Your browser is not supported. Please update it.

8 March 2023

Ford rejoins autonomous driving race as competition heats up

This week Ford established Latitude AI, a wholly owned subsidiary to develop what it calls a “hand-free, eyes off driver assist system” for its next-generation vehicles.

Latitude AI will be a team of 550 people hired from Argo AI – a now defunct company that was majority owned by Ford and the Volkswagen Group at 42% each, alongside Lyft and other smaller investors. The team will be working with data from Ford’s BlueCruise system, which has accumulated more than 50 million miles of driving data to work from. Ford’s executive director of advanced driver assistance technologies (ADAS) – Sammy Omari – will serve as the CEO of Latitude.

Ford is quite late to the game in establishing its own dedicated autonomous driving unit, following on from a litany of incumbents, startups, and tech companies looking for a lateral move into the automotive industry. We’re going to go through a few of them and outline how their approaches differ, but first it’s important to understand what the different levels of driver assistance mean.

ADAS has 6 levels from 0 to 5, where level 0 means a vehicle has no automation whatsoever, and level 5 is completely automated driving with no need for driver input, or even a steering wheel for that matter. For context, Tesla has admitted that its “full self-driving” system – despite the name – is at level 2 which is effectively smart cruise control that still needs constant supervision to not crash into anything.

During Tesla’s investor day on the 1st of March it said that the company will be producing a dedicated robotaxi model on its new platform so it can operate a ridesharing service like Uber without having issues like pesky labor laws. The company has been promising “full self-driving” for years and years now, and while it is edging closer to not embellishing its system’s capabilities it still has a long way to go.

General Motors is pursuing the fully vertically integrated model, developing the necessary software itself for its own vehicles rather than working with specialists that may or may not be more suited towards software development. GM is doing this through its wholly owned subsidiary Cruise, which has already suffered one recall of 80 vehicles due to a crash last June that left 2 people injured. GM has shown that it doesn’t want to rely on other companies to develop or supply its own technology where possible, the company’s expansion into semiconductor manufacturing and raw material supply ownership make great examples of vertical integration.

Companies dedicated to developing their own autonomous driving systems without the vehicle have had a rough year. Alphabet’s Waymo has undergone 2 rounds of layoffs totaling 209 jobs lost this year, Aurora has supposedly been considering acquisition, Nuro has laid off 20% of its workforce.

It isn’t all doom and gloom though; Intel recently acquired Mobileye for $15 billion and has announced that it will be launching a full-scale driverless delivery service this year in partnership with Udelv. Mobileye is targeting a consumer cost of $5,000 for its Chauffeur system when it launches in 2026.

Motional and Lyft plan to launch a robotaxi service in Las Vegas this year using the Hyundai IONIQ 5 as their vehicle of choice while Amazon’s Zoox has recently started offering similar driverless taxi services in California.

Software giant Nvidia is also working on its own self-driving system alongside Bosch, the company already has agreements in place with Mercedes-Benz and Audi for future integration for both passenger and freight operations. The biggest competitor from China currently seems to be Baidu, the internet search giant, which has already launched a driverless taxi service in Wuhan.

The end goal to all this research and development will be adaptive full self-driving capabilities that work outside of specific metropolitan areas. A vast majority of pilot projects ongoing at the moment are in areas that have been extensively mapped out in advance and are quite limited in scope, with some limited to just a few miles in any direction. As more miles are accumulated through pilot projects and simulation-based environments full self-driving capabilities will increase significantly, but as we’ve seen in the past with Tesla’s commitments, true level 5 autonomous driving capabilities is an immense technological feat with no end of hurdles in the way.

There’s also the question of regulation and liability and how that will be addressed.