Non-uniform k-Connectivity for Energy-Efficient and Reliable UWSNs

Cagla Tantur Karagul, TOBB Univ. of Econ. & Tech.

Mehmet Burak Akgun, TOBB Univ. of Econ. & Tech.

Huseyin Ugur Yildiz, TED University

Bulent Tavli, TOBB Univ. of Econ. & Tech.

​

Motivation

• Water cover >70% of the Earth Surface but knowledge of the oceans is still limited
• UWSNs used for coastal/environmental/defense
• Domination of energy limitations and reliability
• Need balance between reliability&energy

Problem Background

• Uniform k-connectivity costs high energy consumption
• High k usage in the network reduces network lifetime sharply
• Not every node in the network needs same connectivity value

Contribution

• Non-uniform k-connectivity
• Higher k selection for critical nodes
• MILP formulation
• Coastal UWSN evaluation

System Model

• 12 nodes, 3 km linear topology
• Two different BS placements
• 10 power levels
• Data + control packets

System Model

Energy Model

• Absorption derived from Thorp’s formula
• Transmission + reception cost
• Data & control package transmission
• Distance-based power selection

MILP Summary

• Objective: minimize
• Flow conservation constraints
• Node-disjoint/link-disjoint paths constraints
• Non-uniform k constraints

Connectivity Configurations

• Uniform k=1,2,3
• Mixed connectivity groups

Results

BS at x = 1.5 km

BS at x = 0 km

Uniform k Findings

• k=2: up to 2.72x energy increase
• k=3: up to 6.17x energy increase

Compared to 1-connected case

Non-uniform k Findings

• Config II: +1.91x
• Config IV: +3.32x
• Config V: +4.16x

Compared to Config I

Control Frequency Effect

• ψ increase from 0.25 to 4→ 2.6–5.6x energy

BS Deployment Effect

• BS at center reduces energy
• 2.22–4.05x energy increase when BS located at x = 0

Implications

• Non-uniform k improves energy efficiency
• Provides high relability for necessary nodes

​

Conclusion

• Balanced energy–reliability solution
• Future work: MILP decomposition, relaxation techniques, mobile BSs