Comparing lifetime learning methods for morphologically evolving robots

TL;DR

This paper compares three lifetime learning algorithms for morphologically evolving robots, aiming to improve the adaptation process and mitigate body-brain mismatches caused by stochastic recombination.

Contribution

It introduces a comparative analysis of three algorithms focusing on efficiency, efficacy, and sensitivity in the context of evolving robot morphologies and controllers.

Findings

One algorithm showed higher efficiency in quick adaptation.

Another demonstrated greater efficacy in optimizing robot performance.

Sensitivity analysis revealed differences in robustness to morphological variations.

Abstract

Evolving morphologies and controllers of robots simultaneously leads to a problem: Even if the parents have well-matching bodies and brains, the stochastic recombination can break this match and cause a body-brain mismatch in their offspring. We argue that this can be mitigated by having newborn robots perform a learning process that optimizes their inherited brain quickly after birth. We compare three different algorithms for doing this. To this end, we consider three algorithmic properties, efficiency, efficacy, and the sensitivity to differences in the morphologies of the robots that run the learning process.