Xiaomi EV begins rolling out enhanced HAD driver-assist system for upgraded experience
Xiaomi EV announced today that it has begun rolling out the Xiaomi HAD (Hyper Autonomous Driving) Enhanced Edition driver assistance system to eligible vehicles, delivering an upgraded experience.
The enhanced Xiaomi HAD incorporates reinforcement learning algorithms and a world model, enabling the system to continuously explore more optimal driving strategies, the company said on Weibo today.
By incorporating reinforcement learning algorithms, the system can repeatedly explore within a virtual environment constructed by the world model, receiving rewards or penalties to derive optimal driving strategies.
The world model enables large-scale cluster management, allowing the system to simultaneously explore over 100 routes and rapidly accumulate experience across vast scenarios, it added.
With these capabilities, the enhanced Xiaomi HAD delivers smoother acceleration and deceleration, more decisive lane changes, and more precise road condition recognition, according to the company.
Xiaomi EV reiterated that assisted driving is not autonomous driving, and drivers must remain fully attentive at all times.
Xiaomi EV univeiled the enhanced Xiaomi HAD system on the opening day of the 2025 Guangzhou auto show on November 21, introducing reinforcement learning and a world model to build upon the 10-million-clip version.
Over recent years, China's leading EV manufacturers have targeted end-to-end technology as the advanced solution for assisted driving capabilities. In the past year, multiple companies have begun shifting toward world models to simulate human driver behavior.
Xiaomi EV began rolling out its end-to-end technology-based Xiaomi HAD assisted driving feature to all eligible vehicles in late February this year, subsequently positioning the free use of this feature as a key selling point.
By late July, Xiaomi HAD upgraded to a version trained on 10 million clips, delivering significant enhancements to its assisted driving capabilities.
