Scholarlink

<p>The vehicle weigh-in-motion (WIM) system is an effective means to measure the weight of moving vehicles. To improve the weighing accuracy, this paper proposes a WIM model based on subtraction-average-based optimizer (SABO) algorithm optimized back propagation (BP) neural network. Firstly, the structure and principles of the designed WIM system are introduced. Subsequently, the sampling signals of the WIM system are decomposed and reconstructed using the ensemble empirical mode decomposition (EEMD) algorithm. Then the BP neural network model optimized by SABO algorithm to establish the SABO-BP vehicle WIM model. After parameter configuration and training, the model is utilized to predict the vehicle total weight and axle load. Through comparative analysis in different models, the SABO-BP vehicle WIM model exhibits fast convergence speed and high accuracy, with an average error less than 1 % in predicting total weight. Taking two-axle vehicles as an example, calibration is carried out by comparing our WIM measurement results with static vehicle weight. The experiment results shows that the peaks of the axle load waveform restored by the SABO-BP WIM model are closest to static axle load.</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

The effect of EEMD and SABO-BP algorithms on the accuracy of vehicle weigh-in-motion (WIM)

  • 期刊论文

  • 发表日期：

    2025年7月1日

查看全文
  • Suan Xu,
  • Liang Hong,
  • Lushuai Qian,
  • Kaixing Hong,
  • Yaqiong Fu,
  • Tao Zeng

摘要

The vehicle weigh-in-motion (WIM) system is an effective means to measure the weight of moving vehicles. To improve the weighing accuracy, this paper proposes a WIM model based on subtraction-average-based optimizer (SABO) algorithm optimized back propagation (BP) neural network. Firstly, the structure and principles of the designed WIM system are introduced. Subsequently, the sampling signals of the WIM system are decomposed and reconstructed using the ensemble empirical mode decomposition (EEMD) algorithm. Then the BP neural network model optimized by SABO algorithm to establish the SABO-BP vehicle WIM model. After parameter configuration and training, the model is utilized to predict the vehicle total weight and axle load. Through comparative analysis in different models, the SABO-BP vehicle WIM model exhibits fast convergence speed and high accuracy, with an average error less than 1 % in predicting total weight. Taking two-axle vehicles as an example, calibration is carried out by comparing our WIM measurement results with static vehicle weight. The experiment results shows that the peaks of the axle load waveform restored by the SABO-BP WIM model are closest to static axle load.