At the all virtual Consumer Electronics Show (CES) in January 2021, Volvo Penta announced it was launching an assisted docking system that would give a captain better control when docking a boat to take into account for wind current.
In January 2021, two new electric car challenger companies Rivian and Cruise, announced a combined $4.65 billion in funding for their new electric cars, which would also be autonomous. Cruise started testing its driverless vehicles on the streets of San Francisco in December 2020.
In Gjesdal, Norway, a self-driving pilot from a Finnish self-driving software company, Sensible4, began field-testing cars in January 2021 using its software on the steep hills and harsh winter conditions of the region.
The company wants to push the boundaries of autonomous driving under challenging conditions. It uses Toyota vehicles to complete the 3.3 kilometers of wintery test-drives and map routes in the area. This pilot is a continuation of a pilot that began in Helsinki, Finland, in Spring 2020.
Harri Santamala, CEO and founder of Sensible 4, said the pilots in Norway raise the bar regarding autonomous cars' ability to operate in all weather conditions.
In addition to using Toyota cars, a small Renault Twizy is also being used in the pilot. The self-driving pilot is in collaboration with Spanish Shotl part of the European proof-of-concept research project, Future Automated Bus Urban Level Operation System last-mile public transport.
MORE FOR YOU
But, Geert-Jan van Nunen, Chief Commercial Officer of Teraki, said that the big challenge of autonomous driving is two-fold from a technical perspective.
“First we need to train computer models [..] to get to a high enough safety level that they can be used on the streets, there needs to be a lot of ‘training,’” said van Nunen. “For a computer model, this essentially is the same as with humans: give them so many examples – say pictures of a car - and tell them what it is, that they start recognizing a car when they see a picture of a car.”
Van Nunen says that second, all the processing of cameras, radars and lidars - the eyes and ears of a car - requires a lot of computing power.
“The first reaction is to ‘throw hardware at the problem’: making a car too expensive and emptying the battery too soon. Alternatively, one can select the important information from a scene and leave all other information out,” said van Nunen. “Do this at ‘the car’s fingertips’ - the sensors - and this is called “edge AI,” essentially the smarts in the car.”
Van Nunen believes that edge AI is a crucial area for successful autonomous driving. “Some companies have made impressive inroads to democratizing L4 driving,” added van Nunen.
Moving the industry forward
Van Nunen says the industry has some challenges to overcome to keep moving forward.
“For a car to be autonomous, it needs sensors. And it needs many of them. These sensors, like HD cameras, lidar, radars, produce an enormous amount of data. With large incoming streams of data, the car has to analyze and compute a lot,” said van Nunen. “We can’t stuff cars with big, heavy computers, and we can’t compromise on safety, and that’s why the industry is looking to artificial intelligence (AI) to solve this problem.”
Van Nunen says that means software that helps distinguish very early in the chain of what image is relevant or not relevant. “The sky is not relevant; a pedestrian is. Both AI in the cloud and AI in the car via the edge, are what I expect to bring that solution.
The Link LonkJanuary 30, 2021 at 02:18AM
https://www.forbes.com/sites/jenniferhicks/2021/01/29/forget-the-city-see-how-this-self-driving-pilot-in-norway-tackles-winter-conditions-and-steep-hills/
Forget The City, See How This Self-Driving Pilot In Norway Tackles Winter Conditions And Steep Hills - Forbes
https://news.google.com/search?q=forget&hl=en-US&gl=US&ceid=US:en
No comments:
Post a Comment