Urban Traffic Control

Vehicle routing applied on problematic regions.

Table of Contents

1. About The Project
   • Results
   • Built With
2. Getting Started
   • Prerequisites
   • Installation
3. Traffic Scenario
4. Usage
   • Description
   • Classes
   • Example

About The Project

This project builds upon previous works by introducing "problematic" region identification on which vehicle routing is applied. These regions are considerably smaller and with the combination of previously developed network simplification techniques (edge simplification, TopKA*, DBSCAN route clustering) allows us to effectively work with larger scenarios, such as Dublin and Luxembourg.

Results

This sections describes how results were generated and how to reproduce them, note that the simulator SUMO might give slightly different results depending on your computer and SUMO version.

Folder containing results can be located in the folder root: results_timeout, where scenarios for both Dublin and Luxembourg are generated in each folder, which denotes the timeout of the planner. All these folders have results in terms of spreadsheet, that compares them to the base version (not routed) and zip file that has the new routed scenarios, those can be put inside the scenarios folder (where the "base" non-routed scenarios are located) and directly launched with SUMO. They also contain configuration file (JSON) that can be used to reproduce the results, for more information refer to routing where planning of scenarios is explained more in detail.

(back to top)

Built With

• gcc / g++
• python3.9
• Cmake / Make (for planner installation)

Python libraries: (version can be found in requirements.txt file)

• pip
• wheel
• setuptools
• numpy
• matplotlib
• prompt_toolkit
• traci
• sumolib

(back to top)

Getting Started

To get a local copy up and running follow these example steps.

Prerequisites

1. osmfilter.c (located in UTC/utc/data/osm_filter) has to be compiled

gcc osmfilter.c -O3 -o osmfilter

2. SUMO has to be downloaded and installed along with SUMO_PATH
3. Planner used in this project: Mercury (agile)

Installation

1. Clone the repo

   git clone https://github.com/Matyxus/UTC_Framework

2. In folder where UTC_Framework is located execute:

   pip install -e UTC_Framework

3. Install requirements (in ./UTC_Framework)

   pip install -r requirements.txt

4. Compile planners (expected IPC standard as input -> domain, problem file, result file)
5. Give permission to project enabling it to create, delete: folders, files and execute commands with subprocess library.
6. Execute UTC_Framework/utc/test/test_launcher.py for testing the whole project (checks for missing folders etc.)

(back to top)

Traffic Scenario

SUMO's scenarios consist of following files:

1. ".sumocfg" file (which runs the simulation in SUMO GUI, references other files)
2. ".rou.xml" file (file containing vehicle routes)
3. ".add.xml" file (additional file type, contains vehicles etc.)
4. Optional files:
   1. ".stat.xml" file (contains vehicle statistics, average values of speed, delay, etc.)
   2. ".pddl" problem file (contains the scenario represented in pddl format)
   3. ".pddl" result file (contains the action of objects from pddl problem files, generated by planner)
   4. ".log" file (contains the logging output)
   5. ".json" files (can contain configuration used for this scenario, vehicle statistics, etc.)

Scenarios structure is following (located in data folder):

[scenario]
   └─── additional (contains files associated with ".sumocfg" input -> routes, vehicles, ...)
   └─── config (contains "[scenario].sumocfg" files (usually only single))  
   └─── info (contains confugration of scenario, logging output, etc.) 
   └─── problems* (contains pddl problem files)
   └─── results* (contains results of planner)
   └─── stats* (contains statistics of scenario)
   └─── output* (contains output of planners)

Note: [scenario] is name chosen by user (when generating
planned scenario, new [scenario] name is given by user).
All scenarios have unique name, since it is used as "name-space",
referencing to similary named files e.g. "[scenario].rou.xml", "[scenario].stat.xml", etc...
* Folders denoted by star (*) are created only when their content is created / needed.

(back to top)

Usage

Interaction with the project is defined trough command line by passing pre-defined configuration files, which the user can modify. These configuration files are loaded at the start of the program into global memory and then into specific data classes. Data classes all automatically validating the given input with json schema, so that any invalid types / values and arguments can be detected before the program starts running. All data classes also define default values for all parameters that do not have to be set by user (logging, TopKA*, DBSCAN, ...).

Description

Classes

There are 6 sub-packages in the project:

1. Clustering implements gravitational & similarity clustering
2. Constants contains global constants, including file system
3. Graph models the road network as graph, contains visualization, TopKA* and other algorithms
4. Pddl responsible for planning scenarios
5. Simulator provides support for SUMO's traci and traffic scenarios
6. Utils utility method and classes such as Converter which converts ".osm" files into SUMO's ".net.xml" and others

More in depth explanation of each package can be found in their respective folders, which contain "readme" files.

Example

1. Download maps from OpenStreetMap.
   Save them in UTC/utc/data/maps/osm folder.
2. Convert downloaded maps into '.net.xml' format using Converter.
3. Generate scenario using ScenarioMain.
4. (Optional) Generate sub-graph for road network used in scenario using GraphMain.
   Planners can take long time to generate solutions if road network (or number of cars) is large, for that reason it is advantageous to use smaller networks (sub-graphs).
5. Generate planned scenario using UtcLauncher.
6. Launch scenario and compare vehicle statistics against planned simulation using ScenarioMain.

(back to top)