Node Depth Adjustment Based Target Tracking in UWSNs Using Improved Harmony Search

Meiqin Liu, Duo Zhang, Senlin Zhang, Qunfei ZhangSensors 2017 · DOI

TL;DR

This paper tackles the challenge of accurately tracking moving targets in underwater wireless sensor networks (UWSNs) by introducing a novel node depth adjustment strategy. By dynamically optimizing sensor node depths, the system significantly improves tracking accuracy, especially when targets are distant, outperforming static node configurations.

What Problem Does This Paper Address?

Underwater wireless sensor networks (UWSNs) are crucial for exploring and monitoring underwater environments, including tracking moving targets. However, deploying dense sensor networks is costly and impractical. Existing UWSN tracking schemes often rely on static nodes, which limits their accuracy, especially when targets move far from the sensors. This paper addresses the problem of improving underwater target tracking accuracy in UWSNs by enabling activated sensor nodes to dynamically adjust their depth, rather than remaining static or passively drifting with water currents. The core challenge is to determine the optimal depth for these nodes to maximize the information gathered for accurate tracking.

Methods

The paper proposes a novel approach that combines node depth adjustment with an improved Harmony Search (HS) algorithm for target tracking in UWSNs.

  1. Fisher Information Matrix (FIM) for Accuracy Metric: The authors establish a relationship between the FIM and the depth of sensor nodes. The FIM quantifies the information content of measurements, and a larger FIM indicates better estimation accuracy. By maximizing the determinant of the FIM, they aim to optimize tracking performance.
  2. Optimization Problem Formulation: The problem of determining the optimal node depth is formulated as a dynamic optimization problem. The objective is to maximize the determinant of the FIM, subject to constraints on the nodes' vertical movement range.
  3. Improved Harmony Search (HS) Algorithm: To efficiently solve this complex optimization problem, the authors propose an enhanced HS algorithm. Key modifications include:
  4. Modified Harmony Generation Probability: The probability of generating a new harmony (a set of node depths) is adjusted to favor depths closer to the target, thereby improving search speed and accuracy.
  5. Dynamically Updated Parameters: Parameters like Harmony Memory Considering Rate (HMCR), Pitch Adjusting Rate (PAR), and Bandwidth (BW) are dynamically updated during the iteration process to avoid local optima and enhance performance.
  6. Multi-Sensor Particle Filter (PF): A standard multi-sensor Particle Filter is used for the actual target state estimation, leveraging the measurements from the depth-adjusted nodes.

Datasets

Not discussed in the paper. The evaluation is based on simulations.

Key Insights and Ideas

  • Depth is a powerful control parameter: The paper demonstrates that actively controlling the vertical depth of underwater sensors is a highly effective strategy for improving target tracking accuracy, a concept not previously explored for this specific application.
  • FIM as a proxy for tracking quality: Using the determinant of the Fisher Information Matrix as an objective function provides a mathematically sound way to quantify and optimize the information content of sensor measurements for tracking.
  • Target distance matters: The effectiveness of depth adjustment is most pronounced when targets are farther away, highlighting the limitations of static sensor configurations in challenging scenarios.
  • Harmony Search adaptation: Modifying the Harmony Search algorithm by biasing its search towards more promising solutions (closer to the target) significantly improves its efficiency and effectiveness for this specific optimization task.

Key Results

  • The proposed MOVE scheme (using node depth adjustment) consistently outperforms the STATIC scheme (nodes remain at the surface) in terms of tracking error across various target depths.
  • For instance, with a target depth of -120m, the MOVE scheme achieved an average tracking error of 6.252m, while the STATIC scheme had an error of 7.807m, representing a 19.92% improvement. This improvement becomes more significant as the target moves deeper.
  • The improved HS algorithm shows faster convergence and achieves a better objective function value compared to the traditional HS algorithm, indicating improved searching speed and accuracy. The improved HS converged to -0.9818, while the traditional HS converged to -0.9769.
  • The MOVE scheme's performance is less sensitive to node density (adjacent node distance) compared to the STATIC scheme, suggesting its robustness in both sparse and dense UWSN deployments.

Limitations and Future Work

  • Energy Consumption: The paper acknowledges that while the proposed method improves tracking accuracy, it doesn't explicitly balance this with energy consumption. Future work will focus on finding an optimal trade-off.
  • Node Number Determination: The current work assumes a fixed number of activated nodes (three). Future research will explore adaptive methods to determine the optimal number of nodes for tracking.
  • Complex Environments: The paper notes that node location errors and complex measurement models in real-world underwater environments are challenges that need to be addressed in future work.
  • Multiple Targets: The current scheme is designed for single target tracking. Extending it to multiple targets is a potential future direction.

Summary updated

BibTeX
@article{liu2017node,
  author = {Meiqin Liu and Duo Zhang and Senlin Zhang and Qunfei Zhang},
  title = {{Node Depth Adjustment Based Target Tracking in UWSNs Using Improved Harmony Search}},
  journal = {Sensors},
  year = {2017},
  doi = {10.3390/s17122807},
}
Figure