Based on the FusionWise software platform, the FusionDrive2.0 software platform integrates AD basic suite and cross-board AD suite, thus providing an exclusive solution that meets the needs of positioning, perception, and planning control in the intelligent driving domain.

Algorithm Library for L2/L2+

  • Parking: APA、RPA、AVP
  • L2+: HWA、TJA etc

Automotive Vision Lab

  • Cooperate with NVIDIA to build an automotive vision laboratory, formulate 40 scenes in 7 categories of objective, 48 scenes in 6 categories of subjective, and form 1000+ visual data standards, to create the most stable, safest and most efficient bottom layer of automotive vision for customers software platform

Communication Scheduling Components

  • TSN Protocol Stack
  • SOME/IP Protocol Stack
  • DDS Protocol Stack

FusionWise(AD Version)

  • Communication
  • Network Management
  • Routing
  • Big Data
  • Time Synchronization
  • OTA
  • Log & Trace
  • Persistence
  • Diagnostic
  • XCP
  • Power Management
  • Health Management
  • Function Safety & Security

Automotive Vision Lab

Automotive Vision Lab

Automotive Vision Laboratory takes the vehicle-mounted field as an entry point, and aims at technical difficulties such as poor imaging quality, weak graphics processing ability, weak anti-interference ability, and scene generalization faced by intelligent driving visual perception, and integrates image quality optimization, Image module adaptation and machine algorithm recognition provide visual solutions for intelligent driving and better help car OEMs and Tier1 improve product quality.


Verify the basic performance of the visual module on the Nvidia platform, ensure that the hardware performance of the visual module meets the application requirements, and ensure product stability, adaptability, and authenticity

Use laboratory supporting optical test equipment, high dynamic test kits and other vehicles to more efficiently realize the test and verification of experimental scenarios and improve the pass rate of vehicle road tests

Possess the whole-process capability of effect acceptance; use professional laboratory equipment to output professional evaluation reports with industry credibility for the combination of different hardware, different algorithms, and different