A Q-Learning Based Scheme for Neighbor Discovery and Power Control in Marine Opportunistic Networks.

Area of Science:

• Computer Science
• Electrical Engineering
• Maritime Technology

Background:

• Opportunistic networks face data transmission challenges due to sparse node distribution and complex energy management with mobile nodes.
• Maritime environments, characterized by high vessel mobility, often utilize opportunistic networks, demanding efficient energy solutions.
• Traditional neighbor discovery algorithms (passive listening, active probing) have limitations in energy efficiency and network congestion.

Purpose of the Study:

• To propose a hybrid neighbor discovery method for energy-constrained maritime opportunistic networks.
• To enhance neighbor discovery performance and data transmission efficiency in mobile maritime environments.
• To reduce energy consumption during the critical neighbor discovery phase.

Main Methods:

• Developed a hybrid neighbor discovery approach combining passive and active mechanisms.
• Optimized passive listening duration and implemented Q-learning for adaptive power control.
• Utilized a specialized wireless communication model suitable for maritime conditions.

Main Results:

• The proposed hybrid method significantly enhances neighbor discovery performance.
• Achieved substantial energy savings, up to 29%, during the neighbor discovery process.
• Demonstrated improved network throughput alongside energy efficiency.

Conclusions:

• The hybrid neighbor discovery method effectively addresses energy limitations in maritime opportunistic networks.
• The adaptive power control and optimized listening duration contribute to significant energy savings.
• This approach offers a promising solution for efficient communication in mobile maritime environments.