Design and Simulation Transfer Learning on Image Processing for Determining Condition of Robot Based on Neural Network

Neural Networks or Artificial Neural Networks (ANN) are an interesting topic in the last decade. This is due to the ability of ANN to mimic the nature of the input system. ANN is an information processing system that has similar characteristics to biological neural networks. This research will discuss the application of the Neural Network method for image processing which aims to classify the condition of the robot. The input for dataset are RSTRAIGHT, RTURNRIGHT, and RTURNLEFT, with a total dataset of 300 images. The image was taken using a camera with various conditions of the robot in various positions. The output of simulation is a learning rate precision which will use as a reference for whether the system can recognize the robot’s condition or not. In this study, the learning rate precision output was 98% which was obtained in the 12th iteration with an epoch of 50 and a batch size of 10 using the Adam optimizer.