Morpho-evolution with learning using a controller archive as an inheritance mechanism

TL;DR

This paper introduces a framework combining evolution and learning with a controller archive to improve the inheritance of neural controllers for diverse robot morphologies, enhancing learning efficiency and providing insights into evolution-learning interactions.

Contribution

It presents a novel approach that uses an external archive of learned controllers to better inherit control structures, improving learning speed and effectiveness in evolving diverse robot morphologies.

Findings

Learning speed and magnitude increase over time with the archive.

The approach outperforms random initialization in two tasks and three environments.

Provides new insights into evolution and learning interactions.

Abstract

The joint optimisation of body-plan and control via evolutionary processes can be challenging in rich morphological spaces in which offspring can have body-plans that are very different from either of their parents. This causes a potential mismatch between the structure of an inherited controller and the new body. To address this, we propose a framework that combines an evolutionary algorithm to generate body-plans and a learning algorithm to optimise the parameters of a neural controller. The topology of this controller is created once the body-plan of each offspring body-plan is generated. The key novelty of the approach is to add an external archive for storing learned controllers that map to explicit `types' of robots (where this is defined with respect the features of the body-plan). By learning from a controller with an appropriate structure inherited from the archive, rather than from a randomly initialised one, we show that both the speed and magnitude of learning increases over time when compared to an approach that starts from scratch, using two tasks and three environments. The framework also provides new insights into the complex interactions between evolution and learning.