Scholarlink

<p>Precise object location is the essence of the operation of the contemporary navigation systems that are predominantly based on GPS and INS. The ultimate goal of the proposed work is to predict the accurate position of the objects by improving the performance of the navigation system. The navigation system consists of a global positioning system (GPS) and an Inertial Navigation System (INS). In this work, we have developed a novel Hybrid RNN based Kalman filter. In addition, a Chaotic-Levy-based Particle Swarm Optimization (PSO) enhances the training efficiency of the proposed model. MATLAB is used as the main tool for implementing the proposed model. The hybrid RNN based Kalman filter system, combined with Chaotic-Levy-based PSO, shows excellent performance in predicting object positions accurately. The model achieved a high correlation value of 0.999, a very low Mean Absolute Error (MAE) of 0.03, Mean Absolute Relative Error (MARE) of 0.125, and Mean Squared Error (MSE) of just 0.00139. These results reflect the model’s strong ability to learn from navigation data and make highly reliable predictions across different scenarios.</p>

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

A novel attention based RNN-Kalman filter for accurate position forecasting framework

  • 期刊论文

  • 发表日期：

    2025年7月7日

查看全文
  • N. V. Maheswara Rao,
  • V. B. S. Srilatha Indira Dutt,
  • B. T. Krishna

摘要

Precise object location is the essence of the operation of the contemporary navigation systems that are predominantly based on GPS and INS. The ultimate goal of the proposed work is to predict the accurate position of the objects by improving the performance of the navigation system. The navigation system consists of a global positioning system (GPS) and an Inertial Navigation System (INS). In this work, we have developed a novel Hybrid RNN based Kalman filter. In addition, a Chaotic-Levy-based Particle Swarm Optimization (PSO) enhances the training efficiency of the proposed model. MATLAB is used as the main tool for implementing the proposed model. The hybrid RNN based Kalman filter system, combined with Chaotic-Levy-based PSO, shows excellent performance in predicting object positions accurately. The model achieved a high correlation value of 0.999, a very low Mean Absolute Error (MAE) of 0.03, Mean Absolute Relative Error (MARE) of 0.125, and Mean Squared Error (MSE) of just 0.00139. These results reflect the model’s strong ability to learn from navigation data and make highly reliable predictions across different scenarios.