Pioneer of Progressive AD Route With Full Stack Ability
Approach of Machine Learning First Enable Mass-Deployed Self-Driving
New Generation of AI Algorithms For the Ultimate Goal of AI-First Mobility
Leading Technolog of Visual Perception And Multi-Sensor Fusion

Pioneer of MLF Approach,
Realising Self-Driving Progressively.

All developed for mass deployment and applied to a dozen of models, with the capacity of self-improvement by mass real-world data, finally to realize autonomous driving in complex scenarios.

Machine Learning First

MLF is a new approach of autonomous driving, based on new technical architecture and driven by massive data. It provides excellent scalability and economical efficiency to handle new scenarios and solve the long-tail challenge. Utilizing all sensor inputs, MLF can generate a safe, proper and comfortable route for autonomous vehicle in complex realistic environment, where man-made rules cannot describe everything and react to all situations.


Following MLF, Nullmax’s visual perception technology contains all of software and algorithms required for autonomous driving perception. Some modules offer top-level performance in industry, including Object Detection, Lane Detection, Multi-Object Tracking, Fusion, Prediction, Mapping & Localization (V-SLAM), etc.

One SW Platform

Nullmax build the One SW Platform-BEV-AI Architecture. It supports all solutions and all products, supports perception of driving and parking, supports any sensor configuration.
The BEV-AI Architecture meeting diverse customer needs, has ultra-long sensing distance, support integrated mass-Production solution. And it optimized for onboard deployment, provide better support for easy deployment and continuous upgrading.

MaxPlatform Full-Stack Software for Car

MaxPlatform contains the application layer MaxStack (perception, planning & control) and the middleware MaxOS. On MaxPlatform, Nullmax can build a variety of solutions for different configurations and functions quickly.
  • MaxStack
  • MaxOS
  • Developed completely in-house, with top-level performance.
  • Decoupled and extensible system, adaptable to various configurations and platforms.
  • Advanced architecture for autonomous driving at all levels.
  • Complete toolchain from development to testing.
  • Support all kinds of scenarios and flexible customization.
  • Including perception, planning and control, with algorithm and engineering modules.
  • Top visual perception capability & multi-sensor fusion system.
  • New generation of architecture with more outstanding performance.
  • Modular design, flexible and convenient to customize.
  • Developed for autonomous driving, with characteristics of high-reliability, high-performance and high-concurrency.
  • Plug-in design, faster and more efficient to iterate software.
  • Ported on mainstream computing platforms, support multiple operating systems.
  • Optimized for popular chips.

MaxCloud Data System on Cloud

MaxCloud contains modules such as visualization and analysis, data filter and annotation, model training and evaluation. It extracts experience and knowledge from oceans of real-world data to feed autonomous driving system in a continuous and efficient pattern.