Learning-Based Approximate Nonlinear Model Predictive Control Motion Cueing

TL;DR

The paper tackles the computational bottleneck of nonlinear model predictive control-based motion cueing by introducing a learning-based MCA built on differentiable predictive control. It combines a learnable nonlinear joint-space plant with a policy network trained offline to mimic NMPC, enabling real-time inference without horizon shortening or downsampling. Key findings show motion cueing quality on par with exact NMPC while achieving around a 400x reduction in computation time, and demonstrated generalization to unseen vehicles and physics-based simulations. This approach offers a scalable path to high-rate, constraint-aware motion cueing in serial robot-based simulators and opens avenues for disturbance handling and gradient-based controller tuning.

Abstract

Motion Cueing Algorithms (MCAs) encode the movement of simulated vehicles into movement that can be reproduced with a motion simulator to provide a realistic driving experience within the capabilities of the machine. This paper introduces a novel learning-based MCA for serial robot-based motion simulators. Building on the differentiable predictive control framework, the proposed method merges the advantages of Nonlinear Model Predictive Control (NMPC) - notably nonlinear constraint handling and accurate kinematic modeling - with the computational efficiency of machine learning. By shifting the computational burden to offline training, the new algorithm enables real-time operation at high control rates, thus overcoming the key challenge associated with NMPC-based motion cueing. The proposed MCA incorporates a nonlinear joint-space plant model and a policy network trained to mimic NMPC behavior while accounting for joint acceleration, velocity, and position limits. Simulation experiments across multiple motion cueing scenarios showed that the proposed algorithm performed on par with a state-of-the-art NMPC-based alternative in terms of motion cueing quality as quantified by the RMSE and correlation coefficient with respect to reference signals. However, the proposed algorithm was on average 400 times faster than the NMPC baseline. In addition, the algorithm successfully generalized to unseen operating conditions, including motion cueing scenarios on a different vehicle and real-time physics-based simulations.

Figures (5)

• Figure 1: MuJoCo model of motion simulator based on the KUKA KR500 robot arm. Bottom images show the reference frames relevant to motion cueing.
• Figure 2: Architecture and training setup for the DPC-based MCA.
• Figure 3: Results from simulation experiments controlling analytical plant model. Box plots show the performance index distribution achieved by each MCA across the 10 scenarios considered for each aircraft.
• Figure 4: Target and achieved inertial signal profiles for 10 minute simulation in the real-time MuJoCo environment. The target profiles are the references for the MCA -- these references are scaled and gravity compensated versions of the IMU measurements gathered aboard the aircraft.
• Figure 5: Swarm charts for joint positions, velocities and accelerations achieved during motion cueing simulation in Figure \ref{['fig:mujoco_experiments']}.