<![CDATA[Wheeled robot is a robot that moves using wheels under certain conditions so that a good navigation system is needed. Whether or not the robot navigation system is influenced by the software system and the algorithm that is running. Over time, artificial neural network algorithm is much in demand and developed because it has good flexibility to the nonlinear environment. Therefore, in this study a backpropagation artificial neural network algorithm is designed for navigation of wheeled robots. Artificial neural networks have advantages, among others, the level of accuracy is better than pattern recognition algorithms such as fuzzy and KNN. Besides artificial neural networks also have good flexibility in solving problems with multidimensional classes. Backpropagation artificial neural network is a supervised neural network in which there is a network training. In network training there are several factors that influence, including hidden layers, momentum, and learning rate. In addition to good training, robot speed regulation also needs to be considered so that the robot has good stability in its environment. Based on the time and error values ​​in the network training, the number of hidden layer nodes of 5 has the best results on input and output of 4 nodes. Whereas in the PWM motor 150 value setting obtained a fairly good movement stability for the wheeled robot used.]]>
Copyrights © 2020
