A Large-Scale Multi-Hop Localization Algorithm Based on Regularized Extreme Learning for Wireless Networks.

Area of Science:

• Computer Science
• Network Engineering
• Machine Learning

Background:

• Traditional large-scale multi-hop localization algorithms often struggle with complex, non-isotropic network environments.
• Existing methods may require extensive parameter tuning and exhibit limited adaptability.
• Accurate node localization is crucial for various network applications.

Purpose of the Study:

• To propose a novel large-scale multi-hop localization algorithm using regularized extreme learning.
• To develop an algorithm adaptable to complex deployment environments beyond isotropic networks.
• To achieve accurate localization with low computational cost and minimal parameter settings.

Main Methods:

• Formulating the large-scale multi-hop localization problem as a machine learning task.
• Employing regularized extreme learning to model the relationship between hop-counts and physical distances.
• Implementing a three-stage process: data acquisition, model construction, and distributed location estimation.

Main Results:

• The proposed algorithm demonstrates strong adaptability to diverse topological environments.
• Experiments confirm low computational cost and high localization accuracy.
• The method effectively overcomes limitations of traditional algorithms in complex settings.
• Accurate results were achieved without the need for complex parameter configuration.

Conclusions:

• The regularized extreme learning-based algorithm provides an effective solution for large-scale multi-hop localization.
• Its adaptability and accuracy make it suitable for complex and real-world network deployments.
• The algorithm offers a computationally efficient and user-friendly approach to node localization.