Vehicle Traffic Management and Analysis

Minor Project
Traffic Jams have been significant problems in Kathmandu valley especially at traffic junctions and crossroads. Efficient traffic management systems have to be implemented at such places. But, traffic lights on most road intersections operate on a fixed timing schedule that lead to unnecessary delays, deadlock situations and lead to higher fuel consumption. Our goal is to design a system that can detect the the flow of traffic across those junctions, analyze those data, and create an optimal traffic light sequences that can best counteract that situation.

We focused our study in one of most congested traffic junctions - Koteshwor, Jadibuti and Lokanthali, where we often see huge traffic jams and even deadlocks.
The street lighting system is based on the electronic controller that utilizes the traffic density survey data. Data were collected and analyzed at different busy junctions of Kathmandu Valley. Cameras were used to capture video of the specific junction at various times, which was then processed to get the vehicle flow of the corresponding time. The data gives an insight into the number of vehicles entering the junction and the time required to cross it. This is helpful to calculate the stoppage time which was programmed into the system for optimized and efficient traffic management.
Designed Hardware
Deadlock Situtation
                                                        in Jadibuti Junction
Project consisted of six major dissections- Data Collection, Modeling, Data Analysis & Predictions, Hardware Design & Fabrications, WebApp Development, and Testing and Optimization. My roles were solely in WebApp development with few collaborative contributions to Data Collections & Analysis and in Simulations.
1. WebApp
WebApp was individually handled by me to simulate as traffic control panel with purposes as:
  • Monitor Purpose
    • To monitor dynamic traffic simulation based on time sequence
    • To monitor currently operating phase
    • To watch live stream video of the junction
  • Control Purpose
    • To control and reset the phase in case of emergency
WebApp included the features such as:
Only Authenticated Access
The access to web application is possible by only the authenticated user preventing access, modification, traffic hijackings.
Time Basis Simulation
The WebApp could show the current junction, with accessible phase in green color and blocked phases in red color. Also, phase reset action is available to reset the phase in case of emergency. Chart volume density on daily or weekly basis can also be observed
Chart density
Live Stream
Live Stream video from IP camera connected to the junction can be monitored.
Livestreaming of
2. Traffic Data Collections & Analysis & Optimization
Traffic Data were collected in Koteshwor and Jadibuti junctions. Following things were done at the beginning of the project:
  • Existing flow of vehicles per unit time (Vehicle Density)
  • Dimensions of the road
  • Number of junctions
  • Existing Lane rules
  • Existing rules for maximum and minimum speed limits
These manually collected data were taken as a starting point to visualize and record the existing traffic sequence. The traffic sequence was divided into three phases in each junction, whose opening and closing times were recorded. Being a tedious method, manual data collection was just performed as a starting point. Video Analysis method using YOLO (You Only Look Once) backed up by SORT (Simple Online Real Time Tracking) algorithm was used later for data collection
Manual Data Collection
                                                    of Koteshwor and Jadibuti
3. Simulation Using PTV Vissim
To test the efficiency of the timing sequence generated by processing the data collected the simulation software PTV Vissim was used.
3D visualization of
                                                    Simulation of Jadibuti

Team Members