Middle East

Tesla’s camera-only autonomous system causes controversy

Vehicles equipped with the Luminal system

Las Vegas: As Tesla pursues the goal of fully autonomous driving, it is fully betting on cameras and artificial intelligence, avoiding other commonly used tools such as laser detection. According to Tesla’s website, Tesla CEO Elon Musk touted a system built around eight “surround” cameras that feed the car’s “deep neural network.” ..

But like much else related to Tesla, there is controversy. At the gigantic Consumer Electronic Show (CES) in Las Vegas, Luminar Technologies set up a demonstration of two cars moving towards a child’s silhouette at 30 mph.

Cars powered by Luminar’s LIDAR, a laser-based system, will stop before trouble and rival Tesla will take care of the mannequin. Lumar’s experiments, which have not been validated by outside experts, were aimed at “side-by-side comparisons” that allowed people to “actually evaluate differences,” Luminar’s Aaron Jefferson said. Told to.

“In perfect driving conditions, on the right sunny days, the camera can do a lot,” Jefferson said. “The problem lies in these corner cases.” In this case, objects such as fog, sunset lights, or drifting plastic bags can obstruct your view.

Many companies working in autonomous transportation spaces have agreed with Luminar and are focusing on technologies that combine cameras with other systems such as lidar, radio-based radar, or both.

However, Tesla, a car that currently offers driver assistance systems that are less than fully autonomous, takes a different approach under the iconic musk. The masking method is to obtain real-time readings from a camera supplied to an artificial intelligence system built around data collected over many years by Tesla sensors.

Kirian Weinberger, a computer scientist at Cornell University, said: Weinberger, who called Tesla’s approach a “pretty rational strategy,” said in Musk’s view that his system “is as much as an algorithm that actually uses all of these expensive sensors with much less data. It ’s excellent. ”

Technology is “stagnation?”

Automakers are adding systems to assist with braking, parking, and other functions, but full autonomy has proven difficult. According to Sam Abuelsamid of market intelligence company Guidehouse Insights, autonomous driving systems are built around four functions. It recognizes the environment in real time, predicts what will happen next, and plans and executes the next movement of the car.

“We found that predictions were much more difficult than everyone expected, especially for pedestrians and cyclists,” said Abuelsamid, who said that advances in software built around cameras are “slightly stagnant.” I added.

“Elon Musk has sold the story of fully automated driving based on the hardware they had,” he said. “They have already sold the system to hundreds of thousands of customers, so they pushed themselves to the corner.” To other skeptics of Tesla’s approach, France announced the latest LIDAR technology at CES. Includes Jacques Aschenbroich, Chief Executive Officer of Valeo, an automobile supplier of.

“Camera alone isn’t enough, even with large amounts of data,” said Aschenbroich, who places particular emphasis on sensors in urban areas and other areas where car traffic and pedestrians are unpredictable.

“We firmly believe that LIDAR is essential to reach a higher level of autonomy,” he said.

“Every sensor has its strengths and weaknesses,” said Marko Bertogna, a professor at the University of Modena, who is part of the team participating in the CES self-driving car race.

Bertogna believes that regulators are likely to embrace an autonomous system built around redundancy.

“With today’s state-of-the-art technology, systems that rely solely on cameras are still too problematic to safely verify,” he said. -AFP

https://www.kuwaittimes.com/teslas-cameras-only-autonomous-system-stirs-controversy/ Tesla’s camera-only autonomous system causes controversy

Back to top button