Qianli Technology unveils AI-powered 4D millimeter-wave radar with 10 TOPS NPU
Qianli Technology has launched an AI-integrated 4D millimeter-wave radar equipped with a 10 TOPS NPU, adding local perception capabilities to the sensor itself. The company said the product supports three operating modes: a local edge mode that processes data on the radar and outputs structured perception results, a satellite architecture mode that sends raw data at high speed to a central domain controller for multi-source early fusion, and a compatibility mode that runs both paths at the same time. According to the release, the radar uses an end-to-end model across the full chain from antenna design to point-cloud output. Qianli Technology said this delivers a 10x increase in point-cloud density, a 40% improvement in detection range, and a 95% ghost-target elimination rate. The design is aimed at vehicle systems spanning L2 to L4 functional levels, as well as the transition period toward centralized EE architectures. The company is led by Yin Qi, while its affiliated company StepStar provides AI model support. Quantum Bit described the launch as an extension of the "AI + physical terminal" concept into autonomous-driving sensors, with model-driven iteration reshaping the underlying sensor architecture.
