Chongqing- A test vehicle equipped with the world's first human eye bionic visual perception technology was showcased at Chongqing Liangjiang Collaborative Innovation Zone on November 7. Its vision algorithm through a camera can simulate human eyes, making autonomous vehicles sense around just like a human does.
At the testing site, a white truck parked on the spacious urban expressway. At the back of the truck, two test vehicles rush over at 100 kilometers per hour.
One was equipped with the world's first human eye bionic visual perception technology developed by Chongqing Toramon Technology Co., Ltd., which sent out an alarm of collision 100 meters from the truck. Another vehicle was equipped with the world's most advanced driving assistance system but was not responding until a few meters before the collision.
The test shows that Toramon, an unknown enterprise in Southwest China's Chongqing municipality, has mastered the industry-leading visual perception technology.
Toramon is a provider of active safety technology for intelligent driving. The world's first human eye bionic visual perception technology enables cars to detect collision hazards based on motion perception results.
When cars are not trained by AI or encounter unidentified obstacles, this technology improves safety while reducing the cost of advanced driving assistance systems. In July 2021, the company obtained a Pre-A investment valued at tens of millions of yuan.
According to the report, by 2025, the National Advanced Driving Assisted System (ADAS) market size will exceed 225 billion yuan (about 31.8 billion U.S. dollars). As a vital automobile production base in China, Chongqing has an opportunity to take the lead on ADAS in the new track by enterprises like Toramon.
"Statistics show that vehicle perception barriers to the environment cause the majority of automated driving accidents," said Liu Yan, CEO of Toramon Technology.
Liu continued that the working principle of the automated driving system is perception, decision-making, and control. The perception of the driving environment is the most critical part.
To accurately perceive the driving environment, Tesla and Mobileye use cameras to collect information, and AI learning enables them to identify more obstacles. However, AI cannot cover all obstacles. Once AI cannot identify obstacles, the system fails.
Other enterprises use cameras plus radar to collect information. Although it can perceive all obstacles, it also results in vehicles being too sensitive to get false alarms. Dozens of radars also drive up costs.
Tramon's solution comes from wearable devices that do not need radar. Its perception technology imitates the perception mechanism of human eyes. Without seeing the object, the part of the brain responsible for motion and spatial perception can subconsciously avoid obstacles according to the motion information.
Up to now, the company has completed 1.5 million kilometers of road testing without any traffic accidents, and the accuracy of collision warnings exceeds 99%.
Professionals consider it the beginning of a transformation of the visual perception technology path. This September, Tesla proposed the Occupancy Networks scheme for automated driving visual perception, similar to the perception mechanism of human eyes.
"I have studied and worked in Chongqing for more than ten years, and I know that Chongqing is an ideal testing place for automated driving," said Liu. "The complexity of Chongqing's traffic scene is unique in China. As long as the technology works in Chongqing, any other city will be a piece of cake."
Many automated driving teams from all over the country have set their eyes on Chongqing. The annual i-VISTA Smart Connected Car Challenge has become a battlefield for self-driving vehicles.
The Chongqing automobile industry has recently accelerated its transformation into an intelligent network.
In 2021, the fourth national internet of vehicles pilot area was in Liangjiang New Area. In August 2022, Chongqing released the pilot policy for unmanned automated driving.
(Tan Jingyue, as an intern, also contributed to this report.)
Internet illegal and undesirable information can be reported by calling this telephone number：+86-23-67230470