We all hear about Uber, Google and others all working hard to get autonomous cars out on the road. As you can imagine, the Big 3 are also attempting to get something working – quicker than the smaller players.
Last year General Motors bought Cruise Automation based out of San Francisco to ensure that they were at the forefront of this new technology. I have written about about LIDAR, one of the components, and also Uber’s plans for a self driving fleet of trucks and I think that with the miniaturisation of the components, this part of the auto industry will gain strength and become more critical to urban areas.
Cruise Automation was founded in 2013 to push the developments of all the components of an autonomous vehicle. Their first product was the RP-1 that was fitted to an Audi A4. This system was a rooftop “pod” filled with electronics and a computer pack that was fitted in the boot (trunk) of the vehicle. Override controls were provided to the driver as well. California State had given the company a permit to test the equipment however they weren’t allowed to use the kit for normal usage. The pod looked quite space-age, however the founder, Kyle Vogt, felt that more could be done. Initially they wanted to sell these kits at US$10,000 a time and after selling all the initial orders it was decided that more really could be done – i.e. a full autonomous car.
The basis was to be a Nissan Leaf and the challenge was to get the vehicle to “understand” a city scape. The other more important challenge was to work with State and Federal lawmakers to allow the vehicles to be used alongside human drivers. Cash was critical and selling kits with no legal approval probably wasn’t the best way to raise the necessary capital and along came General Motors – a big company with big cash. The Nissan was dropped in favour of a Chevrolet Bolt.
The deal was sealed in 2016, however there was a hiccup: one of their engineers demanded a stake in the company (clearly he wanted a big pay day) and so Cruise Automation had to launch a legal swipe to deal with him. Estimates claim that the buy price was US$1B! This gave the startup a big injection of cash and a facility to build cars to their requirements without having to buy a car and modify it. That is key to the success of this project – it provides all the equipment built in rather than added in later which should reduce the number of errors because the kit can be plumbed in properly to the wiring looms etc.
GM supplied 30 test cars for real world testing around San Francisco, Scottsdale and Detroit (perhaps even around MCity) and they have announced that a further 130 cars will be built for expanded testing. GM have also stated that they are keeping Cruise Automation as a separate division – an autonomous autonomous car division!
Each vehicle has 10 cameras taking a constant 10 frames per second in a 360 degree view. The front has a radar unit and there are LIDAR sensors around the car as well. To connect the vehicle to the outside world and other data sources, there is a 4G LTE system enabling the communications link.
A couple of months ago, Cruise Automation quietly bought Strobe Inc, a source of LIDAR technology. Strobe has done what others are also trying to do, which is to shrink the LIDAR technology down on to a single chip. This means that a vehicle can have many more sensors on them to build a more accurate picture of its surroundings. The advantage for General Motors (and Cruise Automation), is that with the shrinking of the equipment, the core vehicle designers can now build it straight into the vehicle’s structure rather than bolting it on.
GM has a partnership with the ride-sharing system from Lyft – that means we will see the battle intensify between Lyft and Uber with autonomous cars on urban roads. Uber is under attack from many sides, so GM may have hitched its wagon to the right (self driving) engine!