The data-centric wireless sensor networks comprise numerous autonomous tiny nodes forms random topology in nature. Applications oriented WSNs immensely used for monitoring harsh environment. Its unique constrains are distinguishing it from traditional networks by energy, lifetime, fault tolerance, scalability and computational power. When they are deployed randomly sensing & generating vast amount of data, for which they are being used. Network meets Congestion, if huge volume of data passed. To eradicate it, misbehaving nodes are identified and skilled for self-healing without human intervention. Existing approaches focus on controlling link level congestions not node level. Our proposed fusion-centric scheme controls node congestion. To achieve this, selectively concentrate on intermediate level. Misbehaving nodes are identified by using their historic data based on fault level occurred. Lifetime is determined by allocation-rate and more radio signal usage. Our scheme keeps the network without consuming much resource. Node level control is needed for self-configuring WSNs.