Road Test Experiments and Statistical Analysis for Real-Time Monitoring of Road Surface Conditions Abstract— Road information services (RIS) is a major component of the information and communication technologies with the main purpose of RIS-based systems is to monitor road health conditions, weather information and traffic congestion. Considering the road conditions, there are various kinds of road surface types and anomalies with lack of efficient analysis of their behavior on the vehicle sensor measurements. Consequently, there are difficulties in detecting and categorizing the different road types and anomalies. This paper demonstrates road test results for the measurements of inertial sensors mounted in land vehicles while monitoring various road surface types and anomalies. In addition, a wavelet-based feature extraction together with statistical approach for the road types and anomalies are explored in this study. Two road test experiments on two different vehicles performed in Kingston, ON, Canada together with in-depth analysis are discussed in this paper. Keywords—Road information services; vehicular resources; intra-vehicle sensing; Road anomalies; automotive inertial sensors; statistical analysis; wavelet analysis; I. INTRODUCTION Recently, intelligent transportation systems (ITS) has received a significant amount of interest from the information and communication technologies (ICT) community, and at the industrial and academic level as well. According to a study provided by P&S market research in 2014, it was expected that the Global ITS market to increase from $18M to reach $38M in 2020, with a compound annual growth rate (CAGR) of 13.1% between 2015 and 2020 [1]. The evolution of computers, sensors, control, communications and electronics devices can lead to ITS solutions that save lives, time, money, energy, and the environment. Current computers and communication systems technologies are being developed in a manner to improve transportation around the world. The integration of such systems provides several forms of intelligent links between travelers, vehicles, and infrastructures to eliminate the challenges in transportation [2]. Next generation ITS systems, specifically those involved in road traffic monitoring, will be required to provide reports of traffic congestion, road conditions, and driver behavior. Road surface anomalies, as one of the road conditions indicators, contribute to increased risk of traffic accidents, reduced driver comfort and increased wear of vehicles. As a notable evidence of the anomalies effect, traffic accidents resulted from weather or road conditions are 33% of 2M+ accidents reported by Transport Canada between 2001 and 2011 [3]. Moreover, according to the American Automobile Association (AAA), two thirds of U.S drivers are worried of road surface anomalies, with approximately $3 billion a year in car repairs [4]. There are some attempts for monitoring road surface conditions but most of them were depending on 3 rd party sensing infrastructure and with limitation in information variety. Furthermore, road surface conditions monitored by authorities rely on special instrumentation integrated with specific simulation software or even reported manually [5]. Some research work has addressed (RIS) and road surface anomalies. In [6] a system was developed to monitor braking events and pump detections using accelerometers, GPS and audio sensors of a smart phone. For the event localization, they relied on GSM and GPS. In addition, a system proposed in [7] used accelerometers of smart phone to classify and detect various pothole types with a GPS used for localization. RoADS, a system in [8] used smartphone’s accelerometer, gyroscope and GPS sensors to classify road anomalies into three categories named severe, mild and span. Each category contained events that share similar behavior over the collected data. A back-end server in [9] was used to categorize anomalies in three categories. Data collected by vehicle mounted with accelerometer was sent to the back-end server whenever an internet-based connection is available. Most of the available road surface monitoring systems lack full insight into the required aspects that formulate robust monitoring of anomalies. These systems presented few types of anomalies [6] or just focused in details of one type of anomalies [7]. In addition, some solutions that classify road anomalies lack adequate geo-referencing of the detected events as they rely only on GPS, which may introduce large localizing errors especially at high vehicle speed or in urban canyons [8, 10]. Generally, in urban areas and downtown cores, the localization accuracy greatly deteriorates due to GPS satellite signal blockage and multipath [11, 12]. Amr S. El-Wakeel * , Abdalla Osman † , Aboelmagd Noureldin *†‡ and Hossam S. Hassanein ‡† * Electrical and Computer Eng. Dept., Queen’s University, Kingston, ON, Canada, K7L 3N6 {amr.elwakeel, nourelda} @queensu.ca † Electrical and Computer Eng. Dept., Royal Military College of Canada, Kingston, ON, Canada, K7K 7B4 {Abdalla.Osman, Aboelmagd.Noureldin} @rmc.ca ‡ School of Computing, Queen’s University, Kingston, ON, Canada, K7L 3N6 Hossam@cs.queensu.ca 978-1-5090-5019-2/17/$31.00 ©2017 IEEE