Background
A new family of dynamic propulsion-power models has recently been developed for multirotor propulsion. [1]
These models progressively represent aerodynamic dissipation, reversible kinetic-energy exchange, torque-dependent electromechanical dissipation, and speed-proportional losses. Unlike speed-only models, they can describe how the electrical power of a propulsion unit changes during both steady and accelerating operation.
The models have been experimentally validated on a single motor-drive-propeller unit. Their validity for a complete multirotor remains unknown.
In a multirotor, several propulsion units share the same electrical bus and battery. Their combined current demand can approach the limits of the battery, power distribution system, and electronic speed controllers. Additional electrical interactions and dissipation effects could also appear at the system level.
Objective
The objective of this thesis is to extend dynamic propulsion-power modelling from a single propulsion unit to a complete multirotor and use the resulting model for power-aware control.
The work will investigate both conventional and redundantly actuated multirotors. It will combine system modelling, experimental identification, optimization-based control, and flight validation.
Main activities
- Determine whether the existing single-rotor power model accurately predicts the total electrical power of a multirotor.
- Design experiments in which several propulsion units operate simultaneously from a shared battery and electrical bus.
- Measure propeller speeds, motor commands, battery voltage, individual or aggregate currents, and total electrical power.
- Investigate whether additional terms are needed to represent battery voltage variations, shared-bus effects, electronic-speed-controller losses, or other system-level dissipation.
- Identify the resulting multirotor power model and validate it using experimental data not employed for identification.
- Formulate power-aware control and control allocation subject to motor, electronic-speed-controller, bus-current, and battery constraints.
- Develop an optimization-based controller, such as model predictive control, that distributes the requested actuation among the propulsion units.
- Implement and validate the controller on a real multirotor.
- Compare the proposed method with conventional control and control-allocation strategies.
Expected outcomes
- An experimentally validated dynamic power model for a complete multirotor.
- An assessment of when independently combining single-rotor models is sufficient and when system-level effects must be included.
- Identification of the relevant propulsion-unit, bus-current, and battery constraints.
- A power-aware controller or control-allocation method for conventional or redundantly actuated multirotors.
- Flight experiments evaluating whether the proposed method reduces electrical-energy consumption for a prescribed task.
- Experiments evaluating whether explicit power and current management improves trajectory tracking or wrench execution during demanding flight or physical-interaction tasks.
Required or preferred skills
- A solid background in electronics, dynamical systems, control, and optimization.
- Experience with MATLAB/Simulink, Python, C++, or ROS.
- Interest in multirotor propulsion, electrical-power measurement, system identification, and experimental robotics.
- Previous experience with model predictive control, control allocation, or embedded data acquisition is beneficial.
Supervisors
- Antonio Franchi
- Ahmed Ali
- Chiara Gabellieri
Reference
- [1] Nested Power Models for Multirotor Propulsion: From Aerodynamic Drag to Electrical Losses, A. Franchi, A. Saini, A. Ali, and C. Gabellieri, arXiv:2610.08177. Preprint.