A new crop of early-stage startups and some recent venture capital investments indicate that a niche market is emerging in the self-driving vehicle technology space. Unlike companies that are bringing robotaxis to the streets, these startups are taking their technology off-road.
Two recent entrants, Seattle-based Overland AI and New Brunswick-based Potential, are looking to gain first-mover advantage in this self-driving space.
Although these startups apply their technologies in different ways, Overland AI and Potential have something in common in the off-road space. The startup founders aim to tackle one of the more challenging applications of self-driving by building software that doesn't rely on some of the key elements of testing and deployment, such as detailed maps and large amounts of training data. I believe I have cracked the code. You can also rely on her remote assistance.
The US Department of Defense and venture capital investors are taking notice.
Overland AI, which develops self-driving systems designed for military operations such as reconnaissance, surveillance and delivery of electronic warfare packages, won an award of up to $18.6 million from the U.S. Army's Defense Innovation Command in April. The funding will be used to build a prototype autonomous software stack for the Robotic Combat Vehicle (RCV) program over the next two years.
The startup, founded in 2022, raised a $10 million seed round this week led by Point72 Ventures. CEO and founder Byron Boots said the funding will be used to expand Overland's team and continue development of the company's autonomous stack, OverDrive.
Meanwhile, Potential, which develops advanced driver assistance systems (ADAS) that enable ATVs, underground mining vehicles, and passenger vehicles to handle off-road environments, has raised a seed round led by Canadian early-stage VC firm Brightspark Ventures. The company raised an additional 2 million Canadian dollars (approximately 1.5 million USD). This brings Potential's total funding to C$8.5 million (approximately US$6.2 million). The startup has been working on technology development for the past six years and currently has several pilot projects in the powersports, motorcycle and automotive sectors.
Off-road opportunities
Potential AI and Overland AI aren't the only companies looking to take self-driving vehicle technology beyond public roads. The pursuit of high costs in operating commercial robotaxis and self-driving trucks has frustrated dozens of startups over the past few years. As these companies shut down, new startups such as Polymath Robotics, Forterra, Pronto.ai, Bear Robotics, and Outrider are taking a more grounded approach to applying AV technology to warehouse, mining, industrial, and off-road environments. He appeared with a strong sense of ambition.
“We are definitely putting capital into off-road autonomous driving,” Alexei Andreev, managing director of Autotech Ventures, told TechCrunch. “In fact, we are moving away from highway autonomy and fully ramping up off-road autonomy.”
Most of the off-road companies in which Autotech Ventures currently invests are in the agricultural and construction sectors, with products such as self-driving mining vehicles, forklifts, and tractors. Andreev says it is important for these sectors to address labor shortages while increasing productivity and making farms and construction sites safer.
“And when you lay people off, your premiums are immediately reduced. So the ROI for these vertical applications is current and significant,” Andreev said.
Another result is that off-road autonomy has found a friend in defense.
Overland AI: From DARPA to seed funding
Overland AI's off-road autonomous driving software OverDrive is being tested for defense and national security applications. Image credit: Overland AI
The US military could be a great customer when it comes to automating off-road driving. After all, self-driving cars started as his DARPA project, says Jeff Peters, a partner at Ibex Investors. DARPA (Defense Advanced Research Projects Agency) is an agency of the United States Department of Defense focused on advancing technology for military applications.
“While the hype surrounding AVs has moved much of the industry toward greater potential for commercial use, the Department of Defense project continues,” Peters told TechCrunch via email, adding that the autonomous mining startup He noted that SafeAI and self-driving trucking startup Kodiak Robotics are also seeking defense grants. “I think the AV companies (those that are still around) will pursue DoD projects because the DoD will provide significant, non-dilutive funding for the interim period until commercial operation. ”
Overland AI is the latest byproduct of the DARPA program. Boots is a professor of machine learning at the University of Washington and founder of the Robot Learning Institute in the university's School of Computer Science and Engineering, and has a long history of collaboration with the U.S. Army Research Laboratory and DARPA.
Overland was spun out from Boots' research and team involved in DARPA's RACER (Resilient Robotic Autonomy in Complex Environments) program, which aims to develop self-driving vehicles capable of tackling harsh terrain. Ta.
The program is still ongoing. Overland brings together deep technology veterans from Google, Nvidia, Apple, Waymo, Aurora, Embark, and Argo, as well as software engineers who have worked on mission-critical solutions at SpaceX, RTX, and the U.S. Army. Proceed to step two.
“The general idea is that almost every ground vehicle used by the military today has people on board,” Boots told TechCrunch in a video interview. “And you can imagine the safety and tactical advantages of being able to pull that person out of the vehicle.”
Detachment means that vehicles can navigate complex off-road terrain using only on-board sensors (mainly cameras, according to Boots), without relying on maps, GPS, or remote human operators. This means that they must move and calculate autonomously. This means Overland's software needs to understand the shape of the ground, including vegetation and mud, at every step, and how it affects vehicle dynamics.
“The terrain determines how the vehicle moves,” Boots says.
Overland's technology “basically takes sensor data and builds an evolving terrain representation,” Boots explained. The vehicle then uses that digital representation and the target it is trying to reach, which may be several kilometers away, to try to find a route through the terrain towards that target.
“One of the benefits of autonomous systems is that if the system is performing a task and loses its communications link with its ground vehicle, the system continues toward its goal until the communications link is reestablished. Just keep going and try to complete the task,” Boots said.
Currently, most road driving relies on communication links to remote assistance, in part because of the high risk to other road users. That's why you'll see Waymo and Cruise robotaxis bricked up on the streets of San Francisco, waiting to be prompted by remote operators after they've been shut down to meet minimum safety requirements. .
“Military ground systems often must function in unstructured and dynamic terrain. We believe that self-driving technology built for well-defined roads and enclosed sites is We believe it will be a struggle there, and we believe it will take a very strong team to provide operationally relevant ground autonomy in these environments,” said Chris, Partner in Point72 Ventures' Defense Technology Team. Morales said. he told TechCrunch.
Potential of off-road ADAS
Potential's technology, Terrain Intelligence, aims to improve ADAS for off-road use.Image credit: Possibility
“How can we actually support someone who might not be a 100% expert driver, but wants to go off-road and experience more challenging conditions?” Sam Poirier, CEO of Potential, recently asked asked in an interview.
Potential's core platform, called Terrain Intelligence, uses computer vision to enable vehicles to see, interpret, and prepare for the complex terrain and changing road conditions ahead. Terrain Intelligence can read data from a single camera rather than relying on additional sensors such as additional cameras, LIDAR, or radar.
At the most basic level, Potential's off-road ADAS alerts drivers that there are impassable objects ahead or that they should switch to better drive settings based on new terrain.
“The second level is whether you can actually automate the configuration changes that are typically driver-assisted,” Poirier said. “Most vehicles have two-wheel drive, four-wheel drive, sand mode, mud mode, etc. Ultimately, at this stage, it's up to the driver to switch between these, and the driver can decide when to use these different modes. you need to understand.”
The final level of Potential involves using existing sensor data and tweaking their settings to push the limits of performance.
Scott Kunselman, former Jeep chief engineer, auto industry veteran, and advisor to Potential, says, “No matter how much expertise an individual driver has, assistance tools can only do so much. I can't do it on my own.” “Stability control is a good example. To enable stability control, you need an independent brake control. The driver has only one brake pedal, which activates the entire brake system at once. , the stability control can actuate each wheel individually, allowing for the ability to offset the vehicle's yaw, for example.”
Yaw, by the way, is when a vehicle's weight shifts to the right or left from its center of gravity, which can cause a spinout or fishtail.
Potential said it is working with both Tier 1 suppliers and OEMs to license the software and integrate it directly into vehicles. Andreev has suggested that he may focus on business relationships with his Tier 1 suppliers rather than his OEMs, which is less likely to give smaller start-ups a chance.