Topics
Latest
AI
Amazon
Image Credits:Overland AI
Apps
Biotech & Health
clime
Image Credits:Overland AI
Cloud Computing
Commerce
Crypto
Overland AI’s off-road autonomous driving software, OverDrive, is being tested for defense and national security applications.Image Credits:Overland AI
endeavour
EVs
Fintech
Potential’s technology, Terrain Intelligence, aims to improve ADAS for off-roading.Image Credits:Potential
Fundraising
appliance
bet on
Government & Policy
Hardware
layoff
Media & Entertainment
Meta
Microsoft
privateness
Robotics
Security
societal
Space
Startups
TikTok
Transportation
Venture
More from TechCrunch
event
Startup Battlefield
StrictlyVC
Podcasts
Videos
Partner Content
TechCrunch Brand Studio
Crunchboard
get through Us
A new crop of other - stage inauguration — along with some late VC investment — illustrate a niche emerging in the autonomous fomite technology sector . Unlike the troupe bestow robotaxis to metropolis street , these startups are taking their technical school off - road .
Two recent entrant — Seattle - basedOverland AIand New Brunswick , Canada - basedPotential — are poised to get a first - removal company advantage on this segment of self-reliance .
While these startups are applying their tech in different way , Overland AI and Potential do apportion some rough-cut off - road ground . Founders of each startup consider they ’ve cracked the codification to one of the more intriguing software of automate driving by building computer software that does n’t rely on some of the main crutch of examination and deployment — such as detailed map , large swaths of training information and the ability to precipitate back on distant help .
The U.S. Department of Defense and speculation capital investor are taking notice .
Overland AI , which is uprise a ego - driving system designed for military operations like reconnaissance , surveillance and delivering electronic war package , was awarded in April up to$18.6 millionfrom the U.S. Army ’s Defense Innovation Unit . The fund will be used to build a image autonomous software slew for its robotic combat fomite ( RCV ) program over the next two years .
This workweek , the inauguration , which was institute in 2022 , raised a $ 10 million seed round led by Point72 Ventures . The funds will be used to expand Overland ’s team and continue developing OverDrive , the company ’s autonomy stack , according to chief executive officer and founder Byron Boots .
Meanwhile , Potential , which is making advanced number one wood - assistance systems ( ADAS ) that let ATVs , underground mining vehicles and passenger cars to cover off - road environments , has raised a CA$2 million ( ~$1.5 million ) extension to its seed cycle run by Brightspark Ventures , a Canadian former - stage VC . That work Potential ’s entire funding to CA$8.5 million ( ~$6.2 million ) . The startup has spent the last six years educate its technology and is now doing several pilot projects across world power sport , motorcycles and self-propelled .
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
Off-road opportunity
Potential and Overland AI are n’t the only troupe examine to apply autonomous vehicle engineering to domain outside of public street . The high - cost pursuit of commercial robotaxi and ego - driving truck military operation has thwarted dozens of startup over the retiring several years . As those keep out down , a raw batch of startup such as Polymath Robotics , Pronto.ai , Bear Robotics and Outrider have emerged ( along with more established caller like Forterra ) with more grounded ambitions : applying AV tech to storage warehouse , mining , industrial and off - road surround .
“ We are perfectly deploying capital in off - road autonomy , ” Alexei Andreev , managing director at Autotech Ventures , told TechCrunch . “ Actually , if anything , we are staying off from highway autonomy and completely doubled down on off - route autonomy . ”
Most of the off - route companies that Autotech Ventures is investing in today are in the agricultural and structure sectors — ware like autonomous minelaying vehicles , forklifts and tractors . Andreev says for these sphere , it ’s about addressing the labor famine while increasing productivity and making farm and mental synthesis areas safer .
“ And if you slay people , you immediately get a reduction in your indemnity premium . So the return on invested capital for those vertical covering is now and it ’s significant , ” aver Andreev .
Another upshot : Off - road autonomy has found a protagonist in defence mechanism .
Overland AI: From DARPA to seed funding
When it fare to automating off - route drive , the U.S. Army can be a great client . After all , autonomous vehicle started as a DARPA project , says Jeff Peters , a partner at Ibex Investors . DARPA ( Defense Advanced Research Projects Agency ) is a U.S. Department of Defense agency focused on advancing engineering science for military use .
“ The hoopla around AV moved a lot of the industry toward gravid possible commercial-grade applications , but DoD projects have persist , ” Peters told TechCrunch over electronic mail , note that autonomous excavation startup SafeAI and autonomous trucking startupKodiak Roboticshave also pursued defense grants . “ I think AV companionship ( those still around ) will tail DoD project because it offers large , non - dilutive financial backing in the meantime prior to commercial operations . ”
Overland AI is the latest byproduct of the DARPA program . Boots , a prof of motorcar learning at the University of Washington and founder of the Robot Learning Laboratory at the university ’s schoolhouse of computer science and engineering , has a prospicient account of collaborating with the U.S. Army Research Lab and DARPA .
Overland was spin out from Boots ’ research and squad involve in DARPA ’s RACER ( Robotic Autonomy in Complex Environments with Resiliency ) program , the finish of which is to develop ego - driving vehicles that can wield tough terrain .
The programme is still on-going . Overland , which is stacked with deep tech veterans from Google , Nvidia , Apple , Waymo , Aurora , Embark and Argo , as well as software package engineers who ’ve worked on mission - critical solutions at SpaceX , RTX and the U.S. Army , was lately selected to continue on to the 2d stage .
“ The high - tier melodic theme is that currently just about every ground fomite that the war machine use has a someone inwardly of it , ” Boots tell TechCrunch in a video audience . “ And you may imagine if you may just draw the person out of the vehicle , that confers safety and tactical advantages . ”
To perpetrate the person out means vehicle must autonomously navigate complex off - road terrain using only onboard sensors ( chiefly cameras , agree to boot ) and figure , without swear on maps , GPS or distant human operators . That means Overland ’s software has to realise the geometry of the background — include thing like flora and mud — every step of the agency , and how that affects vehicle dynamics .
“ The terrain gets a vote on how the vehicle moves , ” tell charge .
Overland ’s tech “ basically take in the detector data and builds a terrain representation as it goes , ” boot explained . Then the vehicle uses that digital representation “ plus the destination that it ’s trying to get to , which could be several km away , to endeavor to retrieve a route through the terrain towards that finish . ”
“ Part of the benefit of have got an autonomous scheme is that when the system is tasked , if you recede a communicating link to that terra firma fomite , it will continue to move towards its goal and hear to complete the task until the communication connection is reestablished , ” said Boots .
Most on - road drive today bank on that telecommunications link to distant aid , in part because the danger to other road users is high . That ’s why you ’ll seeWaymoand Cruise robotaxis bricked up on the streets of San Francisco , waitress for a remote hustler to give them a nudge after they break drive to play a minimum safety requirement .
“ Military ground system often ask to function in unstructured , dynamical terrain . We believe self - driving technology progress for well - delimitate street and enclosed lots will skin there , and that it takes a very strong squad to deliver operationally relevant primer coat liberty in these surround , ” Chris Morales , collaborator on the Department of Defense tech team at Point72 Ventures , told TechCrunch .
Potential’s potential with off-road ADAS
“ How can you really enable somebody who perchance is n’t the 100 % expert driver , but somebody who wants to go off - roading and experience these more thought-provoking consideration ? ” Sam Poirier , CEO at Potential , asked in a recent consultation .
Potential ’s core platform , called Terrain Intelligence , uses computer vision to help vehicles see , understand and prepare for complex terrain and changing Earth’s surface conditions in advance . Terrain Intelligence can translate data point from a single photographic camera , rather than relying on extra sensors like extra cameras , lidar and radar .
At the most basic grade , Potential ’s off - route ADAS alerts the equipment driver to an unpassable object up ahead or the motivation to switch to a better drive setting found on new terrain .
“ The 2d level is , can we instead actually avail to automate the changes of what are typically number one wood - assisted setting ? ” said Poirier . “ Most vehicle have two - wheel drive , four - wheel drive , sand mode , clay mode , things like that . at long last , at this stage , it ’s up to the number one wood to shift between these … and the number one wood has to realize when to practice these different modes . ”
Potential ’s last grade would involve using exist sensor data and fine - tune up those mise en scene and advertize the demarcation of functioning .
“ There are things that the assistance tools can do that an individual driver — no matter how adept your expertise — can not do on their own , ” said Scott Kunselman , an ex - Jeep chief locomotive engineer , auto industry vet and adviser to Potential . “ Stability controls are a good example because to enable stableness control , you want main pasture brake mastery . The driver only has one brake pedal and actuates the whole pasture brake system at once . Whereas constancy control can individually trigger each wheel and that ’s how you’re able to grow , for example , the ability to offset yaw in a vehicle . ”
Yaw , by the way , is when a vehicle ’s weight unit geological fault from its centre of solemnity to the right hand or left , which can do it to spin out or fishtail .
Potential said it ’s working with both Tier 1 suppliers and OEMs to license its software and desegregate it directly into the vehicles . Andreev suggests likely focus on business dealings with level 1 suppliers rather than OEMs that are less likely to take a chance on a pocket-size inauguration .
