Every automaker and their cousin will lead us into an autonomous driving future… according to their press offices. In reality, I think there are only a few automakers that have deep internal teams that focus on impressive internal innovation to achieve that ultimate goal. In the shorter term, that kind of leadership also allows them to have the best ADAS (advanced driver assistance system) features. Aside from Tesla, XPeng is the top company I’ve been eyeing in this area. I know they have a robust in-house team focused on the full development of autonomous driving technology and I know they have an interesting approach that integrates more types of hardware than Tesla (lidar, radar and cameras rather than just cameras). On the software side, I think they have more software developers focusing on this than anyone else but tesla. In short, I think it’s one of the top companies that could potentially innovate itself into a pole position in this area – temporarily, if not even in the long run. I am not to predict that (I’m not making such a prediction for any automaker, autonomous driving startup, or auto supplier), but it’s a definite possibility.
With that context in mind, I found it interesting to see how XPeng’s CEO, He Xiaopeng, reacted to the most advanced version of XPeng’s “City NGP” (City Navigated Guided Pilot) during a recent test drive of the technology. Watching the video, I was even more impressed to see that he was very open and critical of the weaknesses he saw in the technology, while of course also praising some of the more advanced capabilities. to have Tesla FSD Beta in my own car I know very well some of the biggest challenges on the road for this kind of technology, and you can see He Xiaopeng on some of these exact challenges in the video. He highlights fairly nuanced issues that keep the autonomous driving software from being as smooth and relaxing as a normal driver can be. Watch the video below and I’ll get back to some of these items in more detail.
Just looking at this short video of City NGP’s capabilities, the technology doesn’t look much different from Tesla Autopilot. It seems to be excellent at lane keeping and traffic-aware cruise control (aka adaptive cruise control), and it seems very adept at changing lanes automatically, but with some minor niggles like being too passive at times or the context of a traffic jam not recognize -full traffic jam in which other cars behave differently and you have little space to get into an opening. The VP of XPeng’s Autonomous Driving Center, Wu Xinzhou, notes in the video that a future version of the technology will take note of such traffic situations and ask the driver to assist with changing lanes in those environments. It’s not a perfect robotic axi approach to this challenge, but it seems the right one for this stage of the technology.
He Xiaopeng also notes a moment when a car that enters its lane somewhat aggressively is noticed by the XPeng City NGP suite and the software appropriately provides some extra space for that car to enter. That level of seemingly natural human-like adaptation is impressive and it shows how the XPeng team works to make City NGP as smooth and predictable as possible.
Anyone who has followed my commentary and review of Tesla FSD Beta knows that one of my biggest concerns and main reasons for withdrawal is potholes. No one wants to hit potholes, but the system can’t handle them yet. He Xiaopeng specifically asks about potholes and how the City NGP system handles them, and Wu Xinzhou specifies that it can’t quite recognize and avoid them yet, but the team is working on them and the next version of the technology “certainly Personally, I find it refreshing that they are willing to publicly share these kinds of existing weaknesses as they do (not every company would emphasize these things – and certainly not via CEO test drive video† That kind of authenticity and honesty is valued, and I don’t think just by me. I also find it encouraging that they are focused on solving the pothole problem. I’m really curious how well they do that in the next version of their semi-autonomous driving technology.
Wu Xinzhou also spends some time explaining how lidar helps the XPeng team deal with certain things, such as heavy rain. Again, this is one area where Tesla’s system can struggle a bit as it is 100% visibility based and the visibility challenges in such rain. I’d love to compare the two systems side by side in similar rainy conditions! Maybe someday XPeng can lend us a vehicle for that.
Wu Xinzhou also talks about ways lidar helps the company recognize and avoid static obstacles. I’m not sure if that intentionally casts a slight shadow on Tesla’s approach, but it’s a known issue Tesla vehicles have faced from not using lidar. Of course, Tesla is trying to solve that problem in other ways, but it seems remarkable that XPeng explicitly emphasizes that this is one of the reasons why they consider lidar to be crucial.
For the computer geeks out there, perhaps the most impressive comment is about the computing power XPeng uses for all of this. “We are the only company in the world that has succeeded in developing city-level ADAS functions based on only 20-30 TOPs of computing power,” said Wu Xinzhou. But instead of just patting himself on the back at that truly impressive feat, he jumps straight into self-criticism. “But we also have limitations. The computing capabilities are simply not enough for certain specific situations.” He then goes on to talk about some of the fringe cases they’re working on, such as people lying down, people in wheelchairs, and people squatting.
Overall, I thoroughly enjoyed what seemed like an open, candid look at where XPeng’s semi-autonomous driving technology sits at the moment. I highly encourage you to check it out. If any additional thoughts come to mind while watching the video, let us know in the comments below.
Do you appreciate the originality of CleanTechnica and the coverage of cleantech? Consider becoming one CleanTechnica member, supporter, technician or ambassador — or a patron on Patreon†
