Directed locomotion learning for evolving modular robots of arbitrary morphologies

Neuroevolution is a novel and important task in Artificial intelligence (AI), which is an evolutionary approach. The goal of my research is to implement the directed locomotion learning for evolving modular robots of arbitrary morphologies by neuroevolution method. First, we use computer vision to distinguish the direction towards objects. The challenges are 1) the pose is constantly changing; 2) the algorithm has to work on low performance embedded system. Second, Neuroevolution evolving the different controller for robots of arbitrary morphologies to implement the locomotion towards a directed of the object. This project aims to implement a evolved brain for robots of arbitrary with the features of recognizing direction and locomotion.

Team

Gongjin Lan
PhD Student

Gongjin Lan

Prof. dr. Guszti Eiben
Head of the Group

Prof. dr. Guszti Eiben