ComfRide

A Video Presentation

Commuter Survey

We have conducted a large scale online survey involving more than 300 participants across the globe, who mostly access the public transport system. Responses were obtained from different developed and developing countries including Italy, UK, Netherlands, Norway, Germany, France, USA, India, Nepal, Iran, Pakistan, Sri Lanka and Vietnam. Around 25% participants are female and are aged between 20 to 65 years. In our survey, we design the questionnaire to assess factors like total travel time, traffic congestion, sitting probability, road condition, bus type (AC or non-AC buses), number of stops, break journey, which reflect factors regulating commuter comfort and their interplay for a comfortable journey.

In a nutshell, this survey reveals that there exist multiple parameters (Figure 3), viz., speed of the bus, congestion, probability of getting a seat, etc, which impact commuters’ comfort during the trip, and individuals differ widely in their perception of comfort primarily depending on their age and gender. Moreover, the trend is quite similar all over the world

ComfRide captures diverse features through smartphones, which impact the commuter’s comfort level while using public transport and develops a personalized route recommender employing the fuzzy set theory along with TOPSIS approach, which considers individual comfort level based on spatio-temporal road and route characteristics. Besides this diverse and wide suite of factors across various possible routes, single or multiple breakpoints in the journey is also considered, if that increases comfort.

In order to capture the spatio-temporal dynamics over a wide choice of features, we design the route recommendation algorithm using a specialized compositional model, called Dynamic Input/Output Automata (DIOA). The DIOA ensures that the system is not overwhelmed with the huge amount of data to be processed thus reducing the load on the system. Effectively, ComfRide utilizes DIOA based compositional model to identify the most preferable route efficiently based on the historical information and the context of the query. Moreover, the DIOA also provides a mechanism to dynamically modify the model to suit the personalized preference of a commuter on obtaining a feedback from her after a trip, thus improving the quality of recommendation after every trip.

System Description

The overall system can be divided into two broad modules (Figure 5) – (a) the client module (the smartphone app) which runs the client automaton and (b) the server module (runs over a remote server) that runs the server automaton with various filters as internal signatures. Additionally, ComfRide uses a database to store the historical information that are used by the server module to apply various filters using the internal signatures.