Two-Phase Robust Target Localization in Ocean Sensor Networks Using Received Signal Strength Measurements

Yuanyuan Zhang, Huafeng Wu, Xiaojun Mei, Jiangfeng Xian, Weijun Wang, Qiannan Zhang, Linian Liang2021 · DOI

TL;DR

This paper introduces a novel two-phase localization method (IPM-MM) for ocean sensor networks (OSNs) that robustly estimates target positions using received signal strength (RSS) even when the target's transmission power is unknown. By combining an Interior Point Method (IPM) for an initial estimate and a Majorization-Minimization Tactic (MMT) for refinement, the method achieves superior accuracy in dynamic ocean environments compared to existing techniques.

What Problem Does This Paper Address?

Accurate target localization is crucial for ocean observation and engineering tasks. Ocean sensor networks (OSNs) offer a promising solution, but the underwater environment presents significant challenges. Unlike terrestrial networks, acoustic signals used in OSNs are affected by complex factors like multipath propagation, absorption, and environmental dynamics, degrading localization accuracy. A key issue is the uncertainty in the target's transmission power, which can lead to substantial errors if not accounted for. Existing RSS-based localization methods, while cost-effective and synchronization-free, often struggle with accuracy in these dynamic conditions, especially when the transmission power is unknown. This paper aims to bridge this gap by proposing a method that can accurately locate targets in OSNs without prior knowledge of their transmission power, even in challenging underwater environments.

Methods

The paper proposes a two-phase approach:

  • Phase 1: Interior Point Method (IPM): The original localization problem, formulated using the maximum likelihood criterion, is transformed into an "unknown transmitting-power alternating non-negativity-constrained least square" (UT-ANLS) framework. This framework is then tackled using an IPM. IPM is an effective technique for convex quadratic programming and provides an initial, feasible solution for the UT-ANLS problem. However, due to the nature of the barrier function used in IPM, it may not perfectly approach the constraint boundaries, potentially leading to a sub-optimal solution.
  • Phase 2: Majorization-Minimization Tactic (MMT): To refine the solution obtained by IPM and achieve better accuracy, the MMT is employed. MMT iteratively optimizes the target location and transmission power. It works by constructing a surrogate function that approximates the original objective function, allowing for iterative updates. The solution from IPM serves as the initial guess for the MMT iterations. This combination aims to overcome the limitations of IPM by allowing for better convergence towards the true optimum.

Additionally, the paper derives the Cramer-Rao Lower Bound (CRLB) for the measurement model, providing a theoretical benchmark for the performance of any unbiased estimator.

Datasets

Not discussed in the paper. The evaluation is performed using numerical simulations.

Key Insights and Ideas

  • Joint Estimation of Location and Power: The core innovation is the simultaneous estimation of both the target's location and its unknown transmission power. This is crucial for robust localization in environments where transmission power can fluctuate or is simply not known beforehand.
  • Two-Phase Optimization: The combination of IPM and MMT is a key strategy. IPM provides a good starting point, while MMT refines it, effectively leveraging the strengths of both methods to achieve higher accuracy. This addresses the trade-off between initial feasibility and final precision.
  • UT-ANLS Framework: Transforming the problem into the UT-ANLS framework simplifies the complex, non-convex original problem into a more manageable form for optimization.
  • Ocean Environment Specifics: The paper explicitly considers the challenges of the ocean environment, such as absorption, and incorporates them into the model, making the solution more relevant for OSNs.

Key Results

The paper presents extensive simulation results comparing the proposed IPM-MM method with several existing algorithms (LS, SRWLS, ASM, RNLA, and IPM without MMT) across various scenarios:

  • Robustness to Path Loss Exponent: IPM-MM shows superior performance and robustness as the path loss exponent varies.
  • Performance with Varying SNR: The method consistently performs well with increasing Signal-to-Noise Ratio (SNR), often closely approaching the CRLB.
  • Impact of Anchor Nodes: As the number of anchor nodes increases, IPM-MM's performance improves, demonstrating its scalability and effectiveness with more available data.
  • Sensitivity to Absorption Coefficient: While all methods are affected by the absorption coefficient, IPM-MM maintains robust performance and remains optimal for target location estimation.
  • Effect of Side Length: IPM-MM exhibits relatively low error across different side lengths, indicating its suitability for varying deployment areas.
  • Noise Resilience: In noisy environments, IPM-MM provides the best choice for ensuring good performance in position estimation.
  • Improvement over IPM: The addition of the MMT phase significantly enhances the performance of the IPM-only approach, demonstrating the value of the two-phase strategy. For instance, in one scenario, the performance improvement from single IPM to IPM-MM was 52.5%.

Limitations and Future Work

  • Computational Complexity: While the paper doesn't explicitly detail computational complexity, the use of IPM and MMT suggests a potentially higher computational cost compared to simpler methods, though the simulations show good performance.
  • Assumptions: The model relies on certain approximations and assumes specific noise distributions (Gaussian). Real-world ocean noise can be more complex.
  • Future Work: The paper implicitly suggests that further exploration of different optimization techniques or adaptive strategies for handling more complex environmental dynamics could be beneficial. Investigating the method's performance with different acoustic propagation models or in scenarios with more dynamic target movement could also be valuable.

Summary updated

BibTeX
@misc{zhang2021two,
  author = {Yuanyuan Zhang and Huafeng Wu and Xiaojun Mei and Jiangfeng Xian and Weijun Wang and Qiannan Zhang and Linian Liang},
  title = {{Two-Phase Robust Target Localization in Ocean Sensor Networks Using Received Signal Strength Measurements}},
  year = {2021},
  doi = {10.3390/s21051724},
}
Figure