Generating Collective Motion Behaviour Libraries Using Developmental Evolution

This paper presents an evolutionary framework for generating diverse libraries of collective motion behaviours. It builds upon recent advancements in machine recognition of collective motion and the transformation of random motions into structured collective behaviours. The paper describes the design of the framework, including the use of a fitness function and diversity metrics specifically tailored for this purpose. The proposed framework generates diverse behaviours with distinct collective motion characteristics. Analysing the relationship between genotypic and behavioural diversity, we observed that greater diversity emerges after a moderate number of evolutionary generations. Our findings highlight the effectiveness of task non-specific fitness functions in distinguishing structured collective behaviours in an evolutionary setting.