Predictive Admittance Control for Aerial Physical Interaction
Ayham Alharbat, Chiara Gabellieri, Abeje Yenehun Mersha, Antonio Franchi
AI summary
Problem
Existing admittance control methods for aerial robots react only to instantaneous forces, failing to anticipate future impedance states, which causes poor tracking, instability, and unsafe behavior during physical contact.
Approach
The method embeds a dynamic model of the desired mechanical impedance directly into a Nonlinear Model Predictive Control (NMPC) framework, allowing the controller to jointly predict and optimize both the robot's motion and its impedance response over a prediction horizon while respecting actuator constraints.
Key results
- Up to 90% reduction in impedance tracking error versus state-of-the-art reactive methods
- Elimination of instability and reduced oscillations during physical contact
- Seamless transition between free-flight trajectory tracking and compliant interaction
- Real-time validation on a fully-actuated hexarotor with hard actuator constraints
Why it matters
Enables safer, more precise aerial manipulation for human-robot collaboration and infrastructure inspection in unstructured environments.
Abstract
This paper introduces a novel approach for con- trolling aerial robots during physical interaction by integrating Admittance Control with Nonlinear Model Predictive Control (NMPC). Unlike existing methods, our technique incorporates the desired impedance dynamics directly into the NMPC prediction model, alongside the robot’s dynamics. This allows for the explicit prediction of how the robot’s impedance will respond to interac- tion forces within the prediction horizon. Consequently, our con- troller effectively tracks the desired impedance behavior during physical interaction while seamlessly transitioning to trajectory tracking in free motion, all while consistently respecting actuator constraints. The efficacy of this method is validated through real- time simulations and experiments involving physical interaction tasks with an aerial robot. Our findings demonstrate that, across most scenarios, our method significantly outperforms the state-of- the-art (which does not predict future impedance state), achieving a reduction in tracking error of up to 90%. Furthermore, the results indicate that our approach enables smoother and safer physical interaction, characterized by reduced oscillations and the absence of the unstable behavior observed with the state-of- the-art method in certain situations.