Autonomous-Leveling

Overview

A simulation platform for autonomous soil leveling. Noticed since terramechanics model utilizes GPU parallelization tool--CUDA, so it is required to have NVDIA graphic card to run the code. The code has been tested on system: Ubuntu 22.04.5 LTS.

Download Repo

Click Download Repository on the top right corner to download the whole repository.

Usage

Install Chrono Simulation Engine

Noticed you need to download chrono engine with following link https://drive.google.com/file/d/1EuYlXzaYqnN3c9kwzDyVE9AnakoO-Es1/view?usp=sharing instead of official website page of chrono because it has the latest and faster terramechanics development.

After downloading the zip file, upzip it and try to install the simulation engine

cd chrono-engine && mkdir build && cd build

Run ccmake or cmake gui to build and install the chrono engine. you need to enable these following modules enabled: Vehicle, FSI, VSG. The installation process would be similar as described in official Chrono document: https://api.projectchrono.org/tutorial_install_chrono.html.

Build the chrono demo for autonomous leveling

Once the chrono simulation engine is installed, we can link the built chrono engine and build autonomous leveling demo: sim/soil_leveling.cpp. First, we want to get into unziped folder:

cd <path_to_Autonomous-leveling_folder>

Then we want to build the cpp simulation code:

cd sim && mkdir build && cd build && ccmake ../ -G Ninja

Make sure to set chrono path correctly: Chrono_DIR = <path_to_chrono>/chrono-engine/build/cmake. Then build the code by running:

ninja

After successfully building demo, there will be an executable called: soil_leveling_sim in your folder sim/build/

Conda install to run NN based algorithm

create and activate conda environment

conda create -n autograding python=3.10 && conda activate autograding

install the pytorch related package through conda

conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

install other supporting packages

conda install numpy matplotlib

Runing simulation with our autonomous blade control algorithm:

1. make sure you are in the autograding conda environment. Runing algorithm:

python algo/auto_level_sim.py --pile_height 0.37 

2. after blade control algorithm is spinning, start the chrono simulation from another terminal:

cd sim/ && ./build/soil_leveling_sim --pile_height 0.37

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.