Reducing Delay and Maximizing Lifetime for Wireless Sensor Networks With Dynamic Traffic Patterns

Qiang Wang, Wei Liu, Tian Wang, Ming Zhao, Xiong Li, Mande Xie, Ming Ma, Guoping Zhang, Anfeng Liu

Introduction

Wireless Sensor Networks (WSNs) are often constrained by the "energy hole" phenomenon. In a typical configuration, nodes located near the central data collection point, or "sink," bear the brunt of relaying data from the entire network. These hotspot nodes deplete their energy far faster than peripheral nodes, leading to network failure even when a significant portion of the total energy remains unused.

Traditional data collection protocols often exacerbate this by assuming static traffic patterns and enforcing a uniform transmission radius ($r$) for all nodes. In reality, traffic is dynamic—nodes only transmit when events are detected. This paper introduces the Reducing Delay and Maximizing Lifetime (RDML) strategy to address these inefficiencies by leveraging surplus energy for long-distance transmissions.

Methodology: The RDML Strategy

The RDML strategy breaks the convention of uniform transmission radii to optimize both latency and energy distribution. It consists of two primary components:

1. Mixed Transmission Radii

Instead of a fixed radius, RDML identifies nodes with surplus energy (typically those further from the sink) and allows them to use a larger transmission radius, $mr$ (where $m > 1$). This "shortcut" allows data packets to bypass intermediate relay nodes and reach the sink or a closer relay directly. By "spending" the excess energy of peripheral nodes, the strategy reduces the relay burden on hotspots and decreases the total hop count.

2. Dynamic TDMA Scheduling

To handle unpredictable traffic, the authors developed a scheduling algorithm based on a post-order traversal of the network's routing tree.

The k-ary routing tree topology used for evaluating the RDML strategy.

The algorithm allocates time slots consecutively. If a receiver detects an empty slot, it signifies the sender has no more data, allowing the receiver to enter sleep mode immediately. This minimizes energy waste from "idle listening."

Evaluation and Results

The RDML strategy was evaluated through Matlab simulations using $k$-ary routing trees with varying depths ($d=3$ to $8$) and node degrees ($k=2$ to $5$). The performance was benchmarked against the Advanced Two-Phase Optimization (TPO) algorithm.

Key findings include:

  • Latency: Data collection delay was reduced by up to 22.32%.
  • Network Lifetime: Operational life increased by 0.86% to 84.40%, with the most significant gains seen in networks with even tree depths where load balancing is more symmetrical.
  • Energy Efficiency: Effective energy utilization saw a massive boost, ranging from 23.33% to 88.00%.

Discussion and Future Work

The RDML strategy successfully turns the "energy hole" problem into an opportunity by utilizing wasted peripheral energy to bypass bottlenecks. However, the authors note that increasing the transmission radius expands the interference zone, complicating TDMA slot scheduling. Future research will focus on implementing RDML on physical sensor platforms to validate these simulation results against real-world signal interference and environmental noise.

Summary updated

Abstract

Fast and energy efficient data collection is the most important function of wireless sensor networks (WSNs) which governs the performance of WSNs. Time division multiple access (TDMA) is an effective data collection strategy in which node only schedules the active slot to save energy when transmitting data. However, many previous studies are designed for static traffic pattern such that each node to generate one data packet in a data collection cycle. Moreover, all nodes in the network use the same transmission radius r, so that the network data collection has a larger delay, a lower lifetime, and energy utilization rate. In view of the shortcomings of previous research, this paper proposes a reducing delay and maximizing lifetime (RDML) data collection strategy for WSNs with dynamic traffic patterns. The main innovation point of our work is the data transmission radius is no longer unified as r, while some nodes with energy surplus use a data transmission radius of mr(m > 1). The following network performance can be achieved by adopting the RDML scheme. First, because the nodes in the near sink mr radius area can send data directly to sink, the amount of data carried by hotspots nodes in the original strategy is greatly reduced, and the lifetime is improved. Second, since some nodes can go mr distance to the sink in data transmission, thus accelerating the data routing and reducing the time needed for data collection. Finally, since the remaining energy of nodes is fully utilized to increase the radius of data transmission, the energy utilization rate can be effectively improved. Through the theoretical analysis and experimental results, the RDML scheme can reduce the maximum delay of data collection by 22.32% and improve lifetime as well as the energy utilization rate by 0.86%-84.40%, and 23.33%-88.00%, respectively, compared with previous strategies.
BibTeX
@article{wang2019reducing,
  author = {Qiang Wang and Wei Liu and Tian Wang and Ming Zhao and Xiong Li and Mande Xie and Ming Ma and Guoping Zhang and Anfeng Liu},
  title = {{Reducing Delay and Maximizing Lifetime for Wireless Sensor Networks With Dynamic Traffic Patterns}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2918928},
}
Figure