Constant Time-Delay Leader Following with Neural Networks and Invariant Extended Kalman Filters for Arbitrary Trajectories

2026-06-17

AI Summary

This work proposes a novel approach that integrates sequence-to-sequence (Seq2Seq) neural networks with an invariant extended Kalman filter (IEKF) to enable vehicle platoons to track arbitrary nonlinear trajectories under constant time delays, without inter-vehicle communication, a shared coordinate frame, or global positioning. By operating on the SE(2) manifold, the method leverages Seq2Seq to predict the leader’s relative trajectory and incorporates geometric model predictive control to enhance tracking accuracy. This is the first framework to combine Seq2Seq with IEKF in this context, significantly reducing reliance on expert-designed models while accommodating arbitrary velocity profiles and motion patterns. Experimental results from both simulations and real robotic platforms demonstrate superior tracking performance for long-delay nonlinear trajectories compared to purely IEKF-based or purely learning-based approaches.

Abstract

This paper proposes a constant time-delay trajectory tracking method for vehicle convoys operating without inter-vehicle communication, a common coordinate system, or global positioning. The method integrates a probabilistic sequence-to-sequence (Seq2Seq) neural network with an invariant extended Kalman filter (IEKF) to warm-start the prediction process, allowing accurate estimation of a leader vehicle's relative trajectory on the SE(2) manifold. A geometric model predictive controller is further incorporated to fully exploit the manifold-based trajectory predictions for improved control performance. The system can handle arbitrary nonlinear trajectories with varying speeds and motion profiles while reducing the need for expert-based domain knowledge for the design of trajectory following systems, even under long trajectory delays. The effectiveness of the method is validated through comparisons with a pure IEKF baseline, learning-based methods, and the ground-truth trajectory in kinematic simulations, as well as in experiments using real robotic vehicles.

Problem

• time-delay tracking
• leader-following
• arbitrary trajectories
• vehicle convoys
• SE(2) manifold

Innovation

• Seq2Seq neural network
• invariant extended Kalman filter
• SE(2) manifold
• model predictive control
• time-delay trajectory tracking