A rover that navigates autonoumously, identifies and retrieves samples of interest on a simulated Martian terrain.
The simulator build can be downloaded from these links: Linux, Mac, Windows. You can use training mode and navigate the rover manually without the need of any code.
To setup all dependencies, run the following
cd <PATH TO PROJECT DIRECTORY>/ env
conda env create -f vision-project.yml
source activate vision-project
The rover can then run autonomously by running the following:
cd <PATH TO PROJECT DIRECTORY>/code
python drive_rover.py
After that you can launch the simulator and choose Autonomous mode
You can user dubugging mode by adding the "-debug" argument to the run command.
python drive_rover.py -debug
When debugging mode is running, a window will open next to the simulator showing the different stages of the image processing.
Please note that enabling the debugging mode can decrease FPS and can affect the fidelity.