What are the limitations of the existing AI/ML applications in MANETs?
- Traditional Routing Protocols: AODV, DSR, OLSR dominate early solutions
Q & A Forum
Literature Review & Gap Analyses
Q: What are the limitations of the existing AI/ML applications in MANETs?
- Scalability Limitations: DRL agents often come to failure within extremely high mobility of nodes or very large networks.
- Interpretability Issues: The black-box models predominated by the absence of justifications behind the decisions.
- Deployment Limitations: Memory, power, and compute capabilities are limited from the MANET nodes.
- Training Problems: All these require very large datasets which are often not available for MANETs.
- Adaptability Issues: Most of the AI models do not generalize across scenarios or topological changes within the model.