Designing a web-based multi-modal route planner for Edinburgh
MetadataShow full item record
This paper presents a web-based multi-modal journey planner which assists a large proportion of bus users in Edinburgh to find the optimal route based on their specified starting and end location. Acquired data from the open source database NPTDR contains all of the relevant data for research on Edinburgh bus route. To simplify the network modeling, only bus service of Lothian buscompany was chosen. With a three-tier architecture implementation, DBMS in the designed route planner facilitates the spatial and attributes query during the navigation process. In this project, a relational database was created to store and retrieve all the research data. Numerous studies identify that the peculiarities of bus network lead to the complexity of bus network modeling, which make the route finding process challenging as well. A schedule-based and transfer-based path finding algorithm for the transit network was integrated into the route finding process. Then a selection of best route criteria was utilization to filter the inferior route from the candidate routes. When it comes to the user interface, the designed route planner enables people to interact with the itinerary planning by specifying the origin/destination locations and expected departure day and time. Through the processing of the algorithm, the interface could return the user the suggested route, the bus stop location, as well as the departure time and the estimation of journey times of every trip by retrieving timetable data and calculating based on an average walking speed. In addition to showing more sophisticated real time information, this route planner also displays vivid map images to the users. However, both strengths and drawbacks were identified in the evaluation. Evaluated by users and by comparison with routing results using some other approaches, this project managed to construct the proposed route planner to implement a whole routing process. Inevitably, routing performance is not so satisfying because of the existence of potential errors. Some further improvements should be made for more user interactions or the availability of multiple criteria routing scenarios in the near future.