A Type of Low-Latency Data Gathering Method with Multi-Sink for Sensor Networks

Chao Sha, Jian-mei Qiu, Shuyan Li, Meng-ye Qiang, Ruchuan Wang2016 · DOI

TL;DR

This paper proposes a novel low-latency data gathering method (LDGM) for wireless sensor networks (WSNs) that utilizes multiple mobile sinks and a hierarchical region-based approach. By intelligently selecting "leader" nodes within regions and employing sleep scheduling and sensing radius adjustments, LDGM significantly reduces energy consumption and network redundancy while improving data collection efficiency.

What Problem Does This Paper Address?

Wireless Sensor Networks (WSNs) are crucial for monitoring various environments, but efficiently gathering data from a large number of distributed nodes presents significant challenges. A primary issue is the "hotspot problem," where nodes near a static sink become overloaded with data forwarding, leading to premature failure and network disconnection. To mitigate this, mobile sinks have been introduced, but optimizing their paths and coordination for low latency and balanced energy consumption remains complex. Existing methods often struggle with either high latency, unbalanced energy distribution, or the overhead of managing mobile sink trajectories. This paper aims to address these issues by proposing a multi-sink approach that balances energy consumption and reduces data collection latency.

Methods

The proposed LDGM method introduces several key innovations:

  • Virtual Region Division: The network is divided into "Data Gathering Units" (DGUs) and further grouped into "Data Gathering Areas" (DGAs). This hierarchical structure helps manage data collection efficiently.
  • Leader Selection: Within each DGA, a "leader" node is selected based on its residual energy and distance to other nodes. This leader acts as a local aggregator, reducing the number of nodes that directly communicate with the mobile sinks.
  • Sleep Scheduling and Sensing Radius Adjustment: To further reduce energy consumption and redundancy, nodes can enter a sleep mode, and their sensing radii are dynamically adjusted. This ensures that only necessary nodes are active and that their sensing coverage is optimized.
  • Multi-Sink Trajectory Optimization: The paper outlines strategies for optimizing the movement of multiple mobile sinks to visit traverse points (TPs) within each DGA. This involves considering factors like sink speed, distance between TPs, and data gathering rates to minimize overall delay and prevent buffer overflows.
  • Energy-Efficient Path Selection: Within each DGA, an energy-efficient path for data uploading from active nodes to the leader is determined using a weighted directed graph and Dijkstra's algorithm.

Datasets

Not discussed in the paper. The paper relies on simulations for evaluation.

Key Insights and Ideas

  • Hierarchical Aggregation is Key: By introducing regional leaders, the burden on mobile sinks is significantly reduced, and the number of direct transmissions is minimized. This is a more scalable approach than having every node communicate directly with a sink.
  • Dynamic Sensing is Crucial for Efficiency: Sleep scheduling and adaptive sensing radius adjustment are not just about saving energy; they directly combat data redundancy, which is a major source of wasted energy and increased latency.
  • Multi-Sink Coordination is Non-Trivial: While multiple sinks improve efficiency, their paths and timing need careful management to avoid collisions or idle periods, which the paper addresses through TP optimization.
  • Balancing Latency and Energy is a Trade-off: The paper demonstrates how LDGM achieves a better balance than existing methods, but acknowledges that further improvements in sink coordination could reduce latency even more.

Key Results

  • Energy Efficiency: LDGM shows improved energy efficiency compared to MST and MWST, particularly as the number of nodes increases. Average residual energy of nodes is higher in LDGM.
  • Reduced Redundancy: Sleep scheduling and sensing radius adjustment significantly reduce network coverage redundancy, with reductions of over 10% observed in some cases.
  • Lower Latency: The use of multiple mobile sinks and optimized trajectories leads to reduced data gathering time compared to single-sink methods.
  • Scalability: The method demonstrates good performance with varying numbers of nodes, maintaining stable energy consumption and coverage.

Limitations and Future Work

  • Sink Coordination Complexity: The paper acknowledges that while LDGM reduces latency, the possibility of multiple sinks visiting the same traversal point simultaneously can still extend upload delay in some DGAs. Future work could focus on further improving sink coordination to address this.
  • Real-World Deployment: The study is based on simulations. Real-world deployment might introduce additional challenges not fully captured in the simulation environment.
  • Dynamic Network Changes: The current model assumes static nodes. Future work could explore how LDGM adapts to dynamic network conditions, such as node failures or mobility.

Summary updated

BibTeX
@misc{sha2016type,
  author = {Chao Sha and Jian-mei Qiu and Shuyan Li and Meng-ye Qiang and Ruchuan Wang},
  title = {{A Type of Low-Latency Data Gathering Method with Multi-Sink for Sensor Networks}},
  year = {2016},
  doi = {10.3390/s16060923},
}
Figure