A Routing Protocol for Wireless Sensor Networks with Congestion Control

Qi Li, Zongwu Ke, Duanfeng Xia, Sun Yu-xiaCommunications 2013 · DOI

Abstract

This paper proposes a novel Random Early Detection (RED) protocol that incorporates node energy levels and data packet length to manage network traffic. Additionally, an ant-colony-inspired routing protocol is introduced to mitigate congestion by imitating foraging behavior, where sensor nodes select routes based on pheromone density. Experimental results demonstrate that this dual-strategy approach effectively balances energy consumption across nodes and mitigates network congestion compared to traditional methods.

Introduction

Wireless Sensor Networks (WSNs) serve as the foundational infrastructure for the Internet of Things (IoT). However, they are frequently hindered by an "energy-congestion" trap. Modern real-time applications, such as healthcare monitoring and target tracking, generate significant data volumes that often exceed the limited buffer capacity and battery life of small sensor nodes.

When traffic spikes occur, buffers overflow, leading to transmission delays and substantial energy waste due to retransmissions. Furthermore, if a specific node becomes a primary data conduit, it quickly exhausts its battery, creating network gaps. Traditional congestion control methods often fail to account for varying packet sizes or the remaining energy of nodes, leading to inefficient and "blind" packet dropping.

Methodology

The proposed system integrates two primary mechanisms to enhance network longevity and reliability:

Improved Random Early Detection (RED) Model

Unlike standard RED models that drop packets based solely on the number of items in a queue, this improved model is byte-aware, calculating drop probability based on the actual data length in the buffer. Crucially, it introduces an energy factor ($e$); as a node's residual energy decreases, its drop probability increases. This proactively signals the network to reroute traffic away from dying nodes.

Ant Colony Routing Protocol

Inspired by biological ant colonies, nodes maintain routing tables based on "pheromone density." Paths that successfully deliver data accumulate pheromones, while paths experiencing timeouts or congestion see their pheromone levels reduced exponentially ($Q/2^k$). When selecting the next hop, nodes consider both pheromone density (path quality) and the residual energy of neighbors, preventing the formation of energy "hotspots."

Experimental Setup

The protocol was evaluated using a simulated topology consisting of seven nodes: three sensor nodes ($S_1, S_2, S_3$), three routers ($r_1, r_2, r_3$), and one sink node. Traffic was simulated at a constant rate of 100 bytes per second, with buffer thresholds set between 500KB and 1000KB to test the limits of the congestion control mechanisms.

Key Results

Congestion Mitigation

In comparative simulations, the traditional RED protocol reached a maximum drop probability and remained there, indicating sustained congestion. In contrast, the proposed protocol's drop probability peaked and then significantly decreased. This trend suggests that the ant colony logic successfully identified congestion and rerouted traffic to clearer paths before the buffer reached a critical state.

Comparison of packet drop probability over time between the proposed and traditional protocols.

Energy Balance and Network Lifetime

Under traditional routing, the central router ($r_2$) experienced a sharp decline in energy, eventually exhausting its battery while peripheral routers remained underutilized. The proposed protocol achieved a much more balanced distribution of the load. As the energy of the primary router ($R_2$) reached a specific threshold, the network dynamically shifted traffic to routers $R_1$ and $R_3$, extending the operational life of the entire network.

Energy consumption of nodes over time, showing improved load balancing in the proposed protocol (dotted lines) versus traditional methods (solid lines).

Conclusion

By integrating energy levels directly into congestion control and employing dynamic, nature-inspired routing, this protocol addresses the critical hardware constraints of WSNs. While the current simulation demonstrates clear performance gains in load balancing and congestion avoidance, future work will focus on scaling the protocol for larger, more complex networks and testing its stability in dynamic topologies with mobile nodes.

Summary updated

Abstract

This paper proposed a novel RED protocol which takes the node's energy into account depending on the length of the data packet. It also proposed a routing protocol for wireless sensor networks with congestion control which imitates the ant colony foraging behavior. Sensor nodes choose routings according to the pheromone density. The experiment result shows that the algorithm balances the energy consumption of each node. It mitigated congestion effectively with the proposed routing protocol.
BibTeX
@inproceedings{li2013routing,
  author = {Qi Li and Zongwu Ke and Duanfeng Xia and Sun Yu-xia},
  title = {{A Routing Protocol for Wireless Sensor Networks with Congestion Control}},
  booktitle = {Communications},
  year = {2013},
  doi = {10.4236/cn.2013.53b2030},
}
Figure