Data-Driven Approaches for Modelling Target Behaviour

TL;DR

This paper compares three machine learning-based methods—Gaussian Processes, IMM filter parameter learning, and LSTM networks—for modeling target motion to improve tracking accuracy when true dynamics are unknown or complex.

Contribution

It provides a comparative analysis of three novel data-driven motion modeling approaches against traditional methods in tracking scenarios.

Findings

LSTM-based models outperform traditional filters in complex motion scenarios.

Gaussian Processes provide accurate predictions with less training data.

IMM filter parameter learning offers a good balance between complexity and performance.

Abstract

The performance of tracking algorithms strongly depends on the chosen model assumptions regarding the target dynamics. If there is a strong mismatch between the chosen model and the true object motion, the track quality may be poor or the track is easily lost. Still, the true dynamics might not be known a priori or it is too complex to be expressed in a tractable mathematical formulation. This paper provides a comparative study between three different methods that use machine learning to describe the underlying object motion based on training data. The first method builds on Gaussian Processes (GPs) for predicting the object motion, the second learns the parameters of an Interacting Multiple Model (IMM) filter and the third uses a Long Short-Term Memory (LSTM) network as a motion model. All methods are compared against an Extended Kalman Filter (EKF) with an analytic motion model as a benchmark and their respective strengths are highlighted in one simulated and two real-world scenarios.