Prediction-Based Target Tracking Using Adaptive Input and State Estimation for Real-Time Numerical Differentiation
Shashank Narayan Verma, Dennis Bernstein · 2025
This paper presents a novel target tracking algorithm that uses numerical differentiation to estimate the velocity and acceleration of maneuvering targets. By applying adaptive input and state estimation, the first and second derivatives of noisy, sampled position data are computed with minimal latency. The estimated derivatives are then used to predict the target's position. The predicted motion of the target can be used to distinguish between ballistic and maneuvering targets, as well as to facilitate target interception. The performance of the proposed method is validated through detailed, comparative simulations, demonstrating significant improvements in tracking prediction and efficiency compared to existing methodologies.