Amazon's Zoox self-driving taxi software recalls: Vehicles cannot recognize thick smoke and drive into smoke areas
Zoox, its self-driving company, voluntarily recalled 105 self-driving taxis to fix a software defect - the vehicles were unable to recognize a large amount of thick smoke and would drive straight into the smoke area. The company revealed that an unloaded Zoox autonomous vehicle drove into a smoke-filled fire emergency scene in Las Vegas last month. Prior to this recall, the U.S. National Highway Traffic Safety Administration had required autonomous driving companies to address safety hazards related to vehicles interfering with emergency rescue personnel. A Zoox self-driving taxi rides on Wednesday, Dec. 4, 2024, in San Francisco, California, U.S. Zoox, a wholly-owned subsidiary of Amazon, has started testing electric driverless taxis in the SoMa area of San Francisco. Zoox, a subsidiary of Amazon, announced a software recall for 105 self-driving taxis. The hidden danger is that the vehicle sensing system cannot recognize thick smoke and will drive into areas covered by thick smoke. Zoox filed a recall filing with the National Highway Traffic Safety Administration (NHTSA) on July 8, saying an accident last month led to its discovery of a flaw in the smoke detection system. The company recorded in the report: On June 20, a fire scene in Las Vegas was filled with thick smoke. No isolation cones were placed at the scene. An unloaded Zoox unmanned vehicle drove directly into the area. After detecting the abnormality, the vehicle braked urgently, tried to steer and avoid, and finally came to a complete stop. After the accident, Zoox remote control staff issued a reversing command, and then on-site rescuers placed traffic cones to isolate the accident area. Zoox stated that it has completed a comprehensive investigation of the accident and confirmed that this was the only accident of this type of hidden danger. No one was injured in this incident. Amazon spent $1.3 billion to acquire Zoox in 2020. The company's self-developed autonomous scooter has no steering wheel or accelerator and brake pedals. The four sets of seats in the car face each other, and the layout is similar to that of a shuttle bus. Zoox currently provides free self-driving ride-hailing services in some areas of Las Vegas and San Francisco; it has opened ride-hailing reservations in limited areas in Miami and Austin, Texas, and is simultaneously conducting tests in six other cities in the United States. Prior to the issuance of this voluntary recall, the Director of the National Highway Traffic Safety Administration, Jonathan Morrison, issued a directive to all autonomous driving companies last week, requiring unmanned vehicles to actively avoid emergency rescue personnel. Morrison stated in the official letter that regulatory agencies have discovered obvious common problems: multiple self-driving vehicles frequently interfere with law enforcement and emergency rescue work. Typical accidents include driving into emergency response sites, blocking ambulance and fire truck routes, and being unable to recognize emergency signs such as warning lights, signal torches, thick smoke, open flames, and isolation cones. He asked all autonomous driving R&D and operating companies to "immediately allocate resources to rectify such safety defects" and submit rectification plans to the regulatory bureau before the end of the month. The official letter did not name specific car companies. Last year, Zoox launched several software recalls to fix defects in the algorithm for vehicles traveling across lanes and predicting the trajectories of motor vehicles and pedestrians. Zoox is struggling to catch up with Alphabet's Waymo. Waymo is the self-driving rental service provider with the largest market share in the United States, with a self-driving fleet size of approximately 4,000 units in the United States. Last month, Waymo just completed the recall of about 3,900 self-driving taxis. The hidden danger is that some vehicles drive into high-speed closed construction sections, which greatly increases the risk of collision accidents.