On September 24, 2024, Hyper unveiled its latest intelligent driving system NDA 3.0. Gu Huinan, the principal of Hyper, accompanied Mr. Hu Run, chairman of the “Hurun Rich List”, to experience the Hyper HT equipped with the new intelligent driving system. They demonstrated the performance of NDA 3.0 under diverse complex road conditions via a live stream.
NDA 3.0 applies an end-to-end architecture, which solely relies on a single neural network from input to output. This minimises the transfer of intermediate structured cognition results, enables lossless information transmission, and covers a broad range of application scenarios.
Hyper adopts the advanced LIDAR technology, which excels in dealing with various extreme driving scenarios as well as accurately measuring the distance and speed of obstacles. The integration of LIDAR enhances NDA 3.0’s capabilities of identifying and dealing with obstacles accurately when facing complicated traffic conditions. For instances, NDA 3.0 handles scenarios such as multi-exit roundabouts, night-time head-on illumination, complex construction sections efficiently.
Both static and dynamic recognition, as well as intersection interactions and low-speed manoeuvres, can be accomplished by Hyper’s NDA 3.0. With an end-to-end commuting experience, it effortlessly handles complicated scenarios such as multi-exit roundabouts, night-time head-on illumination, and complex construction sections, just like an experienced driver comprehending the traffic environment. It operates independently of high-precision maps, requiring only to set the navigation; the rest of the driving tasks are undertaken by Hyper’s intelligent driving technology.
In the actual test, Hyper’s NDA 3.0 demonstrated excellent performance. Mr. Hu Run specifically mentioned NDA 3.0’s performance in complex scenarios, including construction sites with sandy soil and crowded town markets. Due to the limited coverage of high-precision maps, combined with road-narrowing construction, driving through construction sites with sandy soil becomes more challenging. However, with Hyper’s NDA 3.0’s static and dynamic obstacle recognition, the Hyper HT can identify and bypass oddly-shaped obstacles such as boxes, stones, mounds of soil, and animals, despite facing complex traffic conditions like constructions and traffic merging.
Most of the town market areas consist of narrow roads filled with illegally parked vehicles on both sides. In the live test, it was evident that the Hyper HT equipped with NDA 3.0, despite facing incoming cars and illegally parked cars along the roadside, did not abandon the autonomous driving mode. Instead, it pre-emptively predicted the movement intentions of vehicles and pedestrians and responded appropriately, managing to navigate through this challenging road section without stopping.
Apart from the above scenarios, the arrival of Hyper NDA 3.0 also exhibits its proficiency in complex special scenarios such as congestion lane-changing, turning, lane marking confusion, tidal lanes, multi-intersection roundabouts, U-turn intersections, and so forth.
Apologies, but your text seems to be incomplete. Please provide the complete sentence or context that needs to be translated.
This article is a translation by AI of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.