Motion cueing study finds 5.4% just-noticeable difference in simulator acceleration

A team led by Erik Gustaf Lilljebjörn at Linköping University investigated motion cueing algorithms (MCAs) for driving simulators used in heavy-truck driveline development. The study, published on arXiv (2606.28055), aimed to determine the smallest perceivable difference in longitudinal acceleration—the just-noticeable difference (JND)—when using two MCA variants tuned for tip-in/launch tests versus a general variant. Participants experienced acceleration profiles on a long linear track simulator and provided responses via a weighted staircase procedure. Psychometric functions and generalized linear models revealed no statistically significant differences in JND between the three MCA variants, with an overall mean JND of 5.4%. Additionally, the mean point of subjective equality was -1.9%, indicating participants perceived the second stimulus in a pair as having higher acceleration. Subjectively, most participants favored the launch-tuned MCAs, suggesting that specialized motion cueing can improve simulator realism for specific driving maneuvers.

This research directly impacts the heavy-truck industry by validating that driving simulators with motion platforms can reliably replicate small acceleration changes critical for driveline assessment. The 5.4% JND provides a benchmark for simulator fidelity, while the preference for tuned MCAs supports the use of virtual prototypes over physical prototypes in early development phases. By enabling human-in-the-loop evaluation of virtual drivelines earlier, manufacturers can reduce costs and iteration cycles. The study also contributes to broader knowledge on human perception in dynamic environments, with implications for any motion-based simulation where accurate force feedback is essential.