Abstract

The Indonesian Robot Contest (KRI) is an event to develop creativity and enrich science and technology in the field of robotics. In KRI, there are several branch categories, one of which is the KRPAI (Indonesian Fire Fighting Robot Contest) category. In the KRPAI category competition, the robot aims to find a room where there is a fire, then the robot will extinguish the fire in the room, and return to the starting point. This requires the robot to know its own actual position so that it can return to the starting point. This can be solved by adding artificial intelligence (Artificial Intelligence) artificial neural networks. One of the neural network models used for the learning process is backpropagation. It should be noted that in artificial intelligence systems, artificial neural networks in order to work optimally in identifying robot positions, it is necessary to do a learning process (learning) first in order to get the optimal weight value. The learning system is carried out by entering 280 position data samples that present 28 positions. Learning data is obtained by reading the distance between the robot and the maze wall and the direction facing the robot. To test the artificial neural network system, a running process is carried out to determine the position of the robot. From 140 trial data conducted, the artificial neural network was able to recognize the position of the robot correctly with an accuracy of 92.85%. The simulation results are expected to be applied to the actual quadpod robot.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call