Wireless sensor network (WSN) is networks with no physical infrastructure made up of mobile nodes connected by wireless links. WSN is susceptible to clone attacks or node replication attacks. By breaching a single sensor node, attackers can create several clones with identical identity (ID) from the hacked node, which they can then distribute throughout the network. But achieving optimal energy consumption has never been easy since sensor nodes (SNs) in WSN have short battery lives. Nonetheless, the process of choosing the cluster head (CH) and forming the ideal number of clusters determines how well clustering works. In this paper, energy-efficient sleep awake aware (EESAA) routing protocol combined with deep learning is a proficient approach for enhancing network lifetime and security aligned with attack by preserving quality of service (QoS). EESAA, clustering, and CH are selected using optimization methods like Enhanced Whale Optimization Algorithm (EWOA), Entropy Dragon Fly Optimization Algorithm (EDFOA), Mean Donkey and Smuggler Optimization (MDSO), and Entropy Dove Swarm Optimization (EDSO). Results of routing protocols are measured using the assessment metrics like packet delivery ratio (PDR), packet loss ratio (PLR), end-to-end delay (E2ED), and energy consumption. Attack detection results are measured using the various parameters such as precision, recall, f-measure, and accuracy.