# A Staged Approach to Evolving Real-world UAV Controllers

## TL;DR

This paper presents a two-stage evolutionary approach for developing more general UAV controllers by initially evolving controllers with physical constraints and then refining them in a less restricted environment, improving generalization and efficiency.

## Contribution

The paper introduces a novel two-stage evolution method that enhances controller generalization for real-world UAVs by combining constrained and less restricted optimization phases.

## Key findings

- Two-stage approach yields more general controllers in fewer generations.
- Controllers evolved with the method perform better in real-world hover tasks.
- Sensitivity analysis confirms robustness of the evolved controllers.

## Abstract

A testbed has recently been introduced that evolves controllers for arbitrary hover-capable UAVs, with evaluations occurring directly on the robot. To prepare the testbed for real-world deployment, we investigate the effects of state-space limitations brought about by physical tethering (which prevents damage to the UAV during stochastic tuning), on the generality of the evolved controllers. We identify generalisation issues in some controllers, and propose an improved method that comprises two stages: in the first stage, controllers are evolved as normal using standard tethers, but experiments are terminated when the population displays basic flight competency. Optimisation then continues on a much less restrictive tether, effectively free-flying, and is allowed to explore a larger state-space envelope. We compare the two methods on a hover task using a real UAV, and show that more general solutions are generated in fewer generations using the two-stage approach. A secondary experiment undertakes a sensitivity analysis of the evolved controllers.