Clustering is an effective way to increase network lifetime but it leads to formation of isolated nodes in the wireless sensor network. These isolated sensor nodes forward data directly to sink and consume more energy which significantly reduces the network lifetime. In this article, we present how to maximize the network lifetime through joint routing and resource allocation with isolated nodes technique (JR-IN) between cluster head and isolated nodes in a cognitive based wireless sensor networks. In JR-IN technique the network area is divided into different layers and cluster size is formulated in each layer such that the size of the cluster remains unequal when it moves towards sink. Hence the cluster size is lager in the outermost layer compared to the cluster size in the inner most layer. To avoid inter cluster collision, we proposed different fixed channel to all the cluster heads in the network. For the intra cluster communication, the cluster member (sensor nodes) will lease the spectrum from the cluster head and forward data to their respective cluster head using TDMA technique. The periodical data gathering of cluster heads and forwarding the data to one hop cluster head may tend to lose energy faster and dies out quickly. We also propose in the JR-IN technique, the isolated nodes in the layer will take charge as a cluster head node and utilizes the resource allocated to the respective cluster head and forward the data to next hop cluster head. Simulation result shows that JR-IN outperforms the existing techniques, maximizes network lifetime and throughput and reduces the end to end delay.