基于GPU集群的雷达探测空间目标并行信号处理系统
By using a GPU cluster-based radar detection system and fiber optic and Ethernet connections, real-time sliding window processing of radar signals was achieved. This solved the problem of low data rate caused by high computational load in existing systems, improved processing efficiency, and provided scalability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING INST OF TECH
- Filing Date
- 2024-01-16
- Publication Date
- 2026-07-17
AI Technical Summary
Existing radar signal processing systems require a huge amount of computation when performing sliding window processing, which cannot meet the needs of real-time processing. Especially when the target trajectory error is large or there is no prior trajectory information, long-term coherent accumulation processing requires traversing and searching all possible trajectory parameter spaces of the target, resulting in a significant reduction in measurement data rate.
A radar detection system based on a GPU cluster is adopted, including a fiber optic data receiving and forwarding module, a fiber optic data distribution module, multiple GPU processing nodes and a comprehensive processing module. Data transmission and computing task allocation are realized through fiber optic and Ethernet connections, and signal processing is performed using the parallel processing capabilities of FPGA and GPU.
It achieves real-time processing of radar signals in sliding window mode, improves data rate, meets the requirements of strong real-time control and ultra-high-speed computing, and the system is scalable, which can increase computing power by adding processing nodes.
Smart Images
Figure CN117872282B_ABST