NACRP : A Connectivity Protocol for Star Topology Wireless Sensor Networks
Abstract
NACRP : A Connectivity Protocol for Star Topology Wireless Sensor Networks
Data upload time is a large portion of mobile data gathering time in wireless sensor networks. Data uploading time can be greatly shortened by equipping multiple antennas on the mobile collector.
However, previous works only treated wireless connectivity as a constant and ignored power control on sensors, which would deviate significantly from real wireless environments.
In this NACRP : A Connectivity Protocol for Star Topology Wireless Sensor Networks paper we propose a new data gathering cost minimization framework for mobile data gathering in wireless sensor networks by jointly considering dynamic wireless connectivity capacity and power control.
Our new framework not only allows the simultaneous uploading of data from sensors to the mobile collector, but also determines transmission power under elastic connection capacities. Under constraints of flow conservation, energy consumption, elastic connection capacity, transmission compatibility, and sojourn time, we study the problem.
We use the subgradient iteration algorithm to solve the problem of minimization. We first relax the Lagrangian dualization problem, then decompose the original problem into several subproblems and present distributed algorithms to derive data rate, link flow and routing, power control and compatibility with transmission.
Conclusion
In this paper, we have designed a cross-layer optimization framework for mobile data collection in WSNs, considering elastic connection capacity and sensor power control. We use subgradient iterative approach to solve the problem and present several distributed subalgorithms with explicit message transmission. Extensive numerical results demonstrate convergence within 50 iterations of the proposed algorithm.
First, performance gains compared to system complexity should be further studied using MMSE receivers. Second, the cost function used in this paper may not fully reflect the overall pricing structure in the network.
Therefore, a more comprehensive model that accounts for aspects of the transmission/reception energy, buffer, encode/decode, moving energy, and human administration costs of the sensor may be considered in the future.
