Automated Contextual Anomaly Detection for Network Interface Bandwidth Utilisation: A Case Study in Network Capacity Management

We present a contextual anomaly detection methodology utilised for the capacity management process of a managed service provider that administers networks for large enterprises. We employ an ensemble of forecasts to identify anomalous network traffic. Stream of observations, upon their arrival, are compared against these baseline forecasts and alerts generated only if the anomalies are sustained. The results confirm that our approach significantly reduces false alerts, triggering rather more accurate and meaningful alerts to a level that could be proactively consumed by a small team. We believe our methodology makes a useful contribution to the applications enabling proactive capacity management.

EFFECT OF THREE-WHEELED VEHICLES ON THE CAPACITY OF A TRAFFIC STREAM

Due to the tremendous development of the number and types of vehicles and a large number of traffic congestions, the phenomenon of the spread of three-wheeled vehicles, which is characterized by ease of movement, has recently appeared in Iraq because of a small space it occupies for its movement. The increase in their numbers exceeds the number of heavy vehicles in most urban areas in Iraq. Therefore, the current study has been devoted to studying the effect of those vehicles with three tires using a simulation program that has been previously developed and has been calibrated to suit the normal sections where the characteristics of these vehicles were included in terms of speed and length. Different ratios were used, ranging from (0-15) % for the first lane only. Some general traffic characteristics were examined, such as velocity-flow and flow-density and density-velocity relationship. The results showed that the presence of this type of vehicle has a significant impact on reducing road efficiency and increasing traffic congestion. The study recommends that a particular lane with a width of 2m on both directions of the roads be designated for the movement of these vehicles and keep them away from interference in the traffic flow of other cars.

Deep Learning and YOLOv3 Systems for Automatic Traffic Data Measurement by Moving Car Observer Technique

Macroscopic traffic flow variables estimation is of fundamental interest in the planning, designing and controlling of highway facilities. This article presents a novel automatic traffic data acquirement method, called MOM-DL, based on the moving observer method (MOM), deep learning and YOLOv3 algorithm. The proposed method is able to automatically detect vehicles in a traffic stream and estimate the traffic variables flow q, space mean speed vs. and vehicle density k for highways in stationary and homogeneous traffic conditions. The first application of the MOM-DL technique concerns a segment of an Italian highway. In the experiments, a survey vehicle equipped with a camera has been used. Using deep learning and YOLOv3 the vehicles detection and the counting processes have been carried out for the analyzed highway segment. The traffic flow variables have been calculated by the Wardrop relationships. The first results demonstrate that the MOM and MOM-DL methods are in good agreement with each other despite some errors arising with MOM-DL during the vehicle detection step due to a variety of reasons. However, the values of macroscopic traffic variables estimated by means of the Drakes’ traffic flow model together with the proposed method (MOM-DL) are very close to those obtained by the traditional one (MOM), being the maximum percentage variation less than 3%.

Determination of passenger car noise equivalent for mid-sized cities in India

The heterogeneity in traffic flow composition increases the complexity of road traffic noise analysis for mid-sized in India. This study aims to determine a passenger car noise equivalent (PCNE) with respect to the average traffic stream speed that represents the number of a particular vehicle category with reference to an identified vehicle based on their noise emission characteristics. In the present study, vehicles are classified as bus, truck, light commercial vehicles (minibus, minitruck), three-wheelers (vikram-rickshaw), two-wheelers (bike/scooter), car, e-rickshaw and auto-rickshaw, and tractor-trailer. Car is taken as a reference vehicle for estimation of PCNE in our study due to its high percentage in traffic stream. Data has been collected on both bituminous and concrete pavement in Kanpur city, India, to analyze the differential effect of pavement on the noise level. As per this study, tractors-trailers, trucks, three-wheelers, and buses had a higher PCNE value, while two-wheelers and cars had almost similar PCNE value. A comparative analysis of PCNE value at concrete pavement is also conducted by considering car running on the bituminous pavement as reference vehicle. The study suggests to employ PCNE value in traffic noise analysis as it converts the divergent traffic volume in terms of the car.

Do Heavy Vehicles Always Have a Negative Effect on Traffic Flow?

The purpose of this study is to analyze the effect of heavy vehicles on traffic flow on a two-lane highway. To achieve this goal, data was obtained from piezosensors on the Seoul–Chuncheon Expressway. Analysis of the data showed that, as everyone knows, the average speed of traffic flows decreases as the proportion of heavy vehicles increases. However, not only the speed decreased, but the speed deviation between vehicles decreased. In other words, it was found that within the traffic group that formed the same platoon, individual vehicles were forced to form similar speeds, resulting in a homogeneous rate. This means that heavy vehicles can be included in the traffic stream, reducing the chances of a vehicle-to-vehicle conflict. This kind of influence can be said to explain that heavy vehicles do not necessarily have a negative effect on traffic flow. In this way, we expect to be able to study ways to manage traffic flow by using the effects of low-speed vehicles.

Lambert W Based Speed Reduction Model in Presence of Pedestrian Movements: Case Studies on Undivided Streets

Pedestrian movements sharing right-of-way with vehicular traffic have adverse impacts on the mobility of the traffic stream. Pedestrian movements both along and across the road often force drivers of approaching vehicles either to stop completely or to slow down and change the existing lane. It ultimately results in a decrease in stream speed. With the aim of determining the influence of pedestrian movements, the present study collected traffic data at a standard section (without pedestrian movements) and both traffic and pedestrian data at a pedestrian section (with considerable pedestrian movements). To determine the speed at the standard section, this paper presents a novel ‘Lambert W function’-based speed prediction model in the context of a two-lane undivided urban road. When stream speeds of the pedestrian section were compared to the stream speeds obtained through the speed prediction model at the same traffic volume condition in absence of pedestrians, a significant reduction was observed. This reduction in stream speed is governed by pedestrian parameters. A new pedestrian parameter ‘lateral position of pedestrian from the edge while walking along the road’ was conceived in this study along with few other parameters to predict Percent Speed Reduction (PSR). Intensities of these pedestrian parameters were observed varying over time which results in a high fluctuation in PSR within a range of 29% to 62%. Finally, this investigation forwards an empirical model of Percent Speed Reduction (PSR) to predict the stream speed in the presence of on-street pedestrian movements on undivided urban roads. The outcome of this paper will help transport planners to estimate the efficiency of pedestrian infrastructure projects before implementation.

Calibration of the empirical fundamental relationship using very large databases

This paper describes a procedure for fitting traffic stream models using very large traffic databases. The proposed approach consists of four steps: (1) an initial treatment to eliminate noisy, inaccurate data and to homogenize the information over the density range; (2) a first fitting of the model, based on the sum of squared orthogonal errors; (3) a second filter, to eliminate outliers that survived the initial data treatment; and (4) a second fitting of the model. The proposed approach was tested by fitting the Van Aerde traffic stream model to 104 thousand observations collected by a permanent traffic monitoring station on a freeway in the metropolitan region of São Paulo, Brazil. The model fitting used a genetic algorithm to search for the best values of the model parameters. The results demonstrate the effectiveness of the proposed approach.