Battery-Aware Reinforcement Learning for Aggressive Quadrotor Flight

2026-09-29

AI Summary

This study addresses the thrust degradation caused by battery discharge, which constrains the aggressive flight performance of unmanned aerial vehicles. To overcome this limitation, we propose a voltage-aware deep reinforcement learning framework that, for the first time, jointly trains a load-transient battery model with firmware-level actuator saturation. By employing a voltage feedback feedforward strategy, the approach exploits residual thrust potential while preserving low-level control authority. We demonstrate the critical role of voltage input in enhancing maneuver precision and achieving efficient sim-to-real transfer. Hardware experiments reveal that, compared to a voltage-agnostic baseline, the proposed method reduces trajectory tracking error by 49% and shortens racing lap time by 10.3%, significantly advancing extreme flight performance.

Abstract

Agile flight tasks such as drone racing and pursuit-evasion require strong acceleration and precise turns, but the available thrust changes as the battery discharges and voltage drops under load. Conservative command limits make this variation easier to tolerate, at the cost of unused performance. We investigate how learned controllers can use that additional thrust while retaining the flight controller's voltage compensation and rate control. Our training simulator couples an identified load-transient battery model to rotor dynamics and firmware saturation. The feedforward policy receives filtered voltage during both training and deployment. Controlled ablations distinguish the benefit of a larger thrust-command range from that of voltage information. On a 38 g Crazyflie Brushless, the resulting policy reduces circle tracking error by 49% relative to stock-authority RL at 3.84 m/s, while preserving easy-task precision. Mean 20-lap race time decreases from 106.22 s to 95.24 s. Compared with a voltage-blind policy with the same increased authority, hardware error and race time are lower by 15.3% and 4.5%, respectively. In simulation, replacing the policy's voltage input with a recording from a different battery condition worsens hard-circle tracking, with a smaller, voltage-dependent effect in racing. Together, these results show where a simple voltage input complements existing actuator compensation in aggressive learned flight.

Problem

• agile quadrotor flight
• battery voltage drop
• thrust authority
• reinforcement learning
• drone racing

Innovation

• Battery-aware reinforcement learning
• Aggressive quadrotor flight
• Load-transient battery model
• Voltage compensation
• Feedforward policy