Date of Completion

2025

Document Type

Honors College Thesis

Department

Computer Science

Thesis Type

Honors College

First Advisor

Joshua Bongard

Second Advisor

James Eddy

Keywords

evolutionary robotics, carcinization, blender, crab, simulation

Abstract

Carcinization, the phenomenon of crustaceans independently evolving crab-like body structures, has been widely observed in nature. Since all common ancestors of modern crustaceans are extinct, modern study of carcinization invites the usage of evolutionary algorithms to determine the factors of carcinization. While the causes of carcinization are hypothesized, a prominent theory invokes the direction of locomotion as a factor. Using evolutionary robotics, we test whether selection for lateral locomotion drives convergent evolution toward crab-like morphologies in silico. We performed the experiment utilizing a lobster-like robot, under a hill climber evolutionary algorithm to select through 10,000 specimens for 30 trials under earth-like open-air conditions. Robot morphologies selected mutations which positively affected lateral locomotion. By quantifying trait thresholds (torso width, tail reduction, leg splay), the level of carcinization was assessed. The robots were then compared against robots evolved to maximize longitudinal locomotion. Our findings show that evolved robots maximizing lateral locomotion had some degree of carcinization, with the majority of robots emerging fully carcinized. Robots maximizing longitudinal locomotion did not exhibit traits of carcinization.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.

Download

Share

COinS