Waymo experiences artificial intelligence and other technologies for its self-driving cars, but the company believes that a variety of laser sensors and radars installed on its cars remains the safest way to operate a robotics service on a large scale-at least at the present time.
“We have conducted a lot of research. We are familiar with what it succeeds, what does not work widely and what we need to do,” said Thirumalai, Vice President of Engineering, on the current plane, Waymo, this week at the AI4 Conference in Las Vegas.
While competitors like Timing Thirumalai pays from Waymo, which pays self-driving cars that depend only on video cameras, the Lidar and Radar mix provides an “additional safety network” to ensure that the company has enough data it needs to make driving decisions “under all circumstances”-including the harsh weather.
Thirumalai was talking on the stage in an interview with luck. Earlier that day, Thirumalai made a single show, describing the AI of Waymo and its safety approach in detail, which allowed the company to expand its operation to five cities by mid -2015 and conduct more than 100 million miles without a driver. In its presentation, Thirumalai’s view is a video clip of how Lidar sensors on Waymo Jaguar I-Pace took a movement of people who seek to jump on the way, even when the car cameras-or a woman were not preparing to roam in a bus that stopped directly on the Robotoki Waymo road. Either way, Robotaxi stopped from Waymo or maneuvering outside the road to avoid infantry contact, according to videos.
The presentation showed the blatant contrast in the methods between Waymo and one of its new competitors, Tesla, which launched only a small roboti service in Austin, with safety drivers in the passenger seat. Tesla, which showed its full technology (FSD) by experimental ride at the AI4 conference, uses only video and artificial intelligence technology for FSD and Tesla Robotaxi, after years of Elon Musk stating that other expensive and unnecessary sensors. Elon Musk said in 2019. “Anyone who relies on Lidar is convicted. He is convicted! (They are) unnecessary expensive sensors.”
Thirumalai will not say directly whether it is considered self -driving systems like Tesla to be safe for public roads. He said that you should think about the “whole process” of how the system is built and tested and then checks its health, and he also said that you cannot compare the Waymo system with another, because there are no similar safety measures. The sub -trip of General Motors, which also used Lidar and Radar, was suspended earlier this year after failing to re -launch after a serious accident in San Francisco. For context, Tesla said it paid 7,000 miles without a driver in The end of JulyCompared to 100 million.
“If we are talking about objective measures, we must look at our safety record statistics, on a large scale, right?” Theromalay said. “When someone actually says: Yes, we have match your safety on your scale with a different system, this is great. We will take that.”
Waymo regularly tests new technology when it is available, according to Thirumalai. As part of this experiment, he said that Waymo has searched how to combine multimedia models such as Gemini into Tech Tech (Waymo has not experienced any other models of artificial intelligence alongside GEMINI from Google, as confirmed by Thiumalai). Robotaxi has published many research papers in multimedia models, including a City traffic simulator With the Girls model as well as Waymo Research around it whateverMultimedia from Waymo for independent driving. Waymo said that training its cars with EMMA helped in things such as discovering organisms and graphs on the roads, saying that there is a “possibility” for Emma as a general model for independent driving applications. However, Emma is expensive, and it can only process a small number of photo frames, and does not include LIDAR sensors or radar – all of which lead to “challenges” to use EMMA as an independent driving model ”
Thirumalai said that the integration of obstetric artificial intelligence into the self -driving stack is a field of “intense research”, and that he believes this will continue. “But there is a lot of work that will be necessary to make the system as simple as possible,” he said.
https://fortune.com/img-assets/wp-content/uploads/2024/10/IMG_0573-e1729178270444.jpeg?resize=1200,600
Source link