IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (Jan 2021)

Vessel Velocity Estimation and Tracking From Doppler Echoes of T/R-R Composite Compact HFSWR

  • Weifeng Sun,
  • Zhenzhen Pang,
  • Weimin Huang,
  • Yonggang Ji,
  • Yongshou Dai

DOI
https://doi.org/10.1109/JSTARS.2021.3071625
Journal volume & issue
Vol. 14
pp. 4427 – 4440

Abstract

Read online

Vessel speed and heading are two important kinematic parameters describing its state of motion. However, due to low spatial resolution of a compact high-frequency surface wave radar (HFSWR), vessel speed and heading cannot always be accurately estimated. Since HFSWR can measure vessel Doppler velocity with relatively high accuracy, it is possible to estimate a vessel's vector velocity based on two Doppler velocities measured along two different directions. In this article, a newly developed T/R-R composite compact HFSWR system is introduced, and a corresponding vessel velocity estimation method which employs a target's radial velocity and elliptical velocity, respectively, measured by the monostatic (T/R) and bistatic (T-R) settings is proposed. First, monostatic and bistatic tracks are independently generated using a multitarget tracking algorithm. Then, the obtained monostatic and bistatic tracks are matched using a track-to-track association method to determine the track pair belonging to each target. Subsequently, the associated track pairs are combined to produce fused tracks for improving positioning accuracy. Finally, vessel vector velocity is estimated based on the radial and elliptical velocities as well as the fused target position. Comparisons of vector velocity estimation results from radar field data with corresponding automatic identification system data demonstrate that the average root-mean-square-errors of the estimated speed and heading are 0.48 km/h and 3.9$^{\circ }$, respectively, which meets the practical requirements of a maritime surveillance system. Moreover, the velocity estimation error is analyzed via theoretical derivation and experimental verification. The proposed method shows good potential in further improving the tracking accuracy.

Keywords