WirelessDT is a Wireless Digital Twin Platform for Wireless Software Applications (WSAs) Evaluation.
In this ICSE2023 DEMO Video, we evaluate a wireless indoor localisation mobile application with two typical prediction algorithms: 1) Kalman Filter-based Trilateration and 2) Deep Recurrent Neural Network, as a case study to demonstrate the capabilities of WirelessDT.
Wireless technology has become one of the most important parts of our daily routine. Besides being used for communication, the wireless signal has been applied to various Wireless Software Applications (WSAs). The signal fluctuation caused by the measurement system or the environmental dynamic can lead to nonnegligible influences on WSA performance. However, quantitative and qualitative real-time evaluation of WSAs is challenging due to limited resources and intractable practical issues. To overcome these challenges, we propose WirelsssDT, a wireless digital twin platform, using digital twin and real-time ray tracing technologies to emulate the signal propagation and generate emulation data for real-time WSAs evaluation.
NOTICES: WirelessDT can only be executed on Window Systems with NVIDIA RTX GPU. The suggested GPU hardware is RTX3070 or RTX4090.
To use WirelessDT, we need to clone both WirelessDT and the WiTracing Unreal Engine (UE) repositories.
Unreal Engine is a private project and is only visible to its subscribers. First, We need to associate the GitHub account with our UE account. For the step for account association, please refer to How do I link my Unreal Engine account with my Github account?
After associating the accounts, we should have access to our Unreal Engine Fork with our built-in WiTracing Engine. Then you can clone the engine code from dev/ue5-witracing branch.
git clone https://github.com/codelzz/UnrealEngine.gitAfter cloning the UE source code, you can refer to the official page Downloading Unreal Engine Source Code for engine setup. After setup the UE source code, we need to compile it to generate the binary files.
git clone https://github.com/codelzz/WirelessDT.gitAfter the project is cloned, we need to compile the project.
- Right click the
.uproject, then selectSwitch Unreal Engine Version..., and choose the one withSource build ..... This will associateWirelessDTwithWiTracing Unreal Engine. - Right click the
.uprojectagain and selectGenerate Visual Studio project file. The solution file end with.slnshould appear. - Open the
.slnfile and compile the project.
After compiling, we can double-click the .uproject file to launch the platform. The DemoOffice_30mx30m map is the default game level which can be executed without Physical Twin. (You can change to the other map from the content fold window under All>Content>Demo>Map. The DemoCorridor_Calibration level is the calibration environment that required Physical Twin.)
Now, you can click Play button in the Editor to start the platform.
We provide configuration of WiTracing Engine through Unreal Engine Editor Command Line Interface. Press <`> button to activate the Command Line Interface, Then type in the prefix r.WiTracing. for available commands:
| Command | Description | Type |
|---|---|---|
| r.WiTracing.TXPower | The transmit power of the transmitter (default 25 dBm) | Int |
| r.WiTracing.Wavelength | The wavelength of the transmit signal (default 0.125 m for 2.4 GHz) | Float |
| r.WiTracing.RXSensitivity | The receiver sensitivity (default -95 dBm) | Int |
| r.WiTracing.SamplePerPixel | The number of WiTracing samples for each pixel per frame | Int |
| r.WiTracing.MaxBounce | The maximum number of bounces for each WiTracing ray | Int |
To change the endpoint for data sharing, we can configure the setting of IP and Port in WiTracingAgent from the UE Editor Outliner window. The data is transmitted in JSON format.
The number of auto-spawn AI can be changed by the Number of AIs in the AISpawn actor. This will impact the number of randomly walking characters in runtime.
The path of the target AI carrying the wireless receiver can be adjusted by changing the trajectory of the AISpline Actor.
Besides the above settings, we open-source our project and provide total freedom for other researchers or developers to customize their virtual environment with UE Editor.
The physical twin hardware and the localization application are required to reproduce the demo.
Here is detailed settings and environment:
| Component | Hardware |
|---|---|
| Wireless Sensor | Seeed Studio XIAO nRF52840 |
| Location Sensor | Intel® RealSense™ Tracking Camera T265 |
| Mainboard | Raspberry Pi 4 Model B |
| Power Supply | 5V/3A Power bank |
The python script for physical twin state synchronization is located at \Source\WiTracingHCI . To run the hardware, we need to upload the script in Raspberry Pi 4 Model B, then execute it by
sudo python app.py(The Arduino program for XIAO nRF52840 is in \Source\WiTracingHCI\WiTracingBLE\RX)
| Component | Setup |
|---|---|
| System | Windows 10 Pro |
| Processor | Intel(R) i7-10870H CPU |
| RAM | 16.0 GB |
| GPU | NVIDIA GeForce RTX 3070 Laptop |
| Parameter | Value |
|---|---|
| Wireless Signal Wavelength | 0.125m (e.g. 2.4GHz) |
| Wireless Transmit Power | 25 dBm |
| Wireless System Gain | -45 dB |
| Wireless Receiver Sensitivity | -95 dBm |
| WiTracing Resolution | 16,384 Pixel (128x128) |
| Maximum WiTracing Sample Per Pixel | 1 |
| Maximum WiTracing Bounces | 3 |
To clone the repository.
git clone https://github.com/codelzz/WirelessDTApp.gitThe application has been tested on iPhone 12 Pro Max with iOS 16.1
- If you encounter any issues while executing, please check your GPU driver whether is appropriately installed. To verify the driver, you can create an empty unreal engine project with ray-tracing enabled and run it to see whether the issue still exists.