During the 2026 Beijing International Automotive Exhibition, Cheng-Tech officially launched InsightRadar, its seventh-generation centralized computing radar , marking another step forward in the evolution of radar perception architecture and data processing.

AI-Powered Centralized Computing Reshapes the Radar Perception Paradigm

Designed for the evolving needs of intelligent driving systems, InsightRadar is driven by AI and enables millimeter-wave radar to shift from target output at the sensor edge to centralized perception and sensor fusion.

Distributed radar nodes capture raw, high-fidelity data and transmit it to the central computing platform, where AI models perform unified processing, fusion, and inference. This architecture improves perception accuracy, consistency, and scalability.

By bringing centralized radar data into the core data layer of the intelligent driving perception stack, the system provides greater information redundancy and supports the deployment of advanced L4 automated driving. The launch of InsightRadar represents a key milestone in Cheng-Tech’s development along this technology path.

At the show, Cheng-Tech also presented a centralized computing radar demonstrator built on the NVIDIA DRIVE AGX Thor platform. Through real-time data and perception visualization, the system demonstrates how radar operates in real-world scenarios.

A Radar Portfolio for Diverse Applications

At Auto China 2026, Cheng-Tech also showcased a broad portfolio of millimeter-wave radar products for passenger vehicles, commercial vehicles, and two-wheelers, addressing intelligent driving requirements across multiple levels and use cases.

Building on its centralized computing architecture, Cheng-Tech continues to expand its product portfolio and technology capabilities, including next-generation high-performance radar systems and multi-dimensional perception solutions. These technologies are also extending into emerging applications such as low-altitude mobility and robotics, enabling richer environmental perception across a broader range of scenarios.