Abstract Underwater acoustic source localisation is essential for marine monitoring, navigation of autonomous underwater vehicles and underwater surveillance. Time Difference of Arrival (TDOA) localisation is attractive because it avoids absolute time synchronisation; however, its accuracy degrades in realistic underwater channels due to multipath, measurement noise and environmental variability (e.g. sound-speed mismatch) as well as sensor geometry limitations. This paper proposes an optimisation-based TDOA localisation framework that integrates: (i) Kalman filtering (KF) for dynamic tracking; (ii) extended Kalman filtering (EKF) to handle nonlinear measurement models; and (iii) nonlinear least-squares (NLS) minimisation to refine the source position. A parametric analysis is also presented by varying key system parameters – primarily noise level and measurement uncertainty – to quantify performance trade-offs in terms of localisation error and convergence behaviour. Simulation results (static and moving source cases) show that LS provides high accuracy for low-noise/static cases, while KF/EKF are more robust for dynamic and high-noise scenarios; EKF achieves the fastest error decay due to explicit nonlinear modelling. These results demonstrate the proposed framework’s effectiveness for robust underwater acoustic source localisation.
Shaik et al. (Thu,) studied this question.