Scholarlink

<p>Manoeuvring delay is one of the critical measures to evaluate the performance of unsignalized intersections on urban roads in mixed traffic conditions, because vehicles entering the conflicting area without following the priority rules due to a lack of awareness and inadequate enforcement of traffic rules. Accurate estimation of manoeuvring delay is essential to evaluate the performance of urban unsignalized intersections in mixed traffic conditions with aggressive driver behaviour. The present study assesses the Manoeuvring delay by developing multiple non-linear regression models (MNLR) and the recurrent neural network (RNN) family models to enhance the prediction accuracy. The models were developed and tested with data collected from four four-legged, unsignalized intersections in Vishakhapatnam. The models are validated using performance metrics such as R<sup>2</sup> value, root mean squared error (RMSE), and mean absolute error (MAE). The results showed that Bidirectional Long Short-Term Memory (BiLSTM) and Bidirectional Gated Recurrent Units (BiGRU) models outperform for all the vehicle types with an average RMSE and MAE of 5.6&#xa0;s and 3.8&#xa0;s across all the movements, and the MNLR model has reasonably good performance with an average RMSE and MAE of 6.2&#xa0;s and 4.6&#xa0;s, respectively. Further, this study compares the proposed models with the existing HCM method and Chandra's (2009) model to evaluate their effectiveness in predicting manoeuvring delay. The results show that the proposed models estimate manoeuvring delay more accurately for various vehicle types across different movements and are useful to improve the performance of unsignalized intersections.</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

Analysis of manoeuvring delay at urban unsignalized intersections under mixed traffic conditions using mathematical and RNNs models

  • 期刊论文

  • 发表日期：

    2025年10月10日

查看全文
  • Borigarla Barhmaiah,
  • V. Sunitha,
  • S. Moses Santhakumar

摘要

Manoeuvring delay is one of the critical measures to evaluate the performance of unsignalized intersections on urban roads in mixed traffic conditions, because vehicles entering the conflicting area without following the priority rules due to a lack of awareness and inadequate enforcement of traffic rules. Accurate estimation of manoeuvring delay is essential to evaluate the performance of urban unsignalized intersections in mixed traffic conditions with aggressive driver behaviour. The present study assesses the Manoeuvring delay by developing multiple non-linear regression models (MNLR) and the recurrent neural network (RNN) family models to enhance the prediction accuracy. The models were developed and tested with data collected from four four-legged, unsignalized intersections in Vishakhapatnam. The models are validated using performance metrics such as R2 value, root mean squared error (RMSE), and mean absolute error (MAE). The results showed that Bidirectional Long Short-Term Memory (BiLSTM) and Bidirectional Gated Recurrent Units (BiGRU) models outperform for all the vehicle types with an average RMSE and MAE of 5.6 s and 3.8 s across all the movements, and the MNLR model has reasonably good performance with an average RMSE and MAE of 6.2 s and 4.6 s, respectively. Further, this study compares the proposed models with the existing HCM method and Chandra's (2009) model to evaluate their effectiveness in predicting manoeuvring delay. The results show that the proposed models estimate manoeuvring delay more accurately for various vehicle types across different movements and are useful to improve the performance of unsignalized intersections.