Table 1 Review of existing Techniques.

Humayun et al.²⁰

Energy optimization in smart cities integrating IoT, 5G, and cloud computing

Model enhances energy use in smart homes, lampposts, smart parking; cloud stores data efficiently; 5G enables fast communication

Primarily simulation-based; real-world deployment not tested

Nagaraju et al.²¹

Secure and energy-aware routing in IoT-enabled WSNs

TEEN variants and multipath routing improve energy efficiency, throughput, storage, network lifetime

Security concerns remain in end-to-end communication; scalability not tested

Luo et al.²²

Energy optimization for continuous data-flow applications

Mixed integer nonlinear programming with max-flow algorithm reduces energy consumption in simulation

Only tested in simulation; real-world efficiency may differ

Sarwar et al.²³

IDS for IoT networks using optimized PSO (IDSBPSO)

Improved anomaly detection accuracy, low computational cost, fewer features required

Limited to IoTID20 and UNSW-NB15 datasets; may not generalize to other datasets

Ahmad et al.²⁴

NIDS for IoT anomaly detection using MI and DNN

Reduced feature set improves detection accuracy by 0.99–3.45%

High false alarm rate possible; dataset-specific performance

Samani et al.²⁵

Load forecasting and deviation detection in IoT-BEMS

Achieved 30% load shedding in LED luminaires using DNN-based model

Limited to one building; may not generalize to other environments

Lydia et al.²⁶

Green IoT routing with DL-based anomaly detection

GEER-DLAD with LSTM improves energy efficiency and detection capability

Limited experimental settings; real-world deployment challenges

Sivakumar et al.²⁷

Energy optimization in Industry 4.0 sensor nodes

Dynamic power management, scheduling, and harvesting improve energy efficiency

Applicability to heterogeneous industrial networks not fully validated

Revanesh et al.²⁸

Prolonging WSN lifetime via neural networks (LEACH, EESR)

Improved energy efficiency and reliability using LM-NN; reduces number of sensors needed

Battery-powered nodes still limited; monitoring in large-scale WSNs challenging

Saheed et al.²⁹

IDS in SCADA systems using ELM + PCA + GWO

Optimized classifiers improve detection performance in SCADA systems

Limited to MSU gas pipeline and water utility datasets; may not generalize

Ramalingam et al.³⁰

Hybrid energy-efficient WSN routing (Fuzzy + ASFO + EHO)

QoS metrics improved; PDR 99.8%, latency 1.12s, throughput 98bps

Simulation-based; heterogeneous WSN scalability not tested

Dhanasekaran et al.³¹

Malware detection in 5G using CNN-LSTM

High accuracy (99.8%) and F1-score (0.9925); outperforms existing models

Single dataset; challenges in heterogeneous real-time network deployment