Real - Time Visualisation of Traffic Dana Using Speed Transition Matrices

Prometna infrastruktura nije uspjela pratiti nagli rast prometne potražnje tokom godina što je rezultiralo zagušenjima pogotovo u gradskim sredinama gdje je koncentracija prometa velika. Promatranjem i analiziranjem postojećeg stanja prometne mreže moguće je doći do određenih poboljšanja prometovanja pogotovo na kritičnim prometnim pravcima poput povećanja kapaciteta prometnice, optimizacije vremena slobodnog prolaza na semaforiziranim cestovnim segmentima, prevencija zagušenja i povećanje sigurnosti. Zbog toga raste interes za programske alate obrade, vizualizacije i interpretacije prometnih podatka u promatranim vremenskim intervalima. Korištenjem vizualizacije prometnih podataka omogućena je detekcija ključnih prometnih parametara koji opisuju promatrani dio prometnice. Za vizualizaciju prometnih podataka u ovom radu korištene su prijelazne matrice brzina. One su prikaz brzina prometovanja automobila prilikom prelaska između dva povezana cestovna segmenta unutar određenih vremenskih intervala. Ulazni podaci korišteni u ovom radu, generirani su koristeći SUMO simulator. Podaci su podijeljeni u dva skupa gdje jedan prikazuje normalan prometni tok, dok drugi skup podataka sadrži zagušeni prometni tok dijela autoceste. Koristeći MongoDB bazu podataka i programski jezik Python za pred obradu podataka, izračunate su prijelazne matrice brzina. Za određivanje stanja prometne mreže, za svaku matricu je izračunata točka centra mase. Pri analizi matrica izdvojene su one koje prikazuju prometno zagušenje. Ovo programsko rješenje omogućuje uočavanje zagušenja simuliranih u pojedinim prometnim scenarijima i određivanje dovoljno velikog vremenskog intervala u kojem je potrebno prikupljati podatke za kvalitetnu procjenu stanja prometne mreže.Transport infrastructure has failed to keep pace with the rise in traffic demand over the years, resulting in congestion, especially in urban areas where traffic concentration is high. By observing and analysing the existing conditions of the transport network, it is possible to achieve certain improvements in traffic flow, especially on critical traffic routes, such as increasing road capacity, optimizing free passage time on traffic lights, preventing congestion and increasing safety. Therefore, there is a growing interest in software tools for processing, visualization and interpretation of traffic data. Using the visualization of traffic data, it is possible to detect key traffic parameters that describe the observed part of the road. Speed Transition Matrices were used in this thesis to visualize traffic data. They are a representation of the speed of a car during the transition between two consecutive road segments within certain time intervals. The input data used in this paper were generated using a SUMO simulator. The data is divided into two sets, one showing the normal traffic flow, while the other contains the congested traffic flow of a part of highway. Using the MongoDB database and the Python programming language for data pre-processing, speed transition matrices were computed. To determine the state of the traffic network, the point of the centre of mass was calculated for each matrix. In the analysis of the matrix, those that show traffic congestion were singled out. The software solution presented in this thesis enables the detection of congestion simulated in individual traffic scenarios and the determination of a sufficiently large time interval in which it is necessary to collect data for the quality of the assessment of the condition of the traffic network