Scholarlink

Mobile source positioning/localization estimation is a crucial problem with wide-ranging applications. Currently, numerical estimation algorithms primarily rely on a single type of measurement method, such as time-of-arrival (TOA) or angle-of-arrival (AOA). This article introduces a hybrid approach that combines AOA and TOA measurements to achieve robust source localization, particularly when there are outlier noises. Instead of using traditional numerical methods, we explore the utilization of neurodynamics models for estimating the location of a mobile source. Our formulation considers several key issues. To handle non-Gaussian noise (outlier noise), we incorporate \(\ell _1\) -norm and \(\ell _2\) -norm loss functions. We first formulate the problem as a constrained optimization problem with inequality and equality constraints. Remarkably, we demonstrate that the inequality constraints can be eliminated. However, the non-differentiability of the \(\ell _1\) -norm loss function poses implementation challenges for the neurodynamics equations. To overcome this obstacle, we propose a smooth approximation for the \(\ell _1\) -norm loss component. Furthermore, we conduct a stability analysis of our proposed neurodynamics model. Finally, we present simulation results to evaluate the performance of our method. These results confirm the superiority of our proposed neurodynamics model over several numerical methods, particularly in scenarios with high measurement noise.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Neural Network Approach for Robust Estimation of Position of a Target Based on TOA and AOA Signals

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Muideen Adegoke,
  • Chi-Sing Leung,
  • John Sum

摘要

Mobile source positioning/localization estimation is a crucial problem with wide-ranging applications. Currently, numerical estimation algorithms primarily rely on a single type of measurement method, such as time-of-arrival (TOA) or angle-of-arrival (AOA). This article introduces a hybrid approach that combines AOA and TOA measurements to achieve robust source localization, particularly when there are outlier noises. Instead of using traditional numerical methods, we explore the utilization of neurodynamics models for estimating the location of a mobile source. Our formulation considers several key issues. To handle non-Gaussian noise (outlier noise), we incorporate \(\ell _1\) -norm and \(\ell _2\) -norm loss functions. We first formulate the problem as a constrained optimization problem with inequality and equality constraints. Remarkably, we demonstrate that the inequality constraints can be eliminated. However, the non-differentiability of the \(\ell _1\) -norm loss function poses implementation challenges for the neurodynamics equations. To overcome this obstacle, we propose a smooth approximation for the \(\ell _1\) -norm loss component. Furthermore, we conduct a stability analysis of our proposed neurodynamics model. Finally, we present simulation results to evaluate the performance of our method. These results confirm the superiority of our proposed neurodynamics model over several numerical methods, particularly in scenarios with high measurement noise.