Actuator saturation and command limits
Compare requested action with the effort the modeled actuator can actually apply.
Research note · November 14, 2026Research question
What happens when a controller asks for more force or torque than the actuator permits?
Method
- Define command limits and force or torque limits separately. Record any actuator dynamics or delays that sit between controller output and applied effort.
- Run a load sweep that crosses the chosen limit. Retain requested command, applied force and system response so saturation can be distinguished from a sensor or solver error.
- Test recovery after returning to a feasible load. Inspect controller memory and overshoot rather than assuming the system immediately returns to its earlier behavior.
Records to retain
- Command and effort limits
- Controller output
- Applied effort and state traces
- Saturation and recovery intervals
Source basis
MuJoCo distinguishes actuator control and force constraints and documents their interaction with actuator and joint behavior. A command value alone does not describe the force that reaches the modeled mechanism.
Interpretation & limits
Limits must come from an explicit model or source. A visually successful motion under unlimited actuation says little about a real device. No hardware performance claim is made.
Next experiments
- Use a bounded controller fixture with a reproducible saturation interval.
- Inspect recovery with and without controller-state reset.
References
Technical reference
Implementation details supporting this note, drawn directly from the maintained documentation. These are technical references, not additional papers or new experimental results. Expand a topic to inspect its method, fields and limits.