# Efficient Knowledge Transfer for Jump-Starting Control Policy Learning of Multirotors through Physics-Aware Neural Architectures

arXiv: https://arxiv.org/abs/2602.15533 · HTML: https://mlsanity.com/paper/2602.15533/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in RL for robot navigation and UAV control, one of 15 directions in Deep reinforcement learning methods.
- Builds on 23 of its 30 references that are in the map.
- First posted on arXiv 17 Feb 2026.

## Placed in

- Deep reinforcement learning methods / [RL for robot navigation and UAV control](https://mlsanity.com/deeprl/rl-for-robot-navigation-and-uav-control/)

## Builds on (most cited first, 23 of 23)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [On the Continuity of Rotation Representations in Neural Networks](https://mlsanity.com/paper/1812.07035/) (2018)
- [OptNet: Differentiable Optimization as a Layer in Neural Networks](https://mlsanity.com/paper/1703.00443/) (2017)
- [Statistical Aspects of Wasserstein Distances](https://mlsanity.com/paper/1806.05500/) (2018)
- [Embodied Intelligence via Learning and Evolution](https://mlsanity.com/paper/2102.02202/) (2021)
- [Jump-Start Reinforcement Learning](https://mlsanity.com/paper/2204.02372/) (2022)
- [Sample Factory: Egocentric 3D Control from Pixels at 100000 FPS with Asynchronous Reinforcement Learning](https://mlsanity.com/paper/2006.11751/) (2020)
- [Control of a Quadrotor with Reinforcement Learning](https://mlsanity.com/paper/1707.05110/) (2017)
- [Regularizing Action Policies for Smooth Control with Reinforcement Learning](https://mlsanity.com/paper/2012.06644/) (2020)
- [Data-efficient Co-Adaptation of Morphology and Behaviour with Deep Reinforcement Learning](https://mlsanity.com/paper/1911.06832/) (2019)
- [Mirage: Cross-Embodiment Zero-Shot Policy Transfer with Cross-Painting](https://mlsanity.com/paper/2402.19249/) (2024)
- [Learning to Fly in Seconds](https://mlsanity.com/paper/2311.13081/) (2023)
- [Cross-Embodiment Robot Manipulation Skill Transfer using Latent Space Alignment](https://mlsanity.com/paper/2406.01968/) (2024)
- [A Learning-based Quadcopter Controller with Extreme Adaptation](https://mlsanity.com/paper/2409.12949/) (2024)
- [Learning Minimum-Time Flight in Cluttered Environments](https://mlsanity.com/paper/2203.15052/) (2022)
- [Aerial Gym Simulator: A Framework for Highly Parallelized Simulation of Aerial Robots](https://mlsanity.com/paper/2503.01471/) (2025)
- [Building an Aerial-Ground Robotics System for Precision Farming: An Adaptable Solution](https://mlsanity.com/paper/1911.03098/) (2019)
- [REvolveR: Continuous Evolutionary Models for Robot-to-robot Policy Transfer](https://mlsanity.com/paper/2202.05244/) (2022)
- [Morpho-evolution with learning using a controller archive as an inheritance mechanism](https://mlsanity.com/paper/2104.04269/) (2021)
- [RAPTOR: A Foundation Policy for Quadrotor Control](https://mlsanity.com/paper/2509.11481/) (2025)
- [Pretraining-finetuning Framework for Efficient Co-design: A Case Study on Quadruped Robot Parkour](https://mlsanity.com/paper/2407.06770/) (2024)
- [Design and Control Co-Optimization for Automated Design Iteration of Dexterous Anthropomorphic Soft Robotic Hands](https://mlsanity.com/paper/2403.09933/) (2024)
- [Semantics-aware Predictive Inspection Path Planning](https://mlsanity.com/paper/2506.06560/) (2025)

## Built on by (most cited first, 1 of 1)

- [Embodiment-conditioned Generalist Control for Multirotor Aerial Robots](https://mlsanity.com/paper/2606.10857/) (2026)

## Abstract (arXiv)

Efficiently training control policies for robots is a major challenge that can greatly benefit from utilizing knowledge gained from training similar systems through cross-embodiment knowledge transfer. In this work, we focus on accelerating policy training using a library-based initialization scheme that enables effective knowledge transfer across multirotor configurations. By leveraging a physics-aware neural control architecture that combines a reinforcement learning-based controller and a supervised control allocation network, we enable the reuse of previously trained policies. To this end, we utilize a policy evaluation-based similarity measure that identifies suitable policies for initialization from a library. We demonstrate that this measure correlates with the reduction in environment interactions needed to reach target performance and is therefore suited for initialization. Extensive simulation and real-world experiments confirm that our control architecture achieves state-of-the-art control performance, and that our initialization scheme saves on average up to 73.5% of environment interactions (compared to training a policy from scratch) across diverse quadrotor and hexarotor designs, paving the way for efficient cross-embodiment transfer in reinforcement learning.