A Genetic Approach for Trajectory Optimization Applied to a Didactic Robot

• Oscar Kremer IFSUL
• Mauro Cunha IFSUL
• Fabiano Moraes IFSUL
• Sthefano Schiavon IFSUL

Resumo

This work presents a genetic and adaptive genetic algorithm for generating optimized trajectories for a didactic robot. Manipulator robots are used in industrial environments to perform repetitive tasks, unhealthy and dangerous jobs, among others. Optimization in the execution of these tasks is essential to make the use of these technologies compatible with the industry 4.0 requirements. The trajectory that a robot performs during its operation is directly related to its wear, consumption, time and even space of security that should be delimited inside of the workspace. The artificial intelligence popularization parallels the expansion of its applications and technologies, such as genetic algorithms, fuzzy systems, reinforcement learning and convolutional networks. By modeling the kinematics of a commercial didactic robot, an adaptive genetic algorithm used to optimize the trajectories is developed, presenting trajectories generated by polynomial interpolation, traditional and adaptive genetic algorithm, thus analyzing the possibility of reducing the total distance covered and other metrics by the technique adopted. Simulation results illustrate the main characteristics of the proposed algorithms.

Palavras-chave: Genetic algorithms, Trajectory, Kinematics, Service robots, Manipulators, Sociology