CONGESTION DETECTION BASED ON ROAD DENSITY PARAMETERS AND VEHICLE SPEED PARAMETERS

• Main Author: Enjat Munajat - NIM 33212016 , M.D.
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/28821
• Institution: Institut Teknologi Bandung

<p align="justify">Congestion is still an important issue that is happening up to now, and continues to be the research of experts to find solutions in the process of delivering information to users. The research ever conducted focuses on congestion, more separately discusses methods for detecting vehicles and methods of determining street areas but not explicitly addressing street congestion. Methods for determining existing congestion are more likely to see congestion conditions that are based on the number of vehicles passing through time units or are based on monitoring of vehicle speed slowing which is easily defined as a congestion condition. Whereas congestion information is also determined by a combination of information related to the level of street density and average speed over a certain period of time. Other research, focusing on the use of GPS technology and sensors (detector) for the passing vehicle detection process and vehicle speed in units of time. However, the technology is still quite expensive and requires user participation to obtain data and information, while still having limitations on actual actual distance. <br /> <br /> <br /> Based on this reasoning, a method that is capable of displaying congestion conditions and information is factual, reliable and does not require user participation. In this paper, we discussed two new methods in the context of congestion detection using street level density and vehicle speed level. The street density level defines how large the street is occupied by the vehicle by calculating the percentage of the road pixel percentage comparison of the pixel of vehicles passing through time units and specific observation areas. Rate the speed of the vehicle, see the average speed of a vehicle that passes on a particular frame in a certain time unit and on a particular observation area. The speed reading process is done after first determining the center of each moving object (vehicle) and taking into account the degree of slope of each moving object (vehicle). It is also done the process of determining the closest distance of the vehicle on each frame so that vehicle tracking is maintained and the speed becomes factual. <br /> <br /> <br /> This combination of two information helps define the level of vehicle congestion on a street. To strengthen the argument of the congestion condition is calculated using Fuzzy, considering the level of congestion cannot be measured exactly, so that the information obtained can be more accepted by the user. Speed validation is done by measuring the velocity of the sample object that passes through the observed region with the rate of speed set in such a way among the four-wheeled vehicles. Based on the recording, the system is very good at analyzing the speed of moving objects. Based on the 4 videos used as samples, the accuracy of speed-readings reached up to 93%. For density level validation, use comparison of visual observation results and system observation results. Observations made from light conditions, jammed to heavy-jammed. The system proved able to respond to highway density results compared to visual observations. So with the two approaches, it is considered that this method is very good for defining the level of congestion. To strengthen the argument, an analysis was conducted to measure the ‘influence of the relationship’ on average vehicle speed on the street density level. The influence of the relationship measured from the negative to the strong positive. The results of the reading on the system show the accuracy rate of the relationship speed and density reached 100% accurate, which then re-validated with manual calculations. <br /> <br /> <br /> Still a discussion for further research is the conditions of night lighting and camera cross-sectional conditions are difficult to stabilize. This condition still leads to errors in detecting moving objects resulting in incorrect reading of the congestion level. <p align="justify"> <br />