Model-Based Systems Engineering - Simulating a futuristic package delivery service using drones.

This project have been done as a part of the course 02223 - Model-Based Systems Engineering at DTU in the Fall of 2022

A video of the running simulator can be found here.

Running the simulator

Before running the project please install the Dependencies and look at the config section

python3 simulation.py 

Dependencies

To run the project please install the following packages by using pip or pip3:

pip3 install pygame==2.1.3.dev8 pygame_gui dependency_injector configparser scikit-learn sklearn numpy pandas matplotlib pandas

Using the UI

• By pressing "U" will zoom in

• by pressing "D" will zoom out

• The arrow keys can be used to explore the environment.

Config

The main config file config.ini contains all the global parameters. the file is grouped by sections '[section]'.

graphics [section]

• scale - the scale the simulation starts at.
• fullscreen (1 = fullscreen, 0 = window)
• window_h (height in px of window)
• window_w (width in px of window)

setup [section]

Simulation environment parameters

• world_size

• ground_size

• road_size=1

• customer_density

  Be aware that unsupported configs does exist. If the simulator can not make a meaningful environment it will simply close right after it starts.

Please look at the section Neighborhood (env) structure to understand the parameters.

Simulation parameters

• number_of_tasks=50 (value from 0-1000, number of packages that is in the truck)
• number_of_drones=10 (value from 0-1000, The number of drones there should be simulated)
• simulation_speed=200 (1-500) (1 = realtime speed, 500 = as fast as the computer can go.)
• auto_close_window=1 (1 or 0, if 1 the window will close when the simulation is done.)
• drone_speed = 200 (value from 50-500)
• truck_speed = 150 (value from 50-500)
• customer_density=0.5 (the possibility of a house owner in the neighborhood to be a customer awaiting a package)
• truck_stop_density=0.05 (the percentage of the travel distance that the truck should consider as a stop pos)
• min_package_weight=200 (the minimum weight of the generated packages)
• max_package_weight=2000 (the maximum weight of the generated packages)

Options for a truck:

• moving_truck = 1 -> the truck moves
• moving_truck = 0 and random_position = 0 -> stationary truck with the optimal position
• moving_truck = 0 and random_position = 1 -> stationary truck with the random position

Definitions (Conventions of the project)

ENV service (env_service.py)

Creates the 2D representation of a neighborhood

Neighborhood (env) structure

• Road : "R"
• Ground: “.”
• Delivery spot: "S"
• Truck: "T"

For instance will the input (world_size=15, ground_size=5, road_size=1, customer_density=0.5) result in the neighborhood below. This representation is called a layout in the code. Please change the parameters in the main config file.

(layout, delivery_sports, number_of_grounds, number_of_customers), truck_pos = create_layout_env(15, 5, 1, 0.5)

 R  R  R  R  R  R  R  R  R  R  R  R  R  R  R  R 
 R  .  .  .  .  R  .  S  .  .  R  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  .  T  .  .  .  .  R  .  .  .  .  R 
 R  S  .  .  .  R  .  .  .  .  R  .  .  S  .  R 
 R  R  R  R  R  R  R  R  R  R  R  R  R  R  R  R 
 R  .  .  S  .  R  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  R 
 R  R  R  R  R  R  R  R  R  R  R  R  R  R  R  R 
 R  .  .  .  .  R  .  .  S  .  R  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  R 
 R  R  R  R  R  R  R  R  R  R  R  R  R  R  R  R 

A layout can be printed out byt using print_layout(layout).

truck path finding

Applying the path-finder to the layout above gives following route.

• Start point = (0,0)

• End point = (len(layout) - 1, len(layout) - 1))

  planner = PatchFinder()

  route = planner.find_path(layout, (0, 0), (len(layout) - 1, len(layout) - 1))

The route is marked with the character "M".

# shown in layout
 M  M  M  M  M  M  R  R  R  R  R  R  R  R  R  R 
 R  .  .  .  .  M  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  .  M  .  .  .  .  R  .  .  .  .  R 
 R  .  .  .  S  M  .  .  .  .  R  S  .  .  .  R 
 R  .  .  .  .  M  .  .  .  .  R  .  .  .  .  R 
 R  R  R  R  R  M  M  M  M  M  M  R  R  R  R  R 
 R  .  .  .  .  R  .  .  .  .  M  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  M  .  .  .  .  R 
 R  .  .  .  .  R  .  .  .  .  M  .  .  .  .  R 
 R  .  .  .  .  R  .  .  S  .  M  .  .  .  S  R 
 R  R  R  R  R  R  R  R  R  R  M  M  M  M  M  M 
 R  .  .  .  .  R  .  S  .  .  R  .  .  .  .  M 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  M 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  M 
 R  .  .  .  .  R  .  .  .  .  R  .  .  .  .  M 
 R  R  R  R  R  R  R  R  R  R  R  R  R  R  R  M 

Definitions

Neighborhood (env) structure

• Road : "R"
• Ground: "dot"
• Delivery spot: "S"
• Truck path : "M"

Data (Experiments)

All data produced is in the folder 'simulation_batcher'.

In this folder each scenario has its own folder with the raw log files. Futhermore, a summary of the runs has been compiled in to a CSV file with the same name.

Plots

Plots can be regenerated from the data by running the following files (with the path to the csv files updated):

• plot1.py
• plot2.py
• plot3_analysis.py
• plot4_analysis.py