Batch Estimation

Batch estimation is a technique for refining estimates of system states by processing all available data simultaneously, aiming for improved accuracy and robustness compared to sequential methods. Current research focuses on developing and comparing various batch estimation approaches, including those based on nonlinear least squares, Gaussian processes, splines, and data-driven methods like Koopman linearization, often applied to problems involving sensor fusion and trajectory estimation. These advancements are significant for improving the accuracy and reliability of navigation systems in diverse applications, such as underwater robotics and autonomous vehicles, particularly in scenarios with noisy or incomplete data.