SY
S. Yuan
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
1 records found
1
Nowadays, autonomous driving has been researched for decades, especially with advanced automotive radar. However, there is no common agreement on the best approach for simulating real-world scenarios of automotive radar. In the academic field of automotive radar research, one of the main obstacles for advanced algorithm development, for instance those based on machine learning, is the lack of real data for developing such algorithms and training them to ensure generalization capabilities.
As a strong supplement to experimental measurements, radar signal simulation has made great progress, keeping pace with advances in computation and software capabilities. With this progress, high-fidelity models have become an essential aspect of radar simulation for achieving realistic results with the help of powerful computers. However, to the best of the author’s knowledge, there is no extensive research combining high-fidelity target models with conventional radar simulation techniques.
Therefore, this thesis analyzes recent radar simulation techniques and proposes a pipeline for an FMCW MIMO automotive radar signal simulator for realistic extended targets and their movement in automotive scenes of interest. Within the proposed simulator, a motion control coordinate system is introduced, allowing high freedom in the selection of the target model. The shadowing and occlusion conditions are also investigated for single- and multiple-target cases to increase the realism of the simulation.
In the end, through analysis of the simulated results with preset ground truth from CAD models, the simulator is shown to be capable of simulating realistic extended targets with high flexibility for the user. ...
As a strong supplement to experimental measurements, radar signal simulation has made great progress, keeping pace with advances in computation and software capabilities. With this progress, high-fidelity models have become an essential aspect of radar simulation for achieving realistic results with the help of powerful computers. However, to the best of the author’s knowledge, there is no extensive research combining high-fidelity target models with conventional radar simulation techniques.
Therefore, this thesis analyzes recent radar simulation techniques and proposes a pipeline for an FMCW MIMO automotive radar signal simulator for realistic extended targets and their movement in automotive scenes of interest. Within the proposed simulator, a motion control coordinate system is introduced, allowing high freedom in the selection of the target model. The shadowing and occlusion conditions are also investigated for single- and multiple-target cases to increase the realism of the simulation.
In the end, through analysis of the simulated results with preset ground truth from CAD models, the simulator is shown to be capable of simulating realistic extended targets with high flexibility for the user. ...
Nowadays, autonomous driving has been researched for decades, especially with advanced automotive radar. However, there is no common agreement on the best approach for simulating real-world scenarios of automotive radar. In the academic field of automotive radar research, one of the main obstacles for advanced algorithm development, for instance those based on machine learning, is the lack of real data for developing such algorithms and training them to ensure generalization capabilities.
As a strong supplement to experimental measurements, radar signal simulation has made great progress, keeping pace with advances in computation and software capabilities. With this progress, high-fidelity models have become an essential aspect of radar simulation for achieving realistic results with the help of powerful computers. However, to the best of the author’s knowledge, there is no extensive research combining high-fidelity target models with conventional radar simulation techniques.
Therefore, this thesis analyzes recent radar simulation techniques and proposes a pipeline for an FMCW MIMO automotive radar signal simulator for realistic extended targets and their movement in automotive scenes of interest. Within the proposed simulator, a motion control coordinate system is introduced, allowing high freedom in the selection of the target model. The shadowing and occlusion conditions are also investigated for single- and multiple-target cases to increase the realism of the simulation.
In the end, through analysis of the simulated results with preset ground truth from CAD models, the simulator is shown to be capable of simulating realistic extended targets with high flexibility for the user.
As a strong supplement to experimental measurements, radar signal simulation has made great progress, keeping pace with advances in computation and software capabilities. With this progress, high-fidelity models have become an essential aspect of radar simulation for achieving realistic results with the help of powerful computers. However, to the best of the author’s knowledge, there is no extensive research combining high-fidelity target models with conventional radar simulation techniques.
Therefore, this thesis analyzes recent radar simulation techniques and proposes a pipeline for an FMCW MIMO automotive radar signal simulator for realistic extended targets and their movement in automotive scenes of interest. Within the proposed simulator, a motion control coordinate system is introduced, allowing high freedom in the selection of the target model. The shadowing and occlusion conditions are also investigated for single- and multiple-target cases to increase the realism of the simulation.
In the end, through analysis of the simulated results with preset ground truth from CAD models, the simulator is shown to be capable of simulating realistic extended targets with high flexibility for the user.