HATTO-UFog: Holistic Attitude–Trajectory–Task Optimization for UAV-Assisted Fog Computing

HATTO-UFog is a UAV-assisted fog computing system that jointly optimizes: (1). Attitude control (2). Trajectory planning (3). Resource allocation (4). Task assignment / offloading to reduce end-to-end latency and energy consumption in a terrain-aware 3D environment. The code for the paper "Energy-Aware Holistic Optimization in UAV-Assisted Fog Computing: Attitude, Trajectory, and Task Assignment".

• Project website: https://shuaijun-liu.github.io/HATTO-UFog.

• Online demo: https://shuaijun-liu.github.io/UAV-Assisted-Fog-Computing-Simulation-Demo.

Note: The online demo is a lightweight, browser-based (Three.js) interactive visualization created for the paper to help reviewers quickly understand the simulation environment in a simplified setting.

Modules

• Framework/: paper-aligned Python simulation framework for Trajectory Planning (adapters: ACS, ACS_DS, CPS_ACO, GA_SCA, TD3) + Attitude Control (FEAR-PID) + Communication + Resource Allocation + Task Offloading, with deterministic Parquet+JSON logging. (Framework/README)
• AirSim/: AirSim-based validation and video capture for Framework outputs (mainline/auxline). Included Unreal scenes: AirSimNH, AbandonedPark, LandscapeMountains, Blocks. (AirSim/README)
• Isaac/(Isaac Sim): higher-fidelity dynamics and disturbance robustness experiments, plus Framework→Isaac replay integration and postprocess tooling. (Isaac/README)

Note: the environment boundary, Framework / AirSim / Isaac have different dependencies and launch modes; follow each module README instead of assuming a single end-to-end command.

Showcases

AirSim - AirSimNH Scene

More showcases in Section Modules.

Example result videos (MP4): AirSimNH · Blocks · AbandonedPark · LandscapeMountains

Isaac Sim

Example result video (MP4).

Framework showcase assets (web-based)

Example result video (MP4).

Quickstart

Option A: pip

From HATTO-UFog/:

python -m pip install -r requirements.txt

If you use the showcase renderer (Playwright):

python -m playwright install chromium

Option B: conda (YAML env)

From HATTO-UFog/:

conda env create -f environment.yaml
conda activate hatto-ufog
python -m pip install -r requirements.txt

Python reference env (toolchain-aligned):

conda env create -f environment-python313.yaml
conda activate hatto-ufog-py313
python -m pip install -r requirements.txt

Start from the framework

Framework (from HATTO-UFog/Framework):

python -m pip install -e .
ufog_network run --config configs/smoke.yaml --output runs

Framework (from HATTO-UFog/):

python -m pip install -r requirements.txt
ufog_network run --config Framework/configs/smoke.yaml --output Framework/runs

Environment

These version pins are a reference baseline for reproducing the reported runs and the curated media artifacts.

Component	Item	Value
Isaac Sim	Version	5.1.0-rc.19+release.26219.9c81211b.gl
Isaac Sim	Built-in Python	3.11.13 ($ISAACSIM_ROOT/python.sh)
Isaac Sim	Official stage asset prefix	Assets/Isaac/5.1/...
AirSim	Vendored PythonClient upstream commit	13448700ec2b36d6aad7a4e0909bc9daf9d3d931 (microsoft/AirSim)
AirSim	Binary environments	AirSim release v1.8.1 (e.g., AirSimNH.zip, Blocks.zip)
GPU / Driver	NVIDIA driver	580.119.02
GPU / Driver	GPU	NVIDIA RTX 5880 Ada Generation
Python	Key packages	numpy 2.4.1, pandas 2.3.3, pyarrow 22.0.0, opencv-python 4.11.0, PyYAML 6.0.3, imageio 2.37.2, imageio-ffmpeg 0.6.0, playwright 1.49.1, msgpack-rpc-python 0.4.1

Demo Gallery

Web-based Demo for Trajectory Planning

Web-based Demo for Attitude Control

AirSim - AbandonedPark Scene

AirSim - LandscapeMountains Scene

AirSim - Blocks Scene