Mon. Sep 25th, 2023
    Designing Hardware for Evolvable Robots: Challenges and Considerations

    This article explores the challenges and considerations in designing hardware for evolvable robots. The researchers from the University of York, Edinburgh Napier University, Vrije Universiteit Amsterdam, University of the West of England, and University of Sunderland highlight the importance of implementing evolved robots in hardware for real-world applications.

    The field of evolutionary robotics focuses on generating robots that adapt to specific tasks by optimizing across successive generations of robots. While most evolutionary robotics work is conducted in simulation due to the time and cost constraints of physical construction, implementing evolved robots in hardware presents new challenges.

    The researchers delve into the design of a system for realizing diverse evolved robot bodies and its interaction with the evolutionary process. They uncover that the hardware implementation introduces constraints that alter the evolutionary space. Factors such as manufacturing and assembly constraints, as well as electrical power delivery limitations, pose hurdles in building certain robot configurations or result in malfunctioning.

    To tackle this challenge, the researchers introduce methods of phenotype filtering and repair. However, these methods are found to degrade the diversity of the robot population and impede exploration of the evolutionary space. Moreover, the hardware constraints were poorly matched with the types of morphological variation that would be most beneficial in the target environment, leading to a reduced ability of the evolutionary process to generate robots with effective adaptations.

    The conclusions drawn from this research emphasize two key points. Firstly, designing a hardware platform for evolving robots requires careful consideration of its impact on the viable phenotype space. Secondly, solely evolving robots in simulation without accounting for hardware implementation details is insufficient, as hardware constraints significantly influence the evolutionary space.

    In summary, this study sheds light on the intricate interplay between hardware constraints and evolution in the context of evolvable robots. It emphasizes the necessity of designing hardware platforms with the viable phenotype space in mind and highlights the importance of considering hardware implementation early on in the evolutionary process.

    Sources:
    Angus M, Buchanan E, Le Goff LK, Hart E, Eiben AE, De Carlo M, Winfield AF, Hale MF, Woolley R, Timmis J, and Tyrrell AM (2023) Practical hardware for evolvable robots. Front. Robot. AI 10:1206055. doi: 10.3389/frobt.2023.1206055