Advanced-driver assistance systems (ADAS) have given rise to circuit technology with new signal processing algorithms, machine learning, artificial intelligence, and computer-vision techniques that have made self-driving cars a reality. The use of perception sensors enforced with artificial intelligence has significantly increased in the automotive industry, with obstacle detection being a significant challenge for level 5 autonomy.This study addresses the economic factor of testing these algorithms on a vehicle. Making a physical prototype for every iteration will be economically challenging and delay the development of a completely autonomous vehicle. One solution to this is testing these algorithms and sensors in a simulated environment, giving us near-accurate results using graphical engines like the Unreal Engine (UE). Unreal Engine being open-source with predefined functions to create a realistic simulated environment, can also import Geographic Information System (GIS) data to develop a 1:1 scale terrain. Several simulators are trying to achieve these accurate simulations, but the problem arises as only a handful remain free of proprietary content. Some open-source simulators like AirSim have a highly precise LiDAR (Light Detection And Ranging) model, but little focus has been given to the RADAR (Radio Detection And Ranging) models, creating a need for the development of sufficiently accurate radar models. MATLAB/SIMULINK has one of such near-exact RADAR model under its automated driving toolbox. However, it is licensed, use of this model is restricted due to proprietary conditions.A LiDAR sensor creates a point cloud of the surrounding environment by emitting laser lights and calculating the time of reflection the light takes to return. However, a LiDAR is more prone to be affected by weather conditions such as rain and dust. Similarly, RADAR uses frequency modulated continuous wave(FMCW), which is not affected by these conditions but is prone to be affected by a different set of environment variables, e.g. temperature. Even though the distortion affects the sensors' detection, their functionality remains similar, allowing us to leverage a LiDAR model with an appropriate distortion model to simulate RADAR detection giving it a more comprehensive application range.This research focuses on creating a radar model on UE using blueprint visual scripting in the interest of off-road vehicles and validating the results by comparing them with the results achieved from the radar model from the automated driving toolbox of MATLAB/SIMULINK. This model will give the developers a radar platform to build upon, which could significantly help develop a radar sensor for automotive applications.

RADAR Modeling for Autonomous Vehicle Simulation Environment With Open-Source Utilities

Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2021/volume/85543

ASME 2021 International Mechanical Engineering Congress and Exposition (IMECE2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.