Visual stabilization with sliding mode control outperforms PID only under actuator saturation
What happened
A paper submitted by Sooyeun Yang and other authors points out that the advantage of boundary-layer sliding mode control (SMC) over causal baselines such as LPF and PID in visual stabilization has a clear activation condition: it holds only when the disturbance speed approaches or exceeds the actuator's rotational speed limit and the per-step actuation budget is constrained (i.e., plant binding). This condition was reached by sweeping simulated rotational speed limits; both collected datasets themselves fall outside this condition. Under plant binding, translation error is on par with the baselines, and SMC outperforms PID in the worst-case angular condition, with Cohen's d up to +0.83. The authors state that the PTZ device's default 100 degrees/second happens to sit right at the crossover point, which explains the earlier result in which PID held a slight lead in the angular setting. Ablation experiments also show that increasing the SMC reaching gain k degrades average tracking performance outside a narrow window.
AI 根据报道生成 · 6 days ago更新
Reporting timeline
Follow the reporting to explore different perspectives.
- arXiv cs.RORobustness verification of visual stabilization under plant-binding conditions using sliding mode control
This study examines the visual stabilization performance of boundary-layer sliding mode control (SMC) relative to causal baselines such as LPF and PID under plant-binding conditions, where disturbance velocity approaches or exceeds the actuator slew limit and the per-step actuation budget is constrained.
Activity for this event
- Current
- 3
- Peak
- 10Nov 5 14:00
- Past 24 hours
- -49%
The trend compares participants observed throughout the same period, so its range may be smaller than the current interest count. Move or tap to inspect each hour; use the arrow keys on a keyboard.