Evolutionary Tabu Inverted Ant Cellular Automata with Elitist Inertia for swarm robotics as surrogate method in surveillance task using e-Puck architecture
| Autor: | Danielli A. Lima, Hamilton J. M. Lopes |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
0209 industrial biotechnology
Computer science business.industry General Mathematics Swarm robotics 02 engineering and technology Ant colony Cellular automaton Tabu search Computer Science Applications 03 medical and health sciences 020901 industrial engineering & automation 0302 clinical medicine Surrogate model Control and Systems Engineering Control theory 030220 oncology & carcinogenesis Genetic algorithm Robot Artificial intelligence business Software |
| Zdroj: | Robotics and Autonomous Systems. 144:103840 |
| ISSN: | 0921-8890 |
| DOI: | 10.1016/j.robot.2021.103840 |
| Popis: | The area of swarm robotics has grown widely in recent years, precisely because its formulation is based on the use of various techniques, ranging from computer networks to controllers. We can employ different types of techniques to carry out the control of a robots team. In this work we will focus on creating techniques based on bio-inspired computing. Within this theme, we will be focused on using cellular automata with synchronous and asynchronous rules and ant colonies optimization. Additionally, we will consider greedy approaches to select the next robot’s state cell and a local Tabu search with a queue of robot movement restrictions. Thus, we have a surrogate model capable of providing the team robot navigation in the surveillance task. We developed two different controllers, a simpler first, based on a precursor model and a second optimized model, based on the previous controller refinement. At the end, we used a genetic algorithm, which received the surrogate model as input for the improvement of our proposed models parameters. In addition, a survey with the evolution of surveillance models using cellular automata in a systematic review of literature will be shown. Experiments were performed to demonstrate the degree of robot team coverage by different environments. We accomplished statistical analysis with the intention of presenting different sizes of robot teams and amounts of pheromone deposited into the environment. In the end, we fulfilled experiments using the empirical simulation methodology of a robots team using the Webots simulator with e-Puck architecture. The results were promising, the robot team performed this task efficiently and the system is highly scalable. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect