Simulation of Graph SLAM in 2D environment using Python. The 2D world is generated with specified parameters like world size and number of objects or landmarks in that environment. Landmarks are generated randomly inside 2D world with defined number of those landmarks, then Robot is placed to move and sense in environment with defined number of steps. Robot is measuring distances to landmarks inside it's measurement range and that data with Robot motion actions is the input for SLAM algorithm which is mapping landmarks and robot positions in a 2D world. SLAM results are estimated Robot and Landmarks positions. For more details see Notebook.
Installation
- Clone repository on your computer:
$ git clone https://github.com/TomislavZupanovic/GraphSLAM-2D.git
- Create and activate virtual environment (for Windows):
$ python -m venv slam
$ slam/Scripts/activate
- Install requirements:
$ pip install -r requirements.txt
- For Linux/MacOS see: Virtual Environments
Run SLAM
- Navigate to repository:
$ cd .../GraphSLAM-2D
- Run script:
$ python run.py --steps 20 --num_landmarks 10
There are options for other parameters but best if default:
--world_size --measurement_range --measurement_noise and --motion_noise
Example results
