A plug-in for "full map" in autonomous driving? A CP game between Xiaopeng and AutoNavi

Do you have to choose between single-vehicle capabilities and V2X?

I don’t think so. In fact, high-precision maps are just the beginning of V2X.

The reason for this question is that Tesla does not rely on maps, but relies on powerful visual capture and computing capabilities to achieve good map navigation functions. I think this is also a helpless move. After all, Tesla has no map data available in China. Since Chinese companies have geographical advantages, why not do it?

Furthermore, full-stack autonomous driving technology and high-precision maps are of great help to users in ensuring accuracy, safety and experience. When changing ramps, the vehicle routing may be more accurate if high-precision maps are used. As mentioned above, the "predictive" ability makes drivers feel more at ease.

Also, don’t forget that highways are just one of many driving assistance systems. The ultimate autonomous driving must be universal. It will gradually be used on ordinary roads and even urban roads. The complexity of these roads is incomparable to highways.

Just like the data collection of high-precision maps, at this stage, we must rely on AutoNavi's professional high-precision data collection vehicles to collect data, as well as supplementary collection after changes. With such a large-scale infrastructure construction and information service capabilities, perhaps only China can deepen V2X...

Perhaps Tesla's current vision + algorithm approach can make map navigation more widely used in a short period of time. However, once a true high-precision map is matched with sufficient visual detection + computing power, the experience will definitely be better and more complete. High-precision maps are also a part of V2X, and a very important part.

This is the era of true autonomous driving.