Paper: AirFogSim: A Light-Weight and Modular Simulator for UAV-Integrated Vehicular Fog Computing

Abstract:

Vehicular Fog Computing (VFC) is significantly enhancing the efficiency, safety, and computational capabilities of Intelligent Transportation Systems (ITS), and the integration of Unmanned Aerial Vehicles (UAVs) further elevates these advantages by incorporating flexible and auxiliary services. This evolving UAV-integrated VFC paradigm opens new doors while presenting unique complexities within the cooperative computation framework. Foremost among the challenges, modeling the intricate dynamics of aerial-ground interactive computing networks is a significant endeavor, and the absence of a comprehensive and flexible simulation platform may impede the exploration of this field. Inspired by the pressing need for a versatile tool, this paper provides a lightweight and modular aerial-ground collaborative simulation platform, termed AirFogSim. We present the design and implementation of AirFogSim, and demonstrate its versatility with five key missions in the domain of UAV-integrated VFC. A multifaceted use case is carried out to validate AirFogSim’s effectiveness, encompassing several integral aspects of the proposed AirFogSim, including UAV trajectory, task offloading, resource allocation, and blockchain. In general, AirFogSim is envisioned to set a new precedent in the UAV-integrated VFC simulation, bridge the gap between theoretical design and practical validation, and pave the way for future intelligent transportation domains.

Citation

If you use AirFogSim in your research, please cite our paper:

@misc{wei2024airfogsimlightweightmodularsimulator,
      title={AirFogSim: A Light-Weight and Modular Simulator for UAV-Integrated Vehicular Fog Computing}, 
      author={Zhiwei Wei and Chenran Huang and Bing Li and Yiting Zhao and Xiang Cheng and Liuqing Yang and Rongqing Zhang},
      year={2024},
      eprint={2409.02518},
      archivePrefix={arXiv},
      primaryClass={cs.NI},
      url={https://arxiv.org/abs/2409.02518}, 
}

License

This project is licensed under the MIT License - see the LICENSE file for details.

How to use Real Data?

First, you need to set the sumo-export_tripinfo to True in config.yaml, and set the tripinfo_output to the path you want to save the tripinfo_output.xml file. The code is like:

sumo:
  sumo_config: "./sumo_wujiaochang/osm.sumocfg"
  sumo_osm: "./sumo_wujiaochang/osm_bbox.osm.xml"
  sumo_net: "./sumo_wujiaochang/osm.net.xml"
  sumo_port: 8813
  export_tripinfo: True # 如果true，则导出tripinfo.xml文件（很大）
  tripinfo_output: "./sumo_wujiaochang/tripinfo.xml"

Then, you can run the simulator to generate the tripinfo_output.xml file. After that, you should transform the tripinfo_output.xml file to the tripinfo_output.csv file. The code is like:

(airfogsim) (base) weizhiwei:~/data/airfogsim_code/$ python airfogsim/utils/xml2csv.py ./sumo_wujiaochang/tripinfo.xml

This will generate the tripinfo_output.csv file in the same directory. Then, you can use the real data in the simulator via yaml settings. The code is like:

traffic:
  traffic_mode: "real" # "real" or "SUMO"
  tripinfo: "./sumo_wujiaochang/tripinfo.csv" # The path to the SUMO tripinfo file

Setup:

1. Install SUMO (tested version is 1.15.0), and set the Environment variable. Once suceed, enter sumo in command line as:

(airfogsim) (base) weizhiwei:~/data/airfogsim_code/$ sumo

there should be:

Eclipse SUMO sumo Version 1.15.0
 Build features: Linux-5.4.0-131-generic x86_64 GNU 9.4.0 Release FMI Proj GUI Intl SWIG GDAL GL2PS Eigen
 Copyright (C) 2001-2022 German Aerospace Center (DLR) and others; https://sumo.dlr.de
 License EPL-2.0: Eclipse Public License Version 2 <https://eclipse.org/legal/epl-v20.html>
 Use --help to get the list of options.

2. Create the simulation environment (suggested to use conda for virtual environment) according to requirements.txt:

pip install -r requirements.txt

3. Run example01_offloading_example.py. Here is the main code:

# 1. Load the configuration file
config_path = 'config.yaml'
config = load_config(config_path)

# 2. Create the environment
env = AirFogSimEnv(config, interactive_mode='graphic')
# env = AirFogSimEnv(config, interactive_mode=None)

# 3. Get algorithm module
algorithm_module = BaseAlgorithmModule()
algorithm_module.initialize(env)
accumulated_reward = 0
while not env.isDone():
    algorithm_module.scheduleStep(env)
    env.step()
    accumulated_reward += algorithm_module.getRewardByTask(env)
    print(f"Simulation time: {env.simulation_time}, ACC_Reward: {accumulated_reward}", end='\r')
    env.render()
env.close()

4. If another version of SUMO is used, please generate SUMO scenarios via the sumo_dir/tools/osmWebWizard.py and change related paths in config.yaml. Tested graphic videos:

Use Berlin as the map:

airfogsim._.mp4

Use 五角场 in Shanghai as the map:

wujiaochang.mp4

5. For more personalized developing guidance, we attempt to build AirFogSim Assistant-GPTs. However, GPT may make mistakes, remember to check the code.

Benchmarks

Refer to the benchmarks directory for more details.

Paper Code

Refer to the paper_code directory for more details.