Challenges in the Interpretation of Crowdsourced Road Condition Data

Nowadays almost everyone has a mobile phone and even the most basic smartphones often come embedded with a variety of sensors. These sensors, in combination with a large user base, offer huge potential in the realization of crowdsourcing applications. The crowdsourcing aspect is especially of interest in situations where users' everyday actions can generate data usable in more complex scenarios. This paper introduces a study that utilizes data collected by smartphone sensors such as an accelerometer and GPS for detecting variations in road surface conditions. The data were collected by a group of users driving on actual roads in western Finland. The paper presents the test setup and preliminary results of the study, including the description of the web user interface used to illustrate the data. Additionally, it provides a discussion on the challenges faced in the implementation of the prototype system and a look at the problems related to the analysis of the collected data. In general, the collected data were discovered to be more useful in the assessment of the overall condition of a road, and less useful for finding specific problematic spots on roads, such as potholes.