Lei Jun: Why the Xiaomi Car won't be a cheap vehicle
Before the launch of the Xiaomi SU7, Lei Jun openly stated that this would not be a cheap car, and in fact, it might even be a bit expensive. As the first car from Xiaomi, what makes the SU7 expensive? Lei Jun broke down the technological highlights of the Xiaomi car into five parts: intelligent driving, intelligent cockpit, super motor, battery, and integrated die-casting. In terms of intelligent driving, Xiaomi has developed its own perception technology with a zoomable BEV (Bird's Eye View) that allows the vehicle to change visual schemes like human eyes in different scenarios, sensing the environment in real-time and dynamically adjusting. For example, in a parking lot environment, the algorithm can divide the entire parking lot into millions of grids, each grid measuring 5cm. Such precision can accurately identify the open or closed status of ground locks. When driving on city roads, the view expands to 160m, and on highways, the vehicle can see up to 250m away. The super-resolution occupancy network technology enables the vehicle to better recognize the appearance of objects in the surrounding scene. It can describe obstacles and events as blocks, and with vector algorithms, these independent blocks are combined into models composed of continuous surfaces. This gives the SU7 the ability to clearly see objects, even accurately identifying the contours of irregular obstacles, such as utility poles, parking bollards, etc. The road large model technology can generate road topology in real-time, accurately creating corresponding driving guidance lines based on road conditions. It's like how a person judges the road in real-time, smoothly navigating around construction or congested areas. Lei Jun showcased recorded scenarios of urban navigation, automatic car-finding in parking lots, and automatic parking in mechanical garages. The road large model allows the vehicle to smoothly reverse park even in mechanical garages with just 0.05m of space left on either side. Besides algorithmic enhancements, these functions also rely on a powerful hardware system. The chip configuration for the SU7's intelligent driving is on par with industry standards: Dual NVIDIA Orin X chips, with a total computing power of 508 TOPS Qualcomm Snapdragon 8295 automotive chip 27 perception hardware components: 1 LIDAR, 11 high-definition cameras, 3 millimeter-wave radars, and 12 ultrasonic radars. Music track: Office by Aylex Source: freetouse.com/music Free Background Music (No Copyright)
Before the launch of the Xiaomi SU7, Lei Jun openly stated that this would not be a cheap car, and in fact, it might even be a bit expensive. As the first car from Xiaomi, what makes the SU7 expensive? Lei Jun broke down the technological highlights of the Xiaomi car into five parts: intelligent driving, intelligent cockpit, super motor, battery, and integrated die-casting. In terms of intelligent driving, Xiaomi has developed its own perception technology with a zoomable BEV (Bird's Eye View) that allows the vehicle to change visual schemes like human eyes in different scenarios, sensing the environment in real-time and dynamically adjusting. For example, in a parking lot environment, the algorithm can divide the entire parking lot into millions of grids, each grid measuring 5cm. Such precision can accurately identify the open or closed status of ground locks. When driving on city roads, the view expands to 160m, and on highways, the vehicle can see up to 250m away. The super-resolution occupancy network technology enables the vehicle to better recognize the appearance of objects in the surrounding scene. It can describe obstacles and events as blocks, and with vector algorithms, these independent blocks are combined into models composed of continuous surfaces. This gives the SU7 the ability to clearly see objects, even accurately identifying the contours of irregular obstacles, such as utility poles, parking bollards, etc. The road large model technology can generate road topology in real-time, accurately creating corresponding driving guidance lines based on road conditions. It's like how a person judges the road in real-time, smoothly navigating around construction or congested areas. Lei Jun showcased recorded scenarios of urban navigation, automatic car-finding in parking lots, and automatic parking in mechanical garages. The road large model allows the vehicle to smoothly reverse park even in mechanical garages with just 0.05m of space left on either side. Besides algorithmic enhancements, these functions also rely on a powerful hardware system. The chip configuration for the SU7's intelligent driving is on par with industry standards: Dual NVIDIA Orin X chips, with a total computing power of 508 TOPS Qualcomm Snapdragon 8295 automotive chip 27 perception hardware components: 1 LIDAR, 11 high-definition cameras, 3 millimeter-wave radars, and 12 ultrasonic radars. Music track: Office by Aylex Source: freetouse.com/music Free Background Music (No Copyright)