Morpho Evolution With Learning Using a Controller Archive as an Inheritance Mechanism

Morpho Evolution With Learning Using a Controller Archive as an Inheritance Mechanism
复制标题

使用控制器存档作为继承机制进行学习的 Morpho 进化

DOI:
10.1109/tcds.2022.3148543
复制
发表时间:
2023
影响因子:
5
通讯作者:
Le Goff L
Le Goff L
中科院分区:
计算机科学3区
文献类型:
--
作者:
Le Goff L

文献摘要

参考文献

被引文献

相似文献

进化机器人的大部分工作集中在为固定的身体计划进化控制器上。然而,以前的研究表明,同时发展控制器和身体计划可以打开许多有趣的可能性。然而,通过进化过程的身体规划和控制的联合优化可能是具有挑战性的丰富的形态空间。这是因为后代的身体结构可能与父母中的任何一个都有很大的不同,导致遗传神经控制器的结构与新身体之间的潜在不匹配。为了解决这个问题,我们提出了一个框架,结合了进化算法来生成身体计划和学习算法来优化神经控制器的参数。一旦生成每个后代的身体平面图,就创建该控制器的拓扑。该方法的关键新奇在于添加了一个外部存档,用于存储映射到机器人显式“类型”的学习控制器(其中这是相对于身体计划的特征定义的)。通过从一个控制器开始学习,该控制器具有从档案中继承的适当结构,而不是从随机初始化的结构,我们表明,与从头开始的方法相比,学习的速度和幅度都随着时间的推移而增加,使用两个任务和三个环境。该框架还为进化和学习之间的复杂相互作用提供了新的见解。
Most work in evolutionary robotics centers on evolving a controller for a fixed body plan. However, previous studies suggest that simultaneously evolving both controller and body plan could open up many interesting possibilities. However, the joint optimization of body plan and control via evolutionary processes can be challenging in rich morphological spaces. This is because offspring can have body plans that are very different from either of their parents, leading to a potential mismatch between the structure of an inherited neural 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 optimize the parameters of a neural controller. The topology of this controller is created once the body plan of each offspring has been 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 to the features of the body plan). By initiating learning from a controller with an appropriate structure inherited from the archive, rather than from a randomly initialized one, we show that both the speed and magnitude of learning increase 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.
DOI: 10.3389/frobt.2019.00009
发表时间: 2019
影响因子: 3.4
作者:
Jelisavcic M;Glette K;Haasdijk E;Eiben AE
通讯作者: Eiben AE
通过现实世界进化实现机器人形态和控制的环境适应
DOI: 10.1162/evco_a_00291
发表时间: 2020
影响因子: 6.8
作者:
T. Nygaard;Charles Patrick Martin;David Howard;J. Tørresen;K. Glette
通讯作者: K. Glette
DOI: --
发表时间: 2016-04
期刊: ArXiv
影响因子: --
作者:
N. Hansen
通讯作者: N. Hansen
使用两阶段方法克服机器人控制和形态学多目标进化中的初始收敛
DOI: --
发表时间: 2017
期刊: EvoApplications
影响因子: --
作者:
T. Nygaard;Eivind Samuelsen;K. Glette
通讯作者: K. Glette
现实世界的进化使机器人形态和控制适应硬件限制
DOI: 10.1145/3205455.3205567
发表时间: 2018
期刊: Proceedings of the Genetic and Evolutionary Computation Conference
影响因子: --
作者:
T. Nygaard;C. P. Martin;Eivind Samuelsen;J. Tørresen;K. Glette
通讯作者: K. Glette