IEEE CASE 2026

A Real-Time Affordable Predictive Trajectory Tracking Controller for Differential Drive Robots via Imitation Learning

Concordia Institute for Information Systems Engineering, Concordia University
Predictive Cyber-Physical System Security Group (PreCySe)

Video Code BibTeX

Overview of expert rollouts, offline policy training, and real-time deployment

Expert long-horizon NMPC rollouts are distilled into a lightweight imitation policy for real-time deployment on differential-drive robots.

Abstract

Nonlinear Model Predictive Control (NMPC) achieves high-accuracy trajectory tracking for mobile robots but is often too computationally demanding for resource-constrained platforms. We present a learning-based predictive tracking controller that imitates an expert NMPC policy to achieve real-time performance on differential-drive robots. Our approach combines (a) state-perturbation augmentation with relative-error states to enrich the expert dataset and (b) a lightweight neural architecture based on 1-D convolutions with residual blocks and a physics-aware loss informed by robot kinematics. Expert state-action pairs are collected in a high-fidelity, physics-enabled simulator. Finally, we evaluate the resulting policy in a real-time experiment using a Quanser QBOT differential-drive platform, demonstrating tracking accuracy comparable to NMPC while significantly reducing computation time.

Video

Real-Time Results

Real-time experiment results on the Quanser QBOT platform — Real-time QBOT experiment: performed trajectory, control inputs, and tracking errors.

Position Tracking Error on QBOT Platform
Method	MSE [m²]	ISE [m²]	IAE [m]
LMPC N=7	0.1058	8.4829	24.6028
NMPC N=2	0.0506	2.4376	10.0803
IL N=20	0.0169	0.8133	4.3281

Onboard Computation Time
Method	Average [s]	Maximum [s]
LMPC N=7	0.0159	0.1050
NMPC N=2	0.0163	0.0356
IL N=20 CPU	0.0013953	0.001955
IL N=20 GPU	0.0009955	0.001394

Funding

This work was supported in part by the Natural Sciences and Engineering Research Council of Canada (NSERC), by Fonds de recherche du Québec (FRQNT), and by the National Cybersecurity Consortium (NCC).

BibTeX

@inproceedings{rajkumar2026ilnmpc,
  title = {A Real-Time Affordable Predictive Trajectory Tracking Controller for Differential Drive Robots via Imitation Learning},
  author = {Rajkumar, Suryaprakash and Lucia, Walter},
  booktitle = {Proceedings of the IEEE International Conference on Automation Science and Engineering (CASE)},
  year = {2026}
}