Intelligent Intersections in Various Vehicle Traffic Networks and the Resulting Changes in Network Efficiency
CSEF · 2007 Mathematics & Software Honorable_mention Award
Overview
Objectives/Goals This experiment is designed to determine if street intersections can learn to increase the efficiency vehicle transport in a network over time. Methods/Materials A laptop and Visual Studio for the C# language were used to create a computer simulation from scratch. The simulation allows the user to #draw# a road network where vehicles are added to the network according to a pattern that models real life. The intersections store data about how many cars are waiting in each direction every time the configuration of the traffic lights changes. Intersections use this data to determine if certain lights should stay green for more or less time in future cycles. In each test, a percentage value is set which determines how drastically the intersections respond to the data they record. Tests were applied in simple road networks with both four-way and three-way intersections. Results Due to favorable values for deviation between trials, the experiment has yielded relatively firm results. The experiment demonstrated that the intersection logic and learning capabilities initially benefited intersections by increasing the number of cars they could process in a one-hour interval. As the #learning# percentage and the time over which the simulation was run increased, intersections actually lost efficiency. Conclusions/Discussion It was concluded that agents which learn slowly over time, rather than responding very rapidly to their environments, are able to execute their tasks most efficiently over time. This behavior - learning best at a modest pace - demonstrates that simulated intersections were learning much like humans do. These results have implications for urban planning and artificial intelligence in computer programs.
Summary statement
This experiment uses a computer simulation and the principles of artificial intellegence to determine if road networks can be made more efficient if intersection control systems learn and adapt to traffic patterns automatically.
Awards (1)
- Honorable Mention
Competition history
- CSEF 2007
Resources
Related projects
ISEF · 2018
Creating a Simulation of an Intersection of Self-Driving Cars Without Traffic Lights
ISEF · 2023
Optimizing Traffic Flow: Implementing Neural Networks and Deep RL Machine Learning Algorithms to Make Traffic Management More Effectual
ISEF · 2019
Developing and Simulating Self-Driving Car A.I. for a Crash Free Autonomous Intersection
CSEF · 2017
Traffic Network Management: Novel Approach Using a Deep-Learning System Driven by Dynamic Data and an Adaptive Algorithm
CSEF · 2018
A Computer Traffic Simulation: Vehicle Throughput across a Grid of Intersections Varying Driver and Car Parameters
CSEF · 2006
Finding Efficiencies of 2 and 1-way Road Networks of Various City Block Sizes and Traffic Levels using Computer Modeling
JSHS · 2020
Developing Deep Learning Networks for Dynamic Traffic Light Control
ISEF · 2015
Self-Driving Car Intersection
Closest projects by meaning, across every fair and year in the corpus.
Browse more like this
Source: California Science & Engineering Fair public projects