Optimized Target Parameter Estimation for Massive MIMO Imaging Radar Systems = 대규모 다중 입출력 영상 레이더 시스템에서 타겟의 최적 파라미터 추정 기법

다국어 초록 (Multilingual Abstract)

Autonomous driving systems require the integration of reliable sensor systems to precisely perform decision-making and control functions based on detection information of the surrounding environment. Among these, radar sensors are gaining attention as a key sensor due to their robustness against adverse weather conditions and varying lighting environments. However, radar sensors have a major limitation in that they offer lower angular resolution compared to LiDAR or camera sensors. This dissertation addresses these challenges by designing imaging radar systems based on various waveform and proposing techniques that enhance target parameter estimation performance using massive multiple-input multiple-output (MIMO) array.
First, to effectively operate a large number of transmit and receive antenna elements, a hybrid MIMO scheme combining time-division and Doppler-division multiplexing is proposed, along with a code-division multiplexing strategy that enables efficient use of time-frequency resources in phase-modulated continuous-wave radar systems. In addition, a sensing algorithm that enhances the angular estimation performance of targets is proposed by utilizing large antenna aperture implemented by MIMO scheme. In particular, the dissertation analyzes the spatial-wideband effect occurs in high-resolution radar systems and verifies the effectiveness of compensation techniques through experimental results. Finally, to enhance radar sensing performance using communication-based waveform for sharing sensor-based perception results with surrounding vehicles, the method that enables more accurate estimation of target parameters is proposed with low computational complexity in orthogonal time-frequency space radar systems. These proposed techniques represent essential signal processing algorithms for various waveform considered in next-generation radar systems, contributing to improved target sensing performance and enhanced system stability.

국문 초록 (Abstract)

자율주행 시스템은 주변 환경에 대한 탐지 정보를 활용한 판단 및 제어 기능을 정밀하게 수행하기 위하여 신뢰성 있는 센서 시스템의 탑재를 요구하며, 그중에서도 레이더 센서는 악천후나 조도 변화에 강인한 핵심 인지 수단으로 주목받고 있다. 하지만, 레이더 센서는 라이다 혹은 카메라 센서 대비 낮은 각도 분해능을 주요한 한계점으로 갖는다. 본 학위 논문에서는 이러한 문제를 극복하기 위해 다양한 신호파를 기반으로 한 영상 레이더 시스템을 설계하며, 대규모 다중-입출력 배열을 기반으로 최적의 타겟 파라미터를 추정하는 기법들을 제안한다.
먼저, 다수의 송수신 안테나 요소를 효과적으로 운용하기 위해, 시간 분할과 도플러 분할 다중화를 결합한 하이브리드 방식을 제안하며, 위상 변조 연속파 레이더 시스템에서 시간-주파수 자원을 효율적으로 활용하는 코드 분할 다중화 전략을 제시한다. 추가적으로 해당 다중-입출력 안테나 전략을 활용하여 넓은 안테나 면적을 확보하였을 때, 높은 대역폭을 사용하면서 발생하는 광대역-공간 효과가 레이더 시스템에 미치는 영향을 분석하고, 이를 보상하는 제안 기법을 실측 결과를 통해 검증한다. 마지막으로, 주변 차량과의 센서 기반 센싱 결과를 공유하는 통신 신호파 기반 레이더 센싱 성능을 향상시키기 위하여, 직교 시간-주파수 다중 분할 방식을 활용하는 레이더 시스템에서 낮은 연산량 기반으로 보다 정확한 타겟 파라미터를 추정하는 기법을 제안한다. 해당 제안 기법들은 차세대 레이더 시스템에서 고려되는 신호파별 필수적인 신호처리 알고리즘들로, 타겟의 센싱 성능을 높이고 레이더 시스템의 안정성을 높이는데 기여할 것으로 기대된다.

목차 (Table of Contents)

• 1 Waveform Consideration for Imaging Radar Systems, 1
• 2 Hybrid TDM-DDM MIMO Scheme in FMCW Radar, 5
• 2.1 Introduction, 5
• 2.2 Signal Model of Range and Velocity Estimation, 7
• 2.3 Signal Model of Proposed Hybrid TDM-DDM Scheme, 8
• 2.3.1 Design Strategy, 8
• 2.3.2 Signal Model of Proposed Scheme, 9
• 2.4 Novel Algorithms for Proposed Hybrid TDM-DDM Scheme, 11
• 2.4.1 Clutter Removal Method, 12
• 2.4.2 Detection and Peak Matching in R-v Domain,13
• 2.4.3 Resolving Doppler Ambiguity by Ambiguity Test, 14
• 2.4.4 Doppler Phase Compensation and 2D-Angle Estimation, 15
• 2.5 Pointcloud Refining, 17
• 2.5.1 Post-Processing for Outlier Removal, 17
• 2.5.2 Multipath Removal Algorithm Using DoA-DoD Map, 18
• 2.6 Computation Complexity Analysis, 21
• 2.7 Experiments, 23
• 2.7.1 Calibration, 23
• 2.7.2 Experiment Environment, 23
• 2.7.3 Results, 24
• 2.8 Conclusion, 25
• 3. Fast-Time CDM-MIMO Scheme in PMCW Radar, 26
• 3.1 Introduction, 26
• 3.2 Signal Model of PMCW Radar, 27
• 3.2.1 Range-Velocity Estimation, 27
• 3.2.2 Received Signal Model of MIMO PMCW Radar, 29
• 3.2.3 Matched Filter-Based Demultiplexing, 30
• 3.2.4 DoA Estimation with Ideal Correlation Function, 31
• 3.3 Correlation Property of Various Binary Sequences, 32
• 3.3.1 Several Sequences for PMCW Radar, 32
• 3.3.2 Pulse-Compressed Signal of CDM-MIMO PMCW Radar, 33
• 3.4 Sparse Signal Representations, 35
• 3.4.1 Problem Formulation, 35
• 3.4.2 Conventional Successive Interference Cancellation, 36
• 3.4.3 Conventional Linear Estimators, 37
• 3.5 Proposed Compressive Sensing-Based Self-Code Interference Cancellation, 39
• 3.5.1 Property of Sensing Matrix, 39
• 3.5.2 Sparsity-Based Compressive Sensing Methods, 40
• 3.5.3 Fast Algorithm Using Property of Circulant Matrix, 42
• 3.6 Results, 43
• 3.6.1 Simulation Setup, 43
• 3.6.2 Convergence Test, 44
• 3.6.3 Performance Evaluation, 47
• 3.6.4 SIngle- and Multi-Target Scenarios, 49
• 3.7 Conclusion, 50
• 4. Spatial-Wideband Effect Compensation in High Resolution Massive MIMO Radar Systems, 51
• 4.1 Introduction, 52
• 4.2 System Model, 53
• 4.2.1 MIMO FMCW Radar Signal Model, 53
• 4.3 Spatial-Wideband Effect in FMCW Radar, 55
• 4.3.1 Signal Model Considering Spatial-Wideband Effect, 55
• 4.3.2 Spatial Effect: Range Migration, 56
• 4.3.3 Wideband Effect: Beam Squint, 57
• 4.3.4 Angle Estimation Performance Analysis, 58
• 4.4 Proposed Method to Compensate Spatial-Wideband Effect, 59
• 4.4.1 Proposed Method: FG-SWC, 60
• 4.4.2 Proposed Method with Reduced Computations: CG-SWC, 61
• 4.4.3 Performance Degradation due to Coarse Angle Grid, 63
• 4.4.4 Analysis for Computational Complexity, 64
• 4.5 Experiment Results, 66
• 4.5.1 Experimental Environment,66
• 4.6 Conclusion, 69
• 5. Enhanced Radar Target Parameter Estimation in OTFS Radar Systems, 70
• 5.1 Introduction, 70
• 5.2 SISO OTFS Radar Signal Model, 72
• 5.2.1 Transmit Signal Model, 72
• 5.2.2 Receiving Signal Model of Conventional 2D-FFT Approach, 73
• 5.2.3 MF Approach for Reduced-CP OTFS System, 74
• 5.3 Signal Model of SIMO OTFS Radar,76
• 5.3.1 Signal Model Without Consideration of Beam Squint Effect, 76
• 5.3.2 Signal Model with Consideration of Beam Squint Effect, 76
• 5.4 Proposed Low-Complexity MF for R-v Estimation, 78
• 5.4.1 Total Search, 80
• 5.4.2 Partial Search, 81
• 5.4.3 Fine-Grid Partial Search, 82
• 5.4.4 Examples, 85
• 5.5 DoA Estimation with Complexity Relaxation, 86
• 5.5.1 Coarse R-v Estimation with Low-Complexity MF, 86
• 5.5.2 Coarse DoA Estimation, 86
• 5.5.3 Proposed DoA Estimation Considering Beam Squint Effect, 86
• 5.5.4 DoA Estimation Performance Analysis, 87
• 5.6 Improvement of Proposed Method in Presence of Range Estimation Error, 88
• 5.6.1 Range Error due to Off-Grid Problem, 88
• 5.6.2 Range Error Compensation and DoA Estimation, 88
• 5.7 Results, 90
• 5.7.1 Simulation Setting, 90
• 5.7.2 Extension of Proposed Method to Multi-Target Scenario, 90
• 5.8 Conclusion, 90
• 6. Conclusion 92
• Abstract (In korean), 103