Simulating sensor noise model for real-time testing in a virtual underwater environment
Junwen Deng, Andrew Filisetti, Hui Sheng Lim, Du Yong Kim, Akram Al‐Hourani · 2024
Given the remote and harsh deployment environments associated with developing Autonomous Underwater Vehicles (AUVs), virtual simulations are necessary to mitigate such risks and facilitate AUV research. To increase the accuracy of such simulations, authentic noise models for AUVs are required to produce reliable results. To fill in this knowledge gap, we propose a simulation framework that integrates MATLAB & Simulink with Unreal Engine 4, which enables researchers to emulate AUVs and test the noise models in realistic underwater settings. In order to validate this framework, we simulate an example AUV with LIDAR and vision sensor and present sensor output after injecting a desired sensor noise model. As such, this framework can facilitate emulating, building and verifying AUV platforms with different sensor noise models, thus leading to an improvement of performance in the current virtual simulation of AUVs.