Parameter identification of a fuzzy model on the handwriting process of an arabic letter

We propose a new approach to recognize cursive characters. In particular, there are comparatively few researches on recognition of Arabic letters. On the basis of Van der Gon model that modeling the handwriting process of human beings, we develop a mathematical model of handwriting process in the same way. Although we must identify some parameters in the model, the movement of forearm in the handwriting process of letters has a certain strong nonlinearity from the mechanical viewpoint as one can predict it easily. The handwriting process can not be explained by a simple linear model as the relation between each output signal of electromyographic and the writing force is expressed by a linear combination of muscle signals. So, we apply a fuzzy identification technique to model the handwriting process. We aim to absorb the nonlinearities contained in the process by using a fuzzy model.