In this work a new protocol able to obtain network cooperation in a wireless sensor network reducing the control overhead and reducing the data distortion is developed. Its aim is to select the minimum number of sensors able to cover the phenomena and to report not correlated data to the sink in order to reduce distortion. The proposed protocol is called Minimal Distortion with Maximal Cover (MDMC) protocol and it tries, on the one hand, to gradually expand the covered phenomenon area and, on the other hand to reduce the distortion that arises from the interaction of the sensors. To reach this goal, the protocol uses two heuristics along with an activation technique, the first ones select the sensors on the basis of their features, whereas the second one decreases the number of sensors which take part in the coalition formation process. The proposed has been compared with a cluster based protocol such as Local Negotiated Clustering Algorithm (LNCA).