ResearchResearch paperRobotics & Embodied AI · Reinforcement Learning1 source · Oct 6, 2026

Nested Power Models for Multirotor Propulsion: From Aerodynamic Drag to Electrical Losses

This work develops a nested sequence of propulsion-power models that starts from aerodynamic power dissipation and progressively introduces a reversible kinetic-energy rate, torque-dependent electromechanical dissipation, and lumped speed-proportional dissipation.

Key points

  • Speed-only aerodynamic power models for multirotor propulsion cannot represent acceleration-dependent effects.
  • Independent estimates of rotational inertia and aerodynamic drag complement predictive validation by assessing whether the models correctly attribute the measured power to reversible kinetic-energy exchange and irreversible dissipation and, within the latter, to aerodynamic and electromechanical losses.
  • The results show that the reversible kinetic-energy rate is necessary but insufficient for accurate dynamic power prediction.
  • The resulting methodology provides a reusable and experimentally verifiable basis for developing and selecting dynamic propulsion-power models for multirotor systems.

Sources (1)

  • [1]Nested Power Models for Multirotor Propulsion: From Aerodynamic Drag to Electrical Losses
    arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 11:29 AM
    This work develops a nested sequence of propulsion-power models that starts from aerodynamic power dissipation and progressively introduces a reversible kinetic-energy rate, torque-dependent electromechanical dissipation, and lumped speed-proportional dissipation.
    Speed-only aerodynamic power models for multirotor propulsion cannot represent acceleration-dependent effects.

Extractive summary: sentences quoted from the sources.

Related