The task of an inteligent garbage truck is to autonomously gather waste from dustbins. The truck should recognize the types of dustbins, look inside the dustbins and check if the garbage is correctly sorted, empty the bins, move the waste to the landfill and offload it in a suitable place.
To solve this task we used:
pygame
foragent environment representation
BFS
algorith to implementuninformed search strategies
in our projectA*
algorith to implementinformed search strategies
in our projectLearning Methods
:Decision Trees
(we used them to make decision if the garbage truck should take the trash or not by specific arguments like:- Trash size
- Trash weight
- Space on the trash yard
- And much more
Neural Networks
(we used them to recognize type of trash by its image)Genetic Algorithms
(we used it to solveTravelling salesman problem
)
- python
- pygame library (For environment representation)
- sklearn library (For Decision Trees)
- keras library (For Neural Networks)
- Run
main.py
- Press
space
to start the project :) - If project didn't load you need to import some libraries: like
pytorch
,pygame
,sklearn
,keras
and some more. To do it a recommend to use pip eg.pip install torch
orpip3 install torch
.
- Move garbage truck with arrow (
up
to move forward,left
andright
to rotate the agent). - Go to specific square on map by pressing left mouse button and then right mouse button.
- Toggle algorith from
A*
toBFS
by pressing A. - Place obstacles on map by selecting square by left mouse button and then pressing
O
. - Place trash bins on map by selecting square by left mouse button and then pressing
T
. - Run whole app by pressing
space
. - Trash images recognition created with neural networks.
- TSP problem solved with genetic algorith.
- Make decission whether to pick up trash or not by many arguments.
- Nice HUD to show all important informations to user.
- Show trash image when collecting.
Project is: completed
Created by :
2021