Abstract
This study explores the influence of material selection scenarios on the evolutionary design of soft robots. Soft robots have vast potential applications that rigid robots cannot accomplish due to their lightweight and flexible nature. Despite rapid development of soft robotics in recent years, its effective automatic design method remains undeveloped. To combat this, evolutionary computation has been proposed for automatic soft robot design. Material selection has been identified as crucial in influencing the outcomes of autonomously designed soft robots, yet understanding of its impacts remains limited, particularly in changing material selection scenarios across generations. In this study, we hypothesized that these changing scenarios could influence the evolved robots. Initial experiments, based on travel distance and energy efficiency, suggested material selection could greatly impact the final robot. Further experiments, adjusting material ratio and rewarding specific material use, identified how certain material selection scenarios led to changes in the structure and performance of the evolved soft robots. These findings contribute to automatic design methodologies in soft robotics.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
The number of generations was set at 10,000 to almost catch the trend of evolution. This number of generations was set in consideration of the balance with the required computational power.
References
Bhatia J, Jackson H, Tian Y, Xu J, Matusik W (2021) Evolution gym: a large-scale benchmark for evolving soft robots. Adv Neural Inf Process Syst 34:2201–2214
Cheney N, MacCurdy R, Clune J, Lipson H (2014) Unshackling evolution: evolving soft robots with multiple materials and a powerful generative encoding. ACM SIGEVOl 7(1):11–23
Chervenski P, Ryan S (2014) Multineat. https://github.com/peter-ch/MultiNEAT
Corucci F, Cheney N, Giorgio-Serchi F, Bongard J, Laschi C (2018) Evolving soft locomotion in aquatic and terrestrial environments: effects of material properties and environmental transitions. Soft Robot 5(4):475–495
Google Inc (2017) Liquidfun. http://google.github.io/liquidfun/
Ha S, Coros S, Alspach A, Bern JM, Kim J, Yamane K (2018) Computational design of robotic devices from high-level motion specifications. IEEE Trans Robot 34(5):1240–1251
Hiller J, Lipson H (2011) Automatic design and manufacture of soft robots. IEEE Trans Robot 28(2):457–466
Hiller JD, Lipson H (2010) Evolving amorphous robots. In: Alife, Citeseer, pp 717–724
Joachimczak M, Suzuki R, Arita T (2014) Fine grained artificial development for body-controller coevolution of soft-bodied animats. In: ALIFE 14: the fourteenth international conference on the synthesis and simulation of living systems, MIT Press, pp 239–246
Joachimczak M, Kaur R, Suzuki R, Arita T (2015) Bringing drawings to life: evolving distributed controllers for hand-drawn soft-bodied robots. In: Proceedings of SWARM 2015: the first international symposium on swarm behavior and bio-inspired robotics (AROB 2021), pp 381–388
Kriegman S, Blackiston D, Levin M, Bongard J (2020) A scalable pipeline for designing reconfigurable organisms. Proc Natl Acad Sci 117(4):1853–1859
Nakamura Y, Suzuki R, Arita T (2021) Evolutionary design of soft robots that achieve energy-efficient locomotion based on co-evolution of morphology, materials, and control. In: Proceedings of the 26th international symposium on artificial life and robotics (AROB 2021), pp 84–89
Pfeifer R, Bongard J (2006) How the body shapes the way we think: a new view of intelligence. MIT Press, Cambridge
Pinskier J, Howard D (2022) From bioinspiration to computer generation: developments in autonomous soft robot design. Adv Intell Syst 4(1):2100086
Sims K (1994) Evolving virtual creatures. In: Proceedings of the 21st annual conference on computer graphics and interactive techniques, pp 15–22
Stanley KO (2007) Compositional pattern producing networks: a novel abstraction of development. Genet Program Evol Mach 8(2):131–162
Stanley KO, Miikkulainen R (2002) Evolving neural networks through augmenting topologies. Evol Comput 10(2):99–127
Wang L, Joachimczak M, Suzuki R, Arita T (2018) Evolving morphology and control for virtual soft-robots with a particle-based soft-body simulation. In: Proceedings of the 23rd international symposium on artificial life and robotics (AROB 2018), pp 206–211
Wang T, Zhou Y, Fidler S, Ba J (2019) Neural graph evolution: automatic robot design. In: International conference on learning representations. https://openreview.net/forum?id=BkgWHnR5tm
Zhao A, Xu J, Konaković-Luković M, Hughes J, Spielberg A, Rus D, Matusik W (2020) Robogrammar: graph grammar for terrain-optimized robot design. ACM Trans Graph (TOG) 39(6):1–16
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was presented in part at the joint symposium of the 28th International Symposium on Artificial Life and Robotics, the 8th International Symposium on BioComplexity, and the 6th International Symposium on Swarm Behavior and Bio-Inspired Robotics (Beppu, Oita and Online, January 25–27, 2023).
Appendix: Results without sliding-type robots
Appendix: Results without sliding-type robots
Here are the results of each experiment without the sliding type. The final generation’s distance traveled without a bonus was 0.593, and with a bonus it was 0.438, while the fitness was 0.684 in Experiment 1. The final generation’s distance traveled and fitness for Experiment 2 are shown in Table 3. Common to both experiments, the overall distance traveled decreased due to the absence of the sliding type, which was more likely to achieve greater distances. Concerning Experiment 1, the difference in distance traveled with and without a bonus became even more significant. Thus, it can be said that the trade-off relationship between energy efficiency improvement by handling inert materials and distance traveled is more stringent. For Experiment 2, the overall trend remained consistent except that the ranks of A–P and P–A switched places when comparing the distances traveled between cases. Consequently, the discussions in Sect. 4.3 remain valid even excluding the sliding type.
Regarding the material proportion, this is shown in Figs. 10 and 11. Just like in Sect. 4.2, we can observe that the constitution changes to increase the material that was granted a bonus due to bonus alterations. When excluding sliding, the usage rate of active is higher overall compared to when it is included. Also, the change in constitution due to bonus alterations results in a larger shift toward active. For instance, when sliding is excluded, the increase in active after the bonus change is notable in B–A. In B–P, both bone and active decrease with the increase in passive after the bonus change with sliding included, while the decrease is almost entirely in bone when sliding is excluded.
In summary, in this model, while the sliding type, a soft robot that efficiently uses this model’s unique specifications, extended the distance traveled while also utilizing inert materials, generally speaking, the importance of active, i.e., the actuator, in movement is higher. If one wishes to utilize inert materials, the trade-off will be of greater importance.
Average of the proportion of material comprising the best adapted individuals in each generation (excluding sliding type) in Experiment 1
Average of the proportion of material comprising the best adapted individuals in each generation (excluding sliding type) in Experiment 2
About this article
Cite this article
Shimaoka, N., Suzuki, R. & Arita, T. Investigating the effects of material ratio scenarios on soft robot design based on morphology–material–control coevolution. Artif Life Robotics (2024). https://doi.org/10.1007/s10015-023-00934-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10015-023-00934-3