Moving Ball Destination Prediction using Artificial Neural Network for Soccer Robots

Designing coordinated and intelligent behavior in dynamic, real-time, uncertain adversarial environment; such as in soccer robot is a challenging task. Using available data and processing them in a way leading to a more intelligent understanding is an essential and efficient factor in the performance of decision making systems and will improve robot’s collective intelligence. One of the important issues in coordinating human reaction to a moving object, for instance, while playing games such as soccer, is the prediction of the future position of that object. Reliable and accurate localization of soccer robots has been a debatable issue for years; hence, high performance methods are available. Using the data achieved from the localization and artificial neural networks, we proposed a method which enables robots to predict the moving ball destination and to make wise-human-like choices in every possible situation of the ball movement. The achieved results show the predictive ability of this method.