UB ScholarWorks

Autonomous Task-Based Evolutionary Design of Modular Robots

Show simple item record

dc.contributor.author Alattas, Reem
dc.date.accessioned 2019-03-04T18:07:32Z
dc.date.available 2019-03-04T18:07:32Z
dc.date.issued 2019-02-27
dc.identifier.citation R.J. Alattas, "Autonomous Task-Based Evolutionary Design of Modular Robots", Ph.D. dissertation, Dept. of Engineering, Univ. of Bridgeport, Bridgeport, CT, 2019. en_US
dc.identifier.uri https://scholarworks.bridgeport.edu/xmlui/handle/123456789/4015
dc.description.abstract In an attempt to solve the problem of finding a set of multiple unique modular robotic designs that can be constructed using a given repertoire of modules to perform a specific task, a novel synthesis framework is introduced based on design optimization concepts and evolutionary algorithms to search for the optimal design. Designing modular robotic systems faces two main challenges: the lack of basic rules of thumb and design bias introduced by human designers. The space of possible designs cannot be easily grasped by human designers especially for new tasks or tasks that are not fully understood by designers. Therefore, evolutionary computation is employed to design modular robots autonomously. Evolutionary algorithms can efficiently handle problems with discrete search spaces and solutions of variable sizes as these algorithms offer feasible robustness to local minima in the search space; and they can be parallelized easily to reducing system runtime. Moreover, they do not have to make assumptions about the solution form. This dissertation proposes a novel autonomous system for task-based modular robotic design based on evolutionary algorithms to search for the optimal design. The introduced system offers a flexible synthesis algorithm that can accommodate to different task-based design needs and can be applied to different modular shapes to produce homogenous modular robots. The proposed system uses a new representation for modular robotic assembly configuration based on graph theory and Assembly Incidence Matrix (AIM), in order to enable efficient and extendible task-based design of modular robots that can take input modules of different geometries and Degrees Of Freedom (DOFs). Robotic simulation is a powerful tool for saving time and money when designing robots as it provides an accurate method of assessing robotic adequacy to accomplish a specific task. Furthermore, it is difficult to predict robotic performance without simulation. Thus, simulation is used in this research to evaluate the robotic designs by measuring the fitness of the evolved robots, while incorporating the environmental features and robotic hardware constraints. Results are illustrated for a number of benchmark problems. The results presented a significant advance in robotic design automation state of the art. en_US
dc.language.iso en_US en_US
dc.subject Modular robots en_US
dc.subject Design optimization en_US
dc.subject Genetic algorithm en_US
dc.subject Evolutionary robotics en_US
dc.title Autonomous Task-Based Evolutionary Design of Modular Robots en_US
dc.type Thesis en_US
dc.institute.department School of Engineering en_US
dc.institute.name University of Bridgeport en_US

Files in this item

This item appears in the following Collection(s)

Show simple item record

Search ScholarWorks

Advanced Search


My Account