Accurate Prediction and Estimation of 3D-Repetitive-Trajectories using Kalman Filter, Machine Learning and Curve-Fitting Method

TL;DR

This paper presents a hybrid approach combining Extended Kalman Filter, machine learning, and curve-fitting to accurately predict 3D repetitive trajectories of high-speed targets, validated in simulation and hardware.

Contribution

It introduces a novel method integrating EKF, machine learning, and curve-fitting for precise 3D trajectory prediction of fast-moving targets.

Findings

Effective in noisy visual environments

Accurate future position estimation for repetitive trajectories

Validated in ROS-Gazebo and hardware implementations

Abstract

Accurate estimation and prediction of trajectory is essential for the capture of any high speed target. In this paper, an extended Kalman filter (EKF) is used to track the target in the first loop of the trajectory to collect data points and then a combination of machine learning with least-square curve-fitting is used to accurately estimate future positions for the subsequent loops. The EKF estimates the current location of target from its visual information and then predicts its future position by using the observation sequence. We utilize noisy visual information of the target from the three dimensional trajectory to carry out the predictions. The proposed algorithm is developed in ROS-Gazebo environment and is implemented on hardware.