This paper presents an opportunistic sensing based solution to detect crowdedness in public transportation buses. The solution uses data of accelerometer and Global Positioning System (GPS) sensors available in smartphones carried by the commuters. These data are used to accurately identify bus boarding event and whether a commuter got a seat during his/her trip. The solution is energy efficient as it uses power hungry GPS very conservatively and keeps it off majority of the times.

The solution is evaluated using data collected over the arterial roads of Ahmedabad and Gandhinagar city. The length of routes varies from 25 to 45 kilometers. The effect of application penetration on crowdedness detection in buses is also evaluated. It is found that the penetration of 8 to 12% in commuter population can detect the crowdedness for more than 80% of route segments on the test routes. Further, the solution results in the energy-saving of about 50% compared to a solution that requires GPS data continuously. We also present the bus scheduling scheme that uses the historical data of bus-crowdedness to schedule the feeder buses on the crowded segments of the route.

本文提出了一种基于机会感知的解决方案，以检测公共交通公交车中的拥挤状况。该解决方案使用通勤者携带的智能手机中的加速度计和全球定位系统（GPS）传感器的数据。这些数据用于准确识别公交车上车事件以及通勤者在旅途中是否就座。该解决方案具有很高的能源效率，因为它非常保守地使用耗电的GPS，并且大多数时候都不会使用它。

使用在艾哈迈达巴德和甘地纳加尔市干道上收集的数据对解决方案进行评估。路线长度从25到45公里不等。还评估了应用程序渗透对公交车拥挤检测的影响。发现通勤人口中8％至12％的渗透率可以检测出测试路线上80％以上的路段拥挤情况。此外，与连续需要GPS数据的解决方案相比，该解决方案可节能约50％。我们还提出了公交车调度方案，该方案使用公交车拥挤的历史数据在路线的拥挤路段上调度接驳公交车。