Current Science

Kranthi Kumar Reddy ¹, B. Anil Kumar ¹, Lelitha Vanajakshi ¹

Affiliations
1 Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600 036, IN

Abstract

Bus travel times are prone to high variability, especially in countries that lack lane discipline and have heterogeneous vehicle profiles. This leads to negative impacts such as bus bunching, increase in passenger waiting time and cost of operation. One way to minimize these issues is to accurately predict bus travel times. To address this, the present study used a model-based approach by incorporating mean and variance in the formulation of the model. However, the accuracy of prediction did not improve significantly and hence a machine learning-based approach was considered. Support vector machines were used and prediction was done using v-support vector regression with linear kernel function. The proposed scheme was implemented in Chennai using data collected from public transport buses fitted with global positioning system. The performance of the proposed method was analysed along the route, across subsections and at bus stops. Results show a clear improvement in performance under high variance conditions.

Keywords

Bus Travel Time, High Variance Conditions, Prediction Accuracy, Support Vector Machines.

Full Text

   

B. Anil Kumar ¹, Rony Gracious ², Chitrak Gangrade ², Lelitha Vanajakshi ²

Affiliations
1 Department of Civil and Environmental Engineering, IIT Patna, Bihta 801 103, IN
2 Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600 036, IN

Abstract

The computation of point-to-point shortest paths on time-dependent transportation networks has many practical applications. Finding the shortest path on transportation networks, taking into account prevailing dynamic traffic conditions, can help solve the problem of traffic congestion in urban areas. This study presents a framework for implementation of the shortest path algorithm on static as well as timedependent city networks to identify the correct match between network complexity, computational requirements and scalability. Dijkstra, bidirectional A*, and A* with landmarks and triangle inequality (ALT) algorithms were selected and implemented based on their reported good performance in earlier studies. The algorithm implementation on both static and dynamic networks was tested on selected networks from Chennai city, India. Among the tested algorithms, ALT performed the best in terms of criteria used in this study. This algorithm is shown to be scalable and can be implemented for any other city network with ease, as demonstrated in this study. The study also discusses techniques for data extraction, cleaning and representation in addition to implementation and comparison of algorithms.

Keywords

Dynamic Networks, Shortest Path Algorithms, Time-dependent City Networks, Transport Planning, Traffic Engineering.

Full Text

   

Sharmili Banik ¹, B. Anil Kumar ², Lelitha Vanajakshi ¹

Affiliations
1 Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600 036, India, IN
2 Department of Civil and Environmental Engineering, Indian Institute of Technology Patna, Patna 801 103, India, IN

Abstract

Travel time reliability (TTR) is an important measure to quantify the variation in travel times. Currently, there is no single reliability metric appropriate across all locations, that is easily understandable and can be used to compare across facilities. Moreover, reliability analysis of facilities from developing countries is limited due to the non-availability of extensive data required for such an analysis. The present study addresses these gaps. It identifies a reliable data source for such analysis of heterogeneous, lane-less traffic, compares existing reliability measures for the data, highlights the advantages and disadvantages, proposes a measure that may be more suitable for such traffic with high variability, and finally illustrates how reliability analysis under such conditions can be done with limited data sources such as GPS-fitted transit vehicles. Using such commonly available data for traffic stream reliability analysis is the ultimate aim of this study. For validation, stream travel time from Wi-Fi scanners is used. The study analyses the performance of various reliability measures and identifies the most suitable ones. Following this, a reliability measure, i.e. capacity buffer index (CBI), is developed to identify the unreliable congested regimes or periods, keeping time taken to travel at capacity conditions as the benchmark. From the results, it has been observed that CBI is in agreement with the real-field conditions in 94% of the cases, whereas it is 75% buffer time index. Finally, the feasibility of using bus probes to measure stream TTR is checked. Results show that bus probes can be an indicator of stream reliability and the developed measure can effectively capture the relationship between stream and bus TTR

Keywords

Bus probes, contingency tables, mixed traffic, travel time reliability, Wi-Fi sensors.

Full Text