RoboSense RS-Fusion-P5 for L4&L5 Autonomous Driving

A Fusion of best-in-class 128 laser beams LiDAR RS-Ruby and blind-spot detection LiDAR RS-Bpearl, the RS-Fusion-P5 solution for RoboTaxi is able to reach a detection range of 200m, and high-precision resolution of 0.1° with full sensing coverage, which can detect all objects with visible level details, such as lane lines, passenger cars, vans, trucks, bike, pedestrian, etc.

△Images of obstacles and road details are clear with large contrast and visible details

The RS-Fusion-P5 adds AI perception algorithm "RS-LiDAR-Algorithms" to empower L4&L5 autonomous vehicles with structured semantic-level perception functions in real time, including ground detection, obstacles detection & classification & tracking, attention object detection, free space detection, etc. It gives outstanding performance in various complex Chinese road conditions.

The RS-LiDAR-Algorithms have been developed by the RoboSense R&D team for more than 10 years and have been jointly verified to be mature and reliable by over 100 partners in tens of thousands of driving scenarios and have overcome various extreme corner cases.

△Mixed traffic scenes with pedestrian and cars on city roads

The RS-Fusion-P5 achieves accurate ground detection and can distinguish ground points from non-ground points, laying a solid foundation for obstacle and road detection, pothole processing, and free space detection.

△The detected ground point cloud data are marked in gray

The RS-Fusion-P5 provides comprehensive and accurate obstacles detection, classification and tracking, ensuring the safety of autonomous driving in complex environments.

△Different colors used in this image is to recognize multiple obstacles' types

The close-up obstacles are the greatest threat to the safety of autonomous driving. The RS-LiDAR-Algorithms bring a new function: attention object detection. It marks the obstacles closest to the LiDAR in every direction as attention objects, including pedestrians, vehicles, guardrails, trees, and curbs, etc.

△The bounding boxes of attention objects are emboldened

Based on the comprehensive and accurate detection of various environmental information, the RS-Fusion-P5 outputs free space detection information to support path planning for self-driving cars.

△The green modules indicate free space detection