Table 6 Strengths and limitations of various approaches.

Privacy solutions

Strong privacy protection (encryption, differential privacy)^5,79,91

Higher privacy levels may reduce accuracy (trade-off)^79,97

Adversarial defenses

Advanced defenses via GANs, CNNs, adversarial learning^85,86,116

High computational overhead and energy consumption^68,85,115

FL

Decentralized and privacy-preserving^87,98,112

Challenges in handling non-IID data, higher convergence time^98,112

Real-world feasibility

Tested in real-world (IoT, GNSS-denied, large-scale systems)^88,89,93

Some solutions remain simulation-based, scalability concerns^82,104,113