Wireless sensor networks (WSNs) have found widespread use in both civilian and military settings in recent years. Due to their decentralized and unprotected setup, sensor networks can be easily compromised. It is important to adopt effective methods to safeguard sensor networks from the many routing protocol attacks that can be conducted against them. Preprocessing, feature selection, and model training are all parts of the proposed methodology. To enhance classifier performance through more effective model training, data preprocessing is essential. The proposed approach utilized k-means and Hierarchical clustering for feature selection. After accumulating the attributes, the models are trained with AlexNet-GRU. The proposed method outperforms other approaches like GRU and AlexNet by a wide margin. This strategy has a 98.64% chance of working.