In the midsummer of 2026, the Mengshi M817 High-Performance Edition successfully accomplished a dual-route assisted driving challenge in Tibet. This vehicle is outfitted with Huawei’s cutting-edge ADS 5 intelligent driving system, which is the first NCA (Navigation-Assisted Cruise) Off-Road Edition in the industry, along with an 896-line LiDAR system.

One of the routes covered a distance of 266 kilometers along the iconic G318 National Highway. It included crossing the challenging Gyatsola Pass, navigating the notorious 108 hairpin turns near Mount Everest, and finally reaching the Gyawula Pass. The other route ventured into a remote and untamed area, traversing 60 kilometers of unmarked gravel roads to reach the Laqiongla Pass, which stands at an impressive altitude of 5,830 meters.

Throughout these extreme conditions—characterized by low oxygen levels, heavy dust, the absence of lane markings, and frequent sharp turns—the assisted driving system maintained stable operation. The 896-line LiDAR system, featuring a dual-optical-path architecture and a multi-sensor fusion matrix, enabled reliable road perception even in dusty environments. It could detect obstacles up to 120 meters away, including those as low as 14 centimeters.

The WEWA 2.0 decision-making architecture replicated human-like driving behaviors through cloud-based multi-agent gaming and online reinforcement learning. When combined with Safety Risk Field technology and the CAS 5.0 All-Dimensional Collision Avoidance System, it ensured safe and precise cornering. The NCA Off-Road Edition, specifically calibrated for off-road conditions and tire performance, operated continuously on unmarked and unpaved roads.

This challenge, which encompassed all terrain levels from 1 to 9, thoroughly validated the system’s technical capabilities in perception, decision-making, and scenario adaptation. It significantly expanded the applicable scenarios of assisted driving, extending its reach from urban paved roads to extreme off-road environments, thereby redefining the boundaries of assisted driving capabilities.