Hugo platform

High efficiency, high performance, low delay and high precision

Sensors

Sensors    

Embedded AI solution

Embedded AI solution

Vehicle control

    Vehicle control

Cognition

Cognition

perception|localization|prediction

Environmental perception and object detection, recognition, and prediction based on deep learning
Use structural deep neural network for highly efficient and comprehensive computation

Modeling

Modeling

semantic 3D environment modeling

Homogenous perception interface for different sensors
Semantic environment reconstruction by high scalability and modularized sensor architecture

Decision

Decision

decision and path planning

Rational end-to-end method in compliance with functional safety
Transparent, traceable and inferenceable

Hugo 1.0

Hugo 1.0

Solution Platform Roadmap

Image input and processing, 720p@30fps;
Vehicle detection distance up to 150m,
and tracking range up to 250m;
Pedestrian detection distance up to 100m,
and tracking range up to 150m;
Lane line detection, curvature > 100m, distance up to 150m;
Types of identifiable lane lines;
Detection of drivable area within 90m with an accuracy rate >96%;
L2 driver assistance features, including FCW, LDW, ACC, AEB, LKA, etc.

Hugo 2.0

Hugo 2.0

Image input and processing, 1080p@30fps;
Simultaneous access to up to 4-channel cameras;
Detection of vehicles, pedestrians,
lane lines and drivable areas;
Detection and recognition of traffic signs including guideboards, road signs, surface marks, traffic sign texts and symbols;
Detection of general roadblocks and road surface defects;
Processing and fusion of laser radar and millimeter-wave radar.

Hugo 3.0

Hugo 3.0

Environmental perception and prediction based on multi-sensor fusion;
Simultaneous access to up to 8-channel cameras and multi-sensor fusion;
Active positioning and odometry independent of high-precision map;
Road environment analysis and driving decision making;
Route planning and vehicle control.