By 2025 the automotive AI sector could be creating as much as $26.5 billion in annual revenue. That’s according to one prediction. The sector is clearly booming, particularly in Israel, which has become a hotbed of talent.
Ottopia is one Israeli start-up that has just closed $3 million in seed funding from investors. Unlike, some autonomous driving companies, it believes there’s still a need for humans in the self-driving process. A bit like air traffic control, but for cars.
“We tackle the most difficult 1% of AV performance with a remote human operator. A remote human is called whenever and wherever the AV isn’t sure what to do on the road. The trigger could also come from passengers in the AV,” Says Co-founder and CEO of Ottopia, Amit Rosenzweig.
Ottopia says it uses a unique combination of artificial intelligence and old fashioned human intelligence to work together to during remote intervention.
“The first elevator was installed more than 160 years ago and we still have emergency buttons in them if they get stuck or break. The same applies for parking lots where you occasionally press a button and talk to a human operator because your ticket didn’t get processed. The same actually applies for any ‘automatic’ system that humans interact with.” says Rosenzweig.
“Assembly lines aren’t really fully automatic, and so on and so forth. AVs are more complex than elevators by several orders of magnitude, thus, for the foreseeable future we’ll need a human backup,” he adds.
Rosenzweig spent 12 years in the Israeli Defence Forces and then commanded an R&D cyber-security section of 50 engineers in unit 8200 of the Israeli Intelligence Corps.
“We invest a lot in smart integration with existing AV stacks which include a great deal of AI. The secret sauce is knowing how to balance between commands that the remote human issues e.g. move forward now and the AV’s knowledge about its surroundings e.g. there’s a pedestrian right in front of the vehicle.”
Another AI start-up in Israel that’s laser focused on the auto sector is Tactile Mobility. Its mission is to help create a safer and enjoyable driving experience based on tactile sensing. It was founded by the veteran technologist Boaz Mizrachi, who’s Tactile Mobility’s CTO, back in 2012.
Boaz boasts over three decades of experience in signal processing, algorithm research and system design in the automotive and networking industries.
“The problem we started with was: get a given vehicle from point A to point B with minimal fuel consumption. For that we needed: surface information (e.g. road grades), vehicle information (e.g. optimal gear and RPM per load) and the ability to measure everything accurately (e.g. accurate momentary fuel consumption).”
His company, then called MobiWize, noticed that this information is missing but that it’s vital for the development of smart vehicles and building trust in autonomous cars.
According to the company, ninety-five percent of data currently used by smart and autonomous vehicles is visual. However, for them to mimic and even surpass human driving capabilities, these vehicles must, like human drivers, not only see the road but also ‘feel’ it.
Data is collected such as wheel speed, wheel angle, RPM, paddles position and gear position using a vehicle’s built in, non-visual sensors.
Physical modeling and then machine learning help create real-time virtual sensors that yield insights in real-time. These actionable insights, are subsequently fed back to the vehicle’s on-board computers, where it is used to make better driving decisions.
Tactile Mobility’s mapping service was launched in the city of Haifa. But the company has already collected over ten million kilometers of road data across four continents. The company is already working with a number of OEMs, including Ford.
According to Mizrachi, the AI-based tactile technology: “Will enable the AV to drive fast safely in all weather conditions. It will provide the missing sense for the virtual driver.”
“The sky’s the limit, literally,” concludes Mizrachi.
Ottopia’s Rosenzveig is similarly brimming with optimism and predicts that there may well be more autonomous robots than autonomous cars in the future. He also thinks that China will have the best infrastructure for AVs including V2X communication, dedicated lanes, or smart management at the city level.