Table 1 Summary of related work on AI-based routing in WSNs with identified gaps.

⁸²

Study the security issues related to AI-based routing algorithms in WSNs.

Assesse11027594s weaknesses and suggests security protocols.

Lacks real-time AI integration for intrusion response and anomaly detection.

⁸³

Evaluate computational burden of AI-based routing.

Optimization suggestions for resource constraints.

Does not present a full modular framework deployable on sensor hardware.

⁸⁴

Investigate ethical and societal risks of AI routing.

Ethical principles for fairness and privacy.

No operational routing model tested with ethical constraints.

⁸⁵

Explore RL in adaptive routing.

RL-based protocol that learns from network states.

Does not integrate global optimization or energy balancing.

⁸⁶

Assess strengths/weaknesses of AI routing methods.

Thematic review of adaptivity and limitations.

Survey only; lacks implementation or hybrid framework proposal.

⁸⁷

Integrate AI-based routing with IoT for interoperability.

AI protocols for smooth IoT communication.

No details on adaptation to constrained energy and QoS needs.

⁸⁸

Apply DL to detect anomalies in WSNs.

DL-powered intrusion detection based on traffic patterns.

Lacks integration with adaptive routing and lightweight models.

⁸⁹

Examine environmental impacts on AI routing.

Routing adjustments based on environmental factors.

No real-time learning mechanism to adapt to volatile conditions.

⁹⁰

Use ensemble learning for routing resilience.

Combines classifiers to boost reliability.

High complexity; no demonstration on embedded sensor nodes.

⁹¹

Combine blockchain with AI-routing for trust.

Blockchain-backed secure routing proposals.

Introduces latency and computation cost not suitable for WSNs.

⁹²

Examine routing under mobility in WSNs.

Mobility-aware protocols with adaptive paths.

Lacks use of learning or predictive mobility handling.

⁹³

Apply game theory to reduce selfish routing.

Cooperative routing through game-theoretic incentives.

Does not explore hybrid AI models or security integration.

⁹⁴

Explore bio-inspired optimization for routing.

Firefly and cuckoo-based optimization schemes.

Lacks integration with learning models and practical tuning.

⁹⁵

Use edge computing to support AI routing.

Offloads computation to edge nodes for scalability.

No joint optimization with local routing decisions or real-time learning.

⁹⁶

Balance energy and delay in routing.

Trade-off optimized routing strategies.

Does not use adaptive learning or predictive energy modeling.

⁹⁷

Apply swarm robotics concepts to WSN routing.

Self-organizing routing via swarm algorithms.

Focused on local behavior; lacks global coordination and learning.