Autonomous Tesla vehicles excel in second Chinese self-driving assessment, yielding fatal consequences from certain manufacturers
Chinese media outlet Dongchedi recently conducted a series of tests on Advanced Driver Assistance Systems (ADAS) in urban driving scenarios. The tests, which took place on actual public roads that were temporarily shut down for the purpose of testing, aimed to simulate real-world situations that an SAE Level 2 driver's aid would encounter.
The tests involved multiple car models from Tesla, Mercedes-Benz, BYD, Xiaomi, Toyota, Xpeng, and others, making the results relevant for a wide spectrum of vehicles available in the Chinese market and beyond.
In the urban driving tests, the Model X from Tesla performed exceptionally well, avoiding collisions in 8 out of 9 tests during urban driving, taking the top spot. However, it failed the reverse test by driving through at high speed and clipping the rear of the car. The Model 3, on the other hand, performed similarly on the reverse test but failed other tests, resulting in a lower ranking compared to the Model X and several other vehicles.
Interestingly, Avatr and Aito had roughly similar performance brand-wide as Tesla did in these tests, so all three (Tesla, Avatr, and Aito) should be considered the winners of this test, along with Toyota's bZ3X which fared much better here than on the highway tests.
However, many of the tested systems failed to notice or react properly in situations where a driver's eyes would have trouble anticipating a collision due to the A-pillar blind spot. The Great Wall Motors Wey Lanshan, which was a top-performer on the highway tests, scored one of the worst in urban driving tests.
The urban tests included road users like scooter riders and children, which should be reflected in the training data that companies use to train their ADAS systems. The tests involved various scenarios such as a vehicle obstructing the path, merging into traffic, scooters and children crossing the road, a broken down car, and unsignaled intersections.
Dongchedi's urban driving scenario tests are important because they provide large-scale, real-world evaluations of ADAS under complex, realistic urban conditions, which involve vulnerable road users and chaotic traffic environments. These tests matter because they expose real ADAS performance in multi-scenario, high-traffic environments rather than simplified or isolated cases, revealing practical limitations or successes.
Lu Guang Quan, from the Beijing University of Aeronautics and Astronautics, pointed out that ADAS systems trained on machine learning can learn poor behaviors from the dataset, and these can be harder to correct than rule-based systems would be. The tests also highlighted challenges in system behavior such as lawbreaking indecision or the ability to handle complex urban conditions.
Such evaluations push manufacturers to improve their technologies, especially in urban areas where accidents can be more severe due to vulnerable road users. The public and industry gain insights into how real-world conditions differ from ideal test environments, supporting better regulatory, technical, and consumer understanding.
The test videos are available on Dongchedi's Youtube channel DCARSTUDIO. Carnewschina assembled a table of the results, which includes the urban driving test results, and linked it in their article. Compared to the previous highway test, there was less inconsistency within vehicle brands this time around. However, many of the tested systems routinely broke basic traffic laws and showed poor driving etiquette.
Zeekr performed among the worst in these urban driving tests, as it did in the highway tests. Xiaomi had middling to disappointing results in the urban driving tests. The test covered 26 cars and 9 scenarios, leading to a total of 234 simulations.
In short, Dongchedi’s urban tests provide critical, realistic benchmarks for ADAS development and evaluation by simulating the complexity and unpredictability of real urban traffic, thus helping advance safer and more effective driver-assist technologies.
- The results of the urban driving tests conducted by Dongchedi can help finance investments in green energy technology for electric vehicles within the automotive industry by revealing practical limitations or successes of ADAS systems, especially in urban areas where accidents involving vulnerable road users can be more severe.
- Technology advancements in electric vehicles, such as electric vehicles equipped with superior ADAS systems, could significantly enhance the transportation sector's sustainability by reducing accidents, thus improving overall safety and efficiency.
- As the finance industry increasingly invests in green energy and technology-driven transportation solutions like electric vehicles, it is essential to consider the need for ADAS systems that perform well in urban scenarios to ensure the safety and comfort of both drivers and other road users.
