Comfort on board of public buses plays an essential role in the level of satisfaction perceived by bus passengers. In order to increase the number of passengers on public transport, discomfort should be minimized. It is clear that discomfort and transport sickness is caused by bumps and in turns. However, the relationship between the magnitude of the lateral acceleration (when the bus is turning) and the feeling of discomfort is less clear. The overall purpose of this study is to find a method which can automatically identify when and where a person will feel discomfort during transportation. This study focuses on the lateral impact on passengers, as several studies have previously analyzed the impact of passengers over bumps.
Passenger comfort is investigated on a selected bus route in Aalborg. The route was traversed by a number of test persons, while the lateral acceleration was recorded with an smartphone accelerometer. The test persons indicate when they feel discomfort during the trip by pressing a button on a separate smartphone.
The data processing has shown that unpleasant events cannot be reliably found from lateral peak accelerations alone. It turns out that the greatest lateral accelerations occurred outside the events. Thus, it is necessary to look for more complex patterns in the data.
The causes of discomfort were found by analyzing notes from the data collection and pictures of event areas taken from Google Maps. The analysis shows that the road surface and road design is often to blame. The causes of events seem to be bumps, turns, cars parked along the route, curbstones, drains, and uneven roads with potholes.
The results also indicate that the feeling of discomfort depend on where the test person is looking during the bus ride. If the test person was very focused on the app, events were entered faster. This was evident when bumps or bends came as a surprise, whereas test persons who tried to take in the surroundings were better prepared for a turn or the route’s alignment. In turn, the tolerance of discomfort seems to be higher for these persons. These factors affect the number of events for particular test persons. One test person is reported as many as 53 discomfort events, while another only reported 25 events for the same route. The average of all six test persons was 35 events.
The main takeaways from this study is thus twofold: Automatic detection of discomfort events is not as easy as looking at peak lateral acceleration, but analyzing discomfort events by asking passengers is a viable way of optimizing bus routes for comfort.