Sensing sub-surface features can make Lidar cheaper, says startup
The company’s claim is that data extracted from scanning underground features (as deep as 3m underneath the road’s surface) is much more reliable than any data acquired optically by Lidars or cameras, which can be affected by fast changing weather conditions and urban landscapes.
Still, this novel sensing technology would presumably add to the costs of autonomous vehicles and data for the maps has yet to be extensively collected to make commercial sense, eeNews Europe argued when catching up with Tarik Bolat, CEO and Co-Founder of WaveSense.
Quite the contrary, thinks Bolat. “With Lidar, the upper views are dedicated to positioning. So we could reduce the laser beam count together with the field of view of Lidar systems. Our GPR would cost under a hundred dollars at production scale, so it is pretty inexpensive in the current autonomous vehicle sensing suite”, explained the CEO. “When you consider the tens of thousands of dollars spent on Lidar, we would be squashing the overall costs”.
“There isn’t much more juice to squeeze out of the current sensor set” added Bolat, saying that a mapping and car-positioning solution totally independent and uncorrelated from today’s sensor fusion tricks was seducing car OEMs.
“In the next few months, we’ll be choosing one or two partners for commercialization into different markets. One will be the after-market for autonomous vehicles by 2020, and then our solution could be integrated in the autonomous vehicle models of 2023/2024” detailed the CEO.
Now, WaveSense is pretty much starting from scratch as far as mapping data is concerned, so how does it plan to grow and expand its service area?
“Unlike camera or Lidar maps, our solution does not require post-processing, we use the raw data straight away. First we’ll be covering the top ten metropole areas where high-value services such as automated parking could be offered. Boston (only a few miles from where the company is based) has about 2400 lane miles to cover. That’s not significant and we can create these maps. We are not going to map every single road in the US but fill the high-value maps first. Same for parking, there are about 30 to 40 parking locations in Boston”.
Bolat says he received a lot of interest from other markets, such as oil & gas operations, airports, large logistics and distribution centers and other environments that typically have a low feature surface environment in contrast to the rich sub-surface environment that GPR has access to.
For these other markets, WaveSense could design bespoke solutions. Utilities can create maps of sub-surface infrastructures and there are other pockets of value, explained Bolat, although those are not the company’s primary market.
“Most of our pilot activities happen in Boston and Detroit, but we are open to mapping partnerships” answered the CEO when asked if he could envisage to deploy WaveSense’s technology into fleets of autonomous vehicles being trialled or tested extensively on the country’s roads. He anticipates that the final product design could be shrunk to a 67x30cm plate about 3cm thick, to be fixed underneath the car.
According to Bolat, success in deploying autonomous vehicles will all be down to safety, and WaveSense is precisely trying to create a valuable brand around safety. “It is all about the utilisation rate of vehicles. The ride-hailing market needs autonomous vehicles able to operate even in very inclement weather” concluded the CEO, referring to fog and snow covered roads where visual markings can’t be relied upon.
WaveSense – https://wavesense.io
Related articles:
Self-driving cars need ground penetrating radar, says WaveSense
Thermal imaging will make autonomous vehicles safer and more affordable
The autonomous vehicle rollout: current state of play
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
