Modelica Drone

Scope & Goals

What started as a course project for drone control modeling, has become a larger project on multi-domain modeling for drones.

Scope

We are hoping to continue to develop it as a Drone library, with the ability of modeling multiple domains (e.g. electrical power supply and management, thermal management, VR-based visualization, control modeling with different approaches - continious, discrete, synchronous - etc.).

Goals

To further enahance the models from different domains, with main modeling focus is on the electrical power and controls.
To validate the model against a real prototype. We have procured an FFT Gyro platform to validate the models against a real prototype under controlled conditions.
To establish sponsored research collaborations with funding bodies and private industry in the field. Please contact: luigi.vanfretti@gmail.com

Please cite our work!

If you use these resources, please cite the repository via the DOI:

We have received acceptance of an abstract discussing this model and we are preparing paper, which we would welcome you to cite as:

M. Podlaski, L. Vanfretti, H. Nademi and H. Chang, “UAV Dynamic System Modeling and Visualization using Modelica and FMI,” The Vertical Flight Society's 76th Annual Forum & Technology Display, Virginia Beach Convention Center on Oct. 6-8, 2020.

We will share an open access version of the paper when it is accepted.

Development

Originally, the repository only included a drone model with 3D visualization (enabled by the Multi Body Library) and FMU export configuration developed using Modelica by Hao Chang and Luigi Vanfretti.

Today, the development is coordinated by Meaghan Podlaski, and we have received contributions from students at Rensselaer Polytechnic Institute involved in the Undergraduate Research Program and the course "Modeling and Simulation for Cyber-Physical Systems". Thank you to all that contribute to this fun and exciting effort!

Acknowledgement

This work was funded in whole or part by by the National Aeronautics and Space Administration through the University Leadership Initiative award for the Center for Cryogenic High-Efficiency Electrical Technologies for Aircraft (CHEETA). The first author was supported by Rensselaer Polytechnic Institute through the Rensselaer Graduate Fellowship 2019-2020.

A. Modelica Model

All sub-systems and simulation cases are contained within the ./DroneSimulation/package.mo package. Each model is saved individually as its own file. The Claytex VySeMA and Modelon Libraries are required to run models with non-ideal power systems using batteries and certain DC motor configurations. The full version of the DLR Visualization library is required to run any of the models using VR in the DroneSimulation.Visualization.mo package.

How to simulate it?

Open ./A_Modelica/DroneSimulation.mo
The package structure is set up with the following sub-packages:
- Examples
- Blocks
- Electrical
- Mechanical
- Sensors
- Gyroscope
- Visualization
- Tests
Under the Examples/DroneWithIdealPowerSystem package, open the model DroneTest.mo and select it as Simulation model. This has inputs that act as change in x, y, z coordinates of the drone and outputs x, y, z from the pseudo-GPS modeled within.
Go to simulation tab of your tool and change the simulation time to 10s then click simulate button.
This simulation should result with the z-coordinate approaching 5 meters, and the other coordinates (x,y) should be around zero. To verify, plot the variables .xgps, .ygps and .zgps of the model by running the Modelica script drone_simulation_setup.mos. You should obtain the result below.
Within Dymola, run the script drone_anymation_setup.mos, and then click on the Play button to see the animation. The red arrows indicate the force of the propellers.

B. FMU

Using the distributed FMU

Within ./B_FMI/ an FMU generated by Dymola 2019 FD01 (64-bit) with both ME and CS (using Cvode) options are included. Note that this model requires a Dymola license to execute in your local machine to run.

Generating the FMU under Dymola 2019 FD01 (64-bit)

It is possible to generate an FMU from the Modelica model to provide inputs to x, y and z coordinate changes. Under the 'Test' sub-package, the controlModuleTest_fmu_inputs model can be used. To generate the FMU using Dymola 2019 FD01 (64-bit) under Windows 10, follow the next steps:

Open ./A_Modelica/DroneSimulation/package.mo
Set as Simulation model the model under DroneSimulation.Examples.DroneWithIdealPowerSystem.DroneTests_FMU
Go to the Simulation Setup menu Simulation>Setup... and provide the following configurations in the different tabs shown below:
Under the Simulation mode, go to the menu Simulation>Translate>FMU, and provide the following settings on the Export FMU window:

Alternatively, issue the following command under the Commands window of Dymola:

translateModelFMU("DroneSimulation.Tests.controlModuleTest_fmu_inputs", false, "", "2", "all", false, 1); = "DroneSimulation_Tests_controlModuleTest_0fmu_0inputs"

C. Running the in Simulink using FMIKIT

Install the FMIKit.
Under the folder ./C_SimulinkWithFMIKit a Simulink model ./drone_model.slx set-up using FMIKit to reproduce the simulation scenario explained above.
Note that you might need to reload the FMU using the one in the directory ./B_FMI/ as explained above.
You should obtain the results shown below.

D. Model Visualization and VR

To take advantage of this feature you need the DLR Visualization

Open the models for visualization under DroneSimulation.Examples.Visualization.ModuleTest_SimVis and select it as Simulation model.
Go to simulation tab of your tool and change the simulation time to 100s then click simulate button.
A window for the DLR SimVis application will open with the drone flying in a desert terrain. The drone can be controlled by the keyboard by default, but it can be reconfigured to control the drone with a joystick.

Contributing

Via pull requests.

Copyright

Licensed under the BSD 3-Clause License.

lochel/Modelica-Drone-3D-FMI