This is the code release for our CoRL 2022 Long Horizon Planning Workshop paper:

Comparison of Model Free and Model-Based Learning-Informed Planning for PointGoal Navigation
Yimeng Li*, Arnab Debnath*, Gregory J. Stein, Jana Kosecka
George Mason University

Project website

@inproceedings{li2022comparison,
  title={Comparison of Model Free and Model-Based Learning-Informed Planning for PointGoal Navigation},
  author={Li, Yimeng and Debnath, Arnab and Stein, Gregory J and Kosecka, Jana},
  booktitle={CoRL 2022 Workshop on Learning, Perception, and Abstraction for Long-Horizon Planning}
}

git clone --branch main https://github.com/yimengli46/bellman_point_goal.git
cd  bellman_point_goal
mkdir output

We use python==3.7.4.
We recommend using a conda environment.

conda create --name lsp_habitat python=3.7.4
source activate lsp_habitat

You can install Habitat-Lab and Habitat-Sim following guidance from here.
We recommend to install Habitat-Lab and Habitat-Sim from the source code.
We use habitat==0.2.1 and habitat_sim==0.2.1.
Use the following commands to set it up:

# install habitat-lab
git clone --branch stable https://github.com/facebookresearch/habitat-lab.git
cd habitat-lab
git checkout tags/v0.2.1
pip install -e .

# install habitat-sim
git clone --recurse --branch stable https://github.com/facebookresearch/habitat-sim.git
cd habitat-sim
pip install -r requirements.txt
sudo apt-get update || true
# These are fairly ubiquitous packages and your system likely has them already,
# but if not, let's get the essentials for EGL support:
sudo apt-get install -y --no-install-recommends \
     libjpeg-dev libglm-dev libgl1-mesa-glx libegl1-mesa-dev mesa-utils xorg-dev freeglut3-dev
git checkout tags/v0.2.1
python setup.py install --with-cuda

You also need to install the dependencies:

habitat==0.2.1
habitat-sim==0.2.1
torch==1.8.0
torchvision==0.9.0
matplotlib==3.3.4
networkx==2.6.3
scikit-fmm==2022.3.26
scikit-image
sknw
tensorboardX

To set up lsp-accel, begin by installing Eigen.
Download the Eigen source code, unzip it, and place it in the lsp_accel directory.
The folder structure should resemble the following:

bellman_point_goal
  └── lsp_accel
        └── eigen

Next, navigate to the bellman_point_goal directory and run:

pip install lsp_accel/

Download scene dataset of Matterport3D(MP3D) from here.
Upzip the scene data and put it under habitat-lab/data/scene_datasets/mp3d.
You are also suggested to download task dataset of Point goal Navigation on MP3D from here
Unzip the episode data and put it under habitat-lab/data/datasets/pointnav/mp3d.
Create soft links to the data.

cd  bellman_point_goal
ln -s habitat-lab/data data

The code requires the datasets in data folder in the following format:

habitat-lab/data
  └── datasets/pointnav/mp3d/v1
       └── train
       └── val
       └── test
  └── scene_datasets/mp3d
        └── 1LXtFkjw3qL
        └── 1pXnuDYAj8r
        └── ....

The code supports
(a) Point Goal Navigation on MP3D test episodes.
All parameters are managed through the configuration file core/config.py.
When initiating a new task, create a new configuration file and save it in the configs folder.

Before executing the demo, download the pre-generated scene_maps, scene_floor_heights, large_scale_semantic_maps, and point_goal_episodes from this link.
Unzip the file and place the folders under bellman_point_goal/output.

The trained learning module is available for download here.
Unzip it and place the folders under bellman_point_goal/output.

To run the demo code of an optimistic planner:

python demo_bellman_point_goal.py

For the LSP-UNet demo, comment out lines 28 and 29 of demo_bellman_point_goal.py, then run:

python demo_bellman_point_goal.py

For the LSP-GT demo, comment out lines 32 and 33 of demo_bellman_point_goal.py, then run:

python demo_bellman_point_goal.py

Demo Results

(b) Large-Scale Evaluation
Initiating the evaluation is a straightforward process. Follow these steps:

1. For desktop evaluation of the optimistic planner, use the following command:

python main_eval.py --config='exp_90degree_Optimistic_PCDHEIGHT_MAP_1STEP_500STEPS.yaml'

2. If you're working with a server equipped with multiple GPUs, choose an alternative configuration file:

python main_eval_multiprocess.py --config='large_exp_90degree_Optimistic_PCDHEIGHT_MAP_1STEP_500STEPS.yaml'

Feel free to customize configurations to meet your evaluation requirements.
Configuration files are provided in the configs folder, following this naming convention:

• Files starting with large_exp run complete PointGoal navigation testing episodes.
• Files with only exp run the first three episodes of each test scene.
• Optimistic in the title signifies running the optimistic planner, while DP signifies running the Bellman Equation formulated point goal navigation.
• NAVMESH uses an oracle mapper, and PCDHEIGHT builds the map on the fly.
• UNet_OCCandSEM_Potential means the potential is computed by UNet with inputs of both occupancy and semantic maps.
• GT_Potential means using ground-truth potential values.
• 500STEPS signifies a maximum of 500 allowed steps.

To generate training data, run the following command:

python data_generator_for_point_goal.py --j=1

You can customize the training hyperparameters using the configuration file exp_train_input_partial_map_occ_and_sem_for_pointgoal.yaml.
Here are a few key options:

• Set the SPLIT parameter to either train or val to generate data for training or validation scenes.
• Adjust PRED.PARTIAL_MAP.multiprocessing to single or mp for single-threaded or multithreaded generation, respectively.

Run the following command to initiate training:

python train_UNet_input_partial_map.py

Customize training hyperparameters using the same exp_train_input_partial_map_occ_and_sem_for_pointgoal.yaml configuration file.
Key options include:

• Adjust BATCH_SIZE, EPOCHS, and NUM_WORKERS based on your computer hardware and GPU memory.