Technology and security pain points still exist, and the popularization of intelligence faces obstacles
When it comes to automotive intelligence, automatic driving is undoubtedly the most representative technology, and its realization requires a variety of technologies to promote together, such as positioning and navigation technology, environmental perception ability, automatic control technology, etc., and automatic driving from Level 1 to Level 5, each step up, the amount of computing increases by an order of magnitude. At present, traditional car manufacturers such as BMW and Mercedes-Benz are developing Level 3 autonomous driving, but there is still some distance from commercialization.
Traditional autonomous driving technology focuses on improving the capabilities of a single vehicle to achieve a higher level of autonomous driving, but installing a higher number or more accurate sensors on the vehicle cannot solve the problem of limited sensing range of a single vehicle. Although the problem of insufficient computing resources can be solved by equipping vehicles with high-performance computing systems, this will greatly increase the cost of self-driving cars and is not conducive to the popularization of self-driving cars. The picture shows the unmanned edge computer T609TW-T609 is a computing platform based on NVIDIA Jetson AGX Xavier core module design, with 32TOPS floating point computing AI processing capabilities, fanless passive heat dissipation, rich I/O interfaces, support 8-channel HD cameras, reserved bottom bracket for easy on-site installation
Based on the above, the industry believes that it is imperative to establish a complete set of support systems for on-board chip-end computing-road network edge computing-data center cloud computing, and edge computing plays an indispensable role in it.
