Future reference prediction in model predictive control based driving simulators
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the limitations of Model Predictive Control (MPC) in driving simulator motion cueing algorithms (MCA). While MPC is effective for managing simulator workspace constraints, traditional implementations assume that future input references (e.g., acceleration) remain constant over the prediction horizon. This assumption is unrealistic for dynamic driving scenarios and leads to conservative control behavior, resulting in higher sensation errors and inefficient use of the simulator’s physical workspace. The authors propose a method to predict future reference signals based on finite input history, thereby improving motion fidelity and reducing motion sickness risks associated with false motion cues. The proposed method integrates an Artificial Neural Network (ANN) with the MPC framework. The ANN is trained to predict future acceleration references using a structured representation of past input data. The input history and future horizon are divided into five regions with varying resolutions, where regions closer to the current time have higher resolution. The average values of these regions form quintuples that serve as inputs and targets for the neural network training. The network architecture consists of an input layer with five neurons, a hidden layer with 36 neurons, and an output layer with five neurons. The model was trained using a dataset of 49,262 input samples. In the simulation, the MPC controller utilizes a control horizon of 3, a prediction horizon of 400, and a sampling time of 10 ms. The ANN provides updated future reference estimates at each sampling time, which are interpolated for intermediate points. Simulation results demonstrate that the proposed method significantly outperforms traditional MPC with constant future reference assumptions. The Root Mean Square (RMS) of the sensation error, which measures the discrepancy between the simulated motion and the ideal real-world sensation, decreased from 1.29 m/s² to 0.85 m/s². Additionally, the proposed method utilized the simulator’s workspace more efficiently. The maximum displacement of the motion platform increased from 1.35 m to 2.78 m, approaching the ±3 m physical limit. This increased utilization allowed the system to generate more realistic motion cues without violating constraints, whereas the traditional method remained conservative due to its inability to accurately predict transient acceleration peaks. The significance of this work lies in enhancing the realism and comfort of driving simulators. By accurately predicting future motion references, the proposed MPC-MCA reduces sensation errors and prevents the conservative behavior typical of standard MPC approaches. This leads to better motion cueing fidelity, which is critical for applications such as driver training, vehicle prototyping, and behavioral studies. The method does not require drivers to follow specific predefined routes, making it adaptable to various driving scenarios. The findings suggest that integrating predictive models like ANNs into control systems can substantially improve performance in constrained dynamic environments.
Key finding
Integrating an Artificial Neural Network to predict future reference inputs in Model Predictive Control significantly reduces sensation error and improves motion cueing realism in driving simulators compared to methods assuming constant future inputs.
Methodology
simulation_modeling
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-28 |
| archive | success | canonical_url | — | — | 1 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-28 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Methodological Resource: tool software
- Theoretical Contribution: computational model