An improved track-before-detect algorithm enhances maritime surveillance with GNSS signals

A novel two-stage Track-Before-Detect (TbD) method has been developed to significantly improve the detection of weak moving targets using reflected Global Navigation Satellite System (GNSS) signals. This approach addresses the challenge of motion model mismatch during long integration times, which traditionally degrades detection performance. The method first corrects range and Doppler migrations to concentrate target energy within individual scans, then recursively combines candidate plots across multiple scans to enhance detection reliability. Simulations and field trials confirm its effectiveness for both maneuvering and non-maneuvering targets, achieving high detection accuracy with substantially reduced computational burden.

Global Navigation Satellite System -reflectometry (GNSS-R) has evolved from mitigating multipath interference to a valuable remote sensing tool, applicable in oceanography, soil moisture monitoring, and deformation measurement. The inherently weak power of GNSS signals reflected from moving targets poses significant detection challenges, especially in maritime environments where signal sources are scarce. Long integration times are often necessary to improve signal-to-noise ratio, but target motion during these periods can cause mismatches with assumed motion models, reducing integration gain and detection reliability. Based on these challenges, there is a critical need to develop advanced detection methods that maintain accuracy while improving computational efficiency.

Researchers from Hohai University and The Hong Kong Polytechnic University published (DOI: 10.1186/s43020-025-00176-7) a study on September 1, 2025, in Satellite Navigation, introducing an improved Track-Before-Detect (TbD) technique for moving target detection using GNSS reflected signals. The method employs a two-stage architecture that combines long-time hybrid integration (LTHI) with kinematic tracking across scans, enabling efficient detection of low-signal targets in noise -heavy environments.

The proposed method operates in two phases. First, it uses LTHI to correct range and Doppler migrations within short-duration scans, concentrating target energy in range-Doppler maps. Then, rather than applying a high threshold for detection, it extracts candidate plots using a low threshold and recursively associates them across scans based on kinematic constraints. This allows the system to accumulate target energy over time while filtering out false alarms. A key innovation is the use of the characteristic of the stack of range-Doppler maps to narrow down plot associations, drastically reducing computational load. Compared to conventional Dynamic Programming TbD, the new method maintains similar detection performance and parameter estimation accuracy—with RMSE values under 4 m in range, 0.15 Hz in Doppler, and 0.045 Hz/s in DFR—while cutting processing time by nearly half in some cases. The method also supports multi-target detection, as demonstrated in simulations with three simultaneous targets.

"This work represents a meaningful advance in passive radar technology," says an expert in satellite navigation and remote sensing. "By intelligently leveraging signal characteristics and motion constraints, the method achieves high detection performance with greatly improved efficiency, making it suitable for real-time maritime surveillance applications."

This technology has significant implications for maritime security, search and rescue, and traffic monitoring in open or remote waters where traditional radar coverage is limited. Its ability to operate with existing GNSS infrastructure reduces deployment costs and increases accessibility. Future work may focus on integrating target detection with localization and velocity estimation across multiple satellites, further enhancing the utility of GNSS-R for comprehensive maritime domain awareness.

###

References

DOI

10.1186/s43020-025-00176-7

Original Source URL

https://doi.org/10.1186/s43020-025-00176-7

Funding information

This work was supported by the National Natural Science Foundation of China (Grant No. 524031511, 42274051).

About Satellite Navigation

Satellite Navigation (E-ISSN: 2662-1363; ISSN: 2662-9291) is the official journal of Aerospace Information Research Institute, Chinese Academy of Sciences. The journal aims to report innovative ideas, new results or progress on the theoretical techniques and applications of satellite navigation. The journal welcomes original articles, reviews and commentaries.