.png){kind=icon}
.png){kind=icon}
Return
Rectangular Geometric Constraints-Based Extended Object Tracking With Rigorous Shape Estimation
.png){kind=icon}
PRE
AI
J
W
I
J
DOI：10.1109/taes.2026.3712708.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Modern high-resolution sensors have enabled extended object tracking (EOT) to simultaneously estimate both kinematic and shape states. Moreover, rectangular objects, such as vehicles, are of particular interest in practical applications. However, most existing EOT methods are not tailored for rectangular objects, leading to nonstrictly rectangular shape estimation. The few specialized rectangular EOT approaches suffer from computational inefficiency and instability under sparse measurements due to their reliance on the sampling and association framework. To address these limitations, in this article, we propose a rectangular geometric constraints (RGC)-based EOT method. First, we develop a measurement model using a customized radial function that enforces RGC. This model directly maps the measurements into minimal rectangular shape parameters (length/width) while avoiding complex sampling and association procedures. Thus, this formulation improves both efficiency and stability. In addtion, our model is general and applicable to three typical object scattering cases: fully observable contour, partially observable contour, and observable surface, accounting for various sensor properties and viewing perspectives. For each case, we analytically derive the likelihood functions that ensure geometric rigor and conjugate prior compatibility. To handle nonlinearities, an RGC-extended Kalman filter (RGC-EKF) is implemented for simultaneous tracking and shape estimation. Simulation and experimental results demonstrate that the proposed RGC-EKF not only achieves low computational complexity, but also outperforms state-of-the-art approaches in terms of tracking accuracy and shape fidelity, particularly under sparse measurements.
Keywords:
Extended object tracking (EOT)
rectangular geometric constraints (RGC)
rigorous shape estimation
Journal
IF:
5.7
Papers:
651
Citations:
2.4W
Organization
Cited Papers