Startups with self-drive vision

Article By : Junko Yoshida

It’s a race now for autonomous vehicles with car OEMS and tech startups joining forces. Here are some startups we discovered.

« Previously: With Ford-Argo AI tieup, R&D experts race in
 

Here's a list of autonomous driving startups, in alphabetical order. The listed business names focus on their "autonomous vehicle" platform instead of specific key components such as lidars, vision or connectivity.

AImotive (Budapest, Hungary), formerly AdasWorks, develops AI-based software for autonomous cars. The company’s latest development includes a power-efficient AI accelerator chip – compliant to work with the spec currently in development at the Khronos Group. (The Khronos Group has been working on an extension to enable Convolutional Neural Network topologies to be represented as OpneVX graphs mixed with traditional vision functions.)

 
Embedded content: https://youtu.be/Nj1UQ-L-Ux0
Source: Embedded Vision Alliance
 

AImotive is a spin-off of Kishonti, a known player in the benchmarking business offering optimization services for the mobile market. Based on its heritage, AImotive knows quite a bit about generating realistic graphics. The company offers a real-time driving simulator – depicting all weather, traffic and lighting conditions – which helps automakers simulate and test self-driving vehicle software.

AImotive works closely with automotive and technology companies (OEMs, chip suppliers, platform providers) to develop the architecture of cars, and devise reliable self-driving technologies.

Founded in July, 2015, the company raised $9.45 million in the two rounds of funding.

AutonomouStuff (Morton, Illinois) provides autonomy-enabling technologies including perception sensors, GPS, and computing.

 
Embedded content: https://youtu.be/qct5bqfu09A
AutonomouStuff has developed automated driving platforms that integrate Nvidia’s DRIVE PX 2. The video was uploaded in Jan., 2017 (Source: Nvidia)
 

The company, founded in 2010, offers platforms, software and services for ADAS and autonomous driving. AutonomouStuff touts its “completely customizable R&D Vehicle Platforms used for advanced driver assistance systems (ADAS), advanced algorithm development and automated driving initiatives.”

Comma.ai (San Francisco), a developer of an aftermarket ADAS unit, surprised the industry last fall by cancelling its first product launch. But the startup doubly shocked the industry late November by deciding to give away the code for his self-driving car project.

 

Figure 4: Source: Comma.ai
 

The startup is now providing the source code for OpenPilot to all comers.

George Hotz, Comma.ai’s founder, who calls his new project the “open-source alternative to Autopilot,” is distributing the Comma.ai code through GitHub.

Phil Magney, founder and principal advisor for Vision Systems Intelligence, explained to us, “Some parts of the [Comma.ai] code cannot be modified, such as Comma’s trained network – a.k.a. inference model.” Other parts are open, like HMI and the control commands, he added. “This platform is for development only. If someone wanted to commercialize, they would need to negotiate. OpenPilot is really the behavior model based on Comma.ai’s trained network.”

In seed funding, Comma.ai received $3.1 million last year. Its investors include Andreesen Horowitz and Techammer.

Drive.ai (Mountain View, Claif.) develops AI software for autonomous vehicles. The goal is to build a hardware and software kit powered by artificial intelligence for carmakers.

Founded in 2015 by an engineering team out of Stanford University’s Artificial Intelligence Lab, the startup says it distinguishes itself through its expertise in robotics and deep learning, and applying machine learning to both driving and human interaction.

 
Embedded content: https://youtu.be/GMvgtPN2IBU
On Feb. 14, 2017, Drive.ai uploaded a video clip of a fully autonomous drive on a rainy night on the streets of Mountain View, CA. (Source: Drive.ai)
 

Drive.ai has raised $12 million thus far. The company’s investors include InnoSpring Seed Fund, Northern Light Venture Capital, and OrizaVengures.

 

Figure 5: On Feb 3, 2017, FiveAI reported that it has begun track testing its first autonomous vehicle prototype. (Photo courtesy of FiveAI)
 

FiveAI (Bristol, the U.K.) aims to build “the world’s most reliable autonomous vehicle software stack.” The startup is developing a host of technologies including sensor fusion, computer vision combined with a deep neural network, policy-based behavioral modeling and motion planning. The goal is to design solutions for a Level 5 autonomous vehicle that’s safe in complex urban environments.

FiveAI was founded last year by Stan Boland. Boland is a serial entrepreneur who co-founded and served as CEO for Element 14 (sold to Broadcom) and Icera Inc. (bought by Nvidia). The 15-member team listed on its web site, includes a number of ex-Icera executives in addition to computer scientists, mathematicians, physicists, PhDs in machine learning and computer vision.

The company raised $2.7 million in seed funding from Amadeus Capital Partners, Kindred and Notion Capital.

Nauto (Palo Alto, Calif.) is working on an autonomous vehicle technology system. Started as an after-market ADAS company, it offers an AI-powered dual camera unit that learns from drivers and the road.

Using real-time sensors and visual data, Nauto focuses on insight, guidance and real-time feedback for its clients, helping fleet managers detect and understand causes of accidents and reduce false liability claims. The system also helps cities with better traffic control and street design to eliminate fatal accidents.

 

Figure 6: Combining real time data with historic trends, Nauto provides fleet managers with the complete picture — ranging from the fleet's performance, to the driver's distraction, and environmental conditions ahead. (Source: Nauto)
 

Founded in March, 2015, Nauto says it seeks to understand what causes near-misses, scrapes, and accidents at every level.

Last year, Nauto said that the company is using Qualcomm’s Snapdragon neural processing engine SDK to run its proprietary deep -learning algorithms.

The company has raised $14.85 million over three rounds of funding. Investors include BMW i Ventures, Draper Nexus Ventures, Index Ventures, Playground Global and Toyota Research Institute.

Oxbotica (Oxford, the U.K.), is a self-driving car company and a spin-out from Oxford University. It specializes in mobile autonomy, navigation, perception, and research into autonomous robotics. As VSI’s Phil Magney noted, Oxford’s Robotics Institute, which has been engaged in the development of self-driving features for some time, “purports to train their algorithms using AI for complex urban environments by directly mapping sensor data against actual driving behavior.”

Thus far, it has developed its own autonomous control system, called Selenium, a vehicle agnostic operating system applicable to anything from forklifts and cargo pods to vehicles. Oxbotica also offers a cloud-based fleet management system that will schedule and coordinate a fleet of autonomous vehicles, enabling smartphone booking, route optimization and data exchange between vehicles without human intervention.

The startup also developed Geni, a development platform that allows prototyping and testing of new algorithms.

Oxbotica was founded in 2014 by Professors Ingmar Posner and Paul Newman, leaders of Oxford University’s Mobile Robotics Group.

 
Embedded content: https://youtu.be/38NIyk4JN-g
The demonstration of Selenium — Oxbotica’s autonomous control system – embedded inside the Geni, Oxbotica’s development platform. Video was uploaded in Nov., 2016. (Source: Oxbotica)
 

Preferred Networks Inc. (Tokyo) works in the fields of natural language processing and machine learning. The startup is focused on “Edge-Heavy Computing” (where devices analyze data locally at the edge of the network) with a goal to apply real-time machine learning technologies to new IoT applications including self-driving cars.

PFN began exclusive collaboration with Toyota Motors in Oct. 2014 to develop self-driving technology using Deep Learning. Toyota subsequently invested 1 billion yen in Dec., 2015.

The company, thus far, has raised $17.3 million. Investors include, Fanuc, NTT and Toyota.

 
Embedded content: https://youtu.be/cldaGudHP3I
Meet Nobuyuki Ota, COO of Preferred Networks, talks about his company's Deep Intelligence for Networks technology which enables machines to learn control policies without human intervention. Ota also talks about his experience in the Cisco Entrepreneurs in Residence program. (Source: Cisco)
 

Zoox (Menlo Park, Calif.) describes itself as “a robotics company pioneering autonomous mobility as-a-service.”

The startup, founded in July, 2014, is developing a fully automated, electric vehicle fleet and the supporting ecosystem required to bring the service to market.

 

Figure 7: Zoox aims to offer a luxury autonomous driving service.
 

Once called a “stealth Uber competitor,” Zoox has raised $290 million. IHS’s Juliussen believes that Zoox is one of the best-funded autonomous driving startups.

Before unveiling its first product, the company’s valuation soared to $1.5 billion last fall. It was cofounded by Australian designer Tim Kentley-Klay and Jesse Levinson, an engineer who worked at Stanford University with Sebastian Thrun, the first director of Google’s self-driving car programme.

 
« First page: Ford eyes self-driving in Argo AI stake

First published on EE Times.

Subscribe to Newsletter