Neuroflight: Next Generation Flight Control Firmware

TL;DR

Neuroflight introduces an open source neural network-based flight controller for UAVs, demonstrating stable flight and aerobatic capabilities on embedded hardware, surpassing traditional PID controllers.

Contribution

It is the first open source neuro-flight controller firmware with a complete toolchain for training and deployment on embedded systems.

Findings

Neural network runs at over 2.67kHz on Cortex-M7

Quadcopter achieves stable flight and aerobatic maneuvers

Addresses simulation-to-reality transfer challenges

Abstract

Little innovation has been made to low-level attitude flight control used by uncrewed aerial vehicles (UAVs), which still predominantly uses the classical PID controller. In this work we introduce Neuroflight, the first open source neuro-flight controller firmware. We present our toolchain for training a neural network in simulation and compiling it to run on embedded hardware. Challenges faced jumping from simulation to reality are discussed along with our solutions. Our evaluation shows the neural network can execute at over 2.67kHz on an Arm Cortex-M7 processor and flight tests demonstrate a quadcopter running Neuroflight can achieve stable flight and execute aerobatic maneuvers.