing (like we are doing). So when there are cars behind you, it is strobing to help make your visual presence bigger.
So a great product, meeting the needs of touring riders well (check out their reviews, they're all positive). And another product making bikers safer.
Now back to CommuteGuardian, it fits a different area than both of these, in that it can actually track the trajectory of cars around you, and know actual risk they present to you, and respond accordingly to warn you and also the driver.
Keep the ideas and questions coming!