Using In-Vehicle Monitoring Data to Assess Road Conditions of Traffic Flows

Developments in adaptive systems for maintenance and repair of automotive vehicles set the task of monitoring the conditions of their operation. One of the main factors determining these conditions is the type of road surface.The article describes the results of identification of the type (and condition) of the road surface obtained by theoretical and experimental methods based on the analysis of vertical accelerations recorded on the vehicle body.The purpose of research was to provide a possibility of continuous monitoring of the type of road surface on which a vehicle is driving, with the subsequent application of the obtained data to correct maintenance intervals. The results of experiments have shown the dependence of the vertical acceleration of the body on the micro-profile of the road surface. The described experimentally obtained profiles of vertical accelerations refer to different types of road surface in different conditions. For quantitative assessment, it is proposed to calculate the average level of accelerations as an integral average over a certain time interval.The results of the experiments have allowed to substantiate the empirical dependence of the average level of accelerations on speed of a vehicle. Based on this dependence, a method is proposed for recalculating the current values of the average levels of accelerations obtained at different speeds into values adjusted to the base speed to ensure the possibility of their comparison.It is shown that based on the values of average acceleration levels obtained through operation monitoring regarding a previously known type of road surface, it is possible to determine its condition. A short algorithm is formulated for practical implementation and assessment of road conditions of traffic flows. As for hardware, it is proposed not to equip a vehicle with additional sensors but to use operational standard accelerometers as part of in-vehicle emergency call systems, e.g., ERA-GLONASS equipment units.

TRAFFIC FLOWS FORECASTING BASED ON MACHINE LEARNING

The article aims to develop a model for forecasting the characteristics of traffic flows in real-time based on the classification of applications using machine learning methods to ensure the quality of service. It is shown that the model can forecast the mean rate and frequency of packet arrival for the entire flow of each class separately. The prediction is based on information about the previous flows of this class and the first 15 packets of the active flow. Thus, the Random Forest Regression method reduces the prediction error by approximately 1.5 times compared to the standard mean estimate for transmitted packets issued at the switch interface.

Incorporating Delay Minimization in Design of the Optimized Arterial Signal Progression

Despite the abundance of studies on signal progression for arterial roads, most existing models for bandwidth maximization cannot concurrently ensure that the resulting delays will be at a desirable level, especially for urban arterials accommodating high turning volume at some major intersections or constrained by limited turning bay length. Extending from those models that aim to address delay minimization in the progression design, this study provides two enhanced progression maximization models for arterials with high turning volumes. The first model aims to select the signal plan that can produce the lowest total signal delays for all movements from the set of non-inferior offsets produced by MAXBAND. Failing to address the impact of potential turning bay spillback at some critical intersections under such a design may significantly degrade the quality of through progression and increase the overall delay. For this reason, the second model proposed in this study offers the flexibility to trade the progression bandwidths within a pre-specified level for the target delay reduction, especially for turning traffic. The evaluation results from both numerical analyses and simulation experiments have shown that both proposed models can produce the desirable level of performance when compared with the two benchmark models, MAXBAND and TRANSYT 16. The second model yielded the lowest average network delay of 117.2 seconds per vehicle (s/veh), compared with 121.7 s/veh with TRANSYT. Moreover, even its average delay of 141.8 s/veh for through vehicles is comparable with that of 141.2 s/veh by MAXBAND, which is designed mainly to benefit through-traffic flows.

Impact of Shared Lanes on Performance of the Combined Flexible Lane Assignment and Reservation-based Intersection Control

Benefiting from opportunities offered by connected and autonomous vehicles (CAVs), a concept called Combined Alternate-Direction Lane Assignment and Reservation-based Intersection Control (CADLARIC) was proposed recently for management of directionally unrestricted traffic flows in urban environments. In CADLARIC, resolution of vehicular conflicts is distributed between links and intersections to prevent intersections from turning into traffic bottlenecks. Although CADLARIC has shown promising results, it has been observed that, once traffic volume on a certain lane reaches “physical capacity,” adding more traffic on that lane degrades performance of the entire system, as each lane is exclusively dedicated to a particular movement. To overcome this problem, Combined Flexible Lane Assignment and Reservation-based Intersection Control (CFLARIC) is proposed, which offers more flexible lane assignment possibilities. While CFLARIC allows left- and right-turning lanes to be shared with through traffic, it is unclear how much through traffic should be assigned to turning lanes. Thus, this study investigates which strategy is the most beneficial when reassigning extra through traffic to the turning lanes. This goal is divided into two objectives: 1. Identify which lanes should be shared, and 2. Find a close-to-optimal amount of through traffic that should be assigned to the identified shared lane. The proposed CFLARIC strategies are compared with Fixed-Time Control (FTC), Full Reservation-based Intersection Control (FRIC), and CADLARIC for multiple demand scenarios. The results show that the best performing CFLARIC strategies outperform FTC, FRIC, and CADLARIC for delay and number of stops, and reduce the number of conflicting situations compared with FRIC and CADLARIC.

Development of an automated systemfor determining transport flow on video using a neural network

This article discusses the problem of determining traffic flows at an intersection by identifying vehicles using neural networks. A model of a neural network used to detect vehicles in the image is described. The paper describes an algorithm for identifying cars at an intersection. The article also describes an algorithm for obtaining the trajectory of the vehicle.

Analysis of air pollution in Primorsky Krai in 2019-2020 according to GMAO/NASA satellite monitoring

The results of modeling variations in atmospheric pollutants over Primorsky Krai in 2019 and 2020 based on GMAO/NASA satellite monitoring data are analyzed. It is shown that average annual concentrations of pollutants in 2020 decreased as compared to 2019: by 20–35% for sulfur dioxide; by 5–20% for sulfates; by 8–20% for carbon monoxide; by 25–40 % for particulate matter PM (1, 2.5, and 10 μm). One of the reasons for the air pollution decline in Primorsky Krai in 2020 is the reduction of anthropogenic load in the context of a decrease in industrial activity and traffic flows both in Primorye and in the adjacent areas of China. Episodes of high pollution in 2019 were formed under influence of the transboundary transport of polluted air masses. Keywords: air pollution, aerosol and chemical elements, transboundary transport, satellite monitoring, Primorsky Krai

Minimization of traffic delay in traffic flows with coordinated control

The method and results of transport research, carried out by field research method, on the determination of the main indicators of traffic flows with significant unevenness of the movement on the arterial street in conditions of coordinated control is reviewed in the paper. Time parameters of traffic light control for which a reduction in traffic delay is achieved in direct and opposite traffic flow by the change of permissive signal depending on traffic intensity are determined using the simulation method. Change (increase) of the duration of the permissive signal provides uninterrupted movement of vehicles` group during their passage of stop-line at traffic light objects. The proposed method can be used on sections of transport networks with coordinated control, where there is significant heterogeneity of traffic flow, and it prevents the dissipation of groups that consist of vehicles with different dynamic characteristics. Such a result is being performed in the case when in the system of automated control, which combines adjacent intersections on an arterial street, fixed-time program control of traffic light signalization is operating. In this condition, there is a possibility to adjust the duration of signals of traffic light groups by correcting the width (permissive signal duration) and angle of inclination (speed of movement) of the timeline in coordination graphs. The scientific novelty of this research is that the method of traffic delay minimization in conditions of coordinated control acquired further development. The essence of the method is in the controlled change of the range of permissive signal duration in conditions of simultaneous control of the speed of movement between adjacent intersections. Practical value is the application of different programs of traffic light control on sections of arterial streets in transport districts where a significant difference of values of traffic intensity by directions is in morning and evening peak periods.

Assessment of the noise level on arterial streets depending on traffic flow indicators

The intensity of traffic flows increases every year and intensifies its impact significantly on urban territory's ecologic situation and natural environment. In Ukrainian cities, the growth of noise pollution is quite a topical problem as it has a significant impact on the environment because of the increase in car numbers. Traffic flow intensity and densification of the built-up area cause the increase of the noise level. As analytical methods of calculation of determination of traffic noise level have several disadvantages, then for assessing the level of noise on arterial streets of Lviv city, research is carried out with consideration of traffic flow indicators. One of the critical criteria that have an impact on noise pollution in the urban environment is traffic flow composition. Conducted research indicates that the speed of traffic flow creates significant noise pollution, especially on those streets where cobblestone is a road pavement. Freight and passenger transport that moves along arterial streets causes the increase of noise level and increases its impact on the city territory. On sections of arterial streets in peak periods of the day with high values of volume-capacity ratio and general share of freight and public transport of more than 30% in a flow, the noise exceeds the acceptable norms and has a harmful impact on the environment. That is why it is necessary to assess noise pollution and the development of protection measures from noise during the planning and construction of some urban regions or residential areas. It is established that to lower the noise level on streets, it is necessary to consider existing intersections of arterial streets and the regime of movement on them. To lower transport noise, it is necessary to distribute freight and transit transport by city streets more rationally, use qualitative public transport, and limit the maximal speed regime on city streets. Further research on the road network will provide an opportunity to assess the noise level distribution of traffic flows in Lviv city. It will help to determine the problem places and propose methods of combating traffic noise.