Artificial Intelligence for Robotics

Neural networks

How about this? We use a neural network to classify the images from our camera. We drive the robot around, and take a picture about four times a second. We record what the robot is doing at each picture – going forward, turning right, turning left, or backing up. We use that information to predict the robot's motion path given the image. We make a convolutional neural network, with the camera image as the input and four outputs – go forward, go left, go right, and back up. This has the advantage of avoiding obstacles and hazards automatically. When we get to the stairs (remember I have stairs going down in my game room that would damage the robot), the robot will know to turn around, because that is what we did in training. ...