انجمن کامپیوتر ایران

This article will introduce a location estimating method for prediction of the behavior of a mobile robot and its position errors after tracking various paths. Firstly, different controllable parameters were investigated by a statistical method and those which are absolutely effective were employed as the inputs for artificial neural network. Afterward, more tests were implemented for providing training data for the connectionist. Finally, the performance of the neural networks was studied by new set of data.