基于模块级流水线加速高斯-牛顿法的高速LiDAR在线多回波测距方法

摘要

在应急救援和无人驾驶中，全波形激光雷达（FW-LiDAR）通过分解目标散射的多次回波来获取多层次的三维成像。为解决实时多目标检测耗时的问题，提出并实现了一种基于模块级流水线加速的高斯-牛顿多回波在线测距方法（GNMO）。建立了GNMO测距方法吞吐率与流水线硬件资源消耗之间的关系，以提供加速设计基础。该高斯-牛顿迭代包含四个功能模块，通过内模块并行流水线和模块间流水线进行优化。在内模块并行流水线中，采用了无除法高斯-约旦消元求解器的矩阵稀疏化方法，显著降低了每个模块的输入到输出的延迟时间。实验结果表明，三子回波在线测距的最大测距率（MRR）可达637.8 kHz，约为此前报告的最先进性能的2.7倍。测距标准差（RStD）在49.6 dB时为3.1 mm，在18.4 dB时为23.9 mm；平均测距误差（MRE）的绝对值低于5.0 mm。成像结果表明，采用所提出的在线测距方法实现的FW-LiDAR能够提供更均匀分布的点和更密集的树木结构特征。

作者

Xiaolu Li 仪器与光电工程学院，北京航空航天大学，北京，中国；基于大数据精准医学先进创新中心，北京，中国 ORCID: 0000-0002-2550-5676

Wenbin Chen 仪器与光电工程学院，北京航空航天大学，北京，中国 ORCID: 0000-0002-7473-4475

Tengfei Bi 仪器与光电工程学院，北京航空航天大学，北京，中国 ORCID: 0000-0002-6466-1065

出版信息

期刊: IEEE Transactions on Instrumentation and Measurement 年份: 2024 卷: 73 页: 1-15 DOI: 10.1109/TIM.2024.3472835 文章编号: 10704719 ISSN: Print ISSN: 0018-9456, Electronic ISSN: 1557-9662

指标

论文引用: 2 总下载: 228

资助

关键词

IEEE 关键词: Distance measurement, Accuracy, Jacobian matrices, Fitting, Field programmable gate arrays, Pipeline processing, Vectors, Sparse matrices, Real-time systems, Imaging

索引词: Ranging, Pipelining, 3D Images, Late Time, Resource Consumption, Sparse Matrix, Throughput Rate, Computational Resources, Point Cloud, Coefficient Matrix, Lookup Table, Jacobian Matrix, Bit Error Rate, Peak Signal-to-noise Ratio, Digital Signal Processing, Levenberg-Marquardt Algorithm, Constant Vector, Distance In Meters, Laser Ranging, Target Interval, Gauss-Newton Method, Design Pipeline, Echo Pulse, Kth Iteration, Avalanche Photodiode Detector, Computational Consumption, Expectation Maximization, Repetition Frequency, Invertible

作者关键词: Field-programmable gate array (FPGA), full-waveform light detection and ranging (FW-LiDAR), Gauss-Newton, module-level pipeline, multiecho ranging

未定义

参考文献

额外参考文献

  1. W. Luo 和 W. Wei, “低成本高分辨率LiDAR系统的非重复扫描,” IEEE Trans. Instrum. Meas., vol. 71, 2022, Art. no. 7005510, doi: 10.1109/TIM.2022.3181272.

  2. C. Liu, X. Li, J. Huang, and L. Xu, “基于B样条的低信噪比LiDAR波形渐进分解,” IEEE Trans. Instrum. Meas., vol. 71, 2022, Art. no. 8501812, doi: 10.1109/TIM.2022.3157008.

  3. C. Liu, “一种针对森林中稠密点云生成的机载 LiDAR 波形的鲁棒去卷积方法，” IEEE Trans. Geosci. Remote Sens., vol. 60, 2022, Art. no. 5700314, doi: 10.1109/TGRS.2021.3061166.

  4. D. Wu, X. Zhong, X. Peng, H. Hu, and Q. Liu, “多模态信息融合用于多样场景下无人地面车辆高鲁棒性和低漂移状态估计，” IEEE Trans. Instrum. Meas., vol. 71, 2022, Art. no. 8505115, doi: 10.1109/TIM.2022.3205687.

  5. D. Bastos, P. P. Monteiro, A. S. R. Oliveira, and M. V. Drummond, “LiDAR 需求与技术综述：自动驾驶，” in Proc. Telecoms Conf. (ConfTELE), Leiria, Portugal, Feb. 2021, pp. 1–6, doi: 10.1109/ConfTELE50222.2021.9435580.

  6. J. Wu, J. A. N. van Aardt, and G. P. Asner, “基于小足迹 LiDAR 波形仿真的信号去卷积算法比较，” IEEE Trans. Geosci. Remote Sens., vol. 49, no. 6, pp. 2402–2414, Jun. 2011, doi: 10.1109/TGRS.2010.2103080.

  7. Y. Zhang, “为天问一探测器着陆 LiDAR 设计与制造的高度集成光学模块，” Opt. Lasers Eng., vol. 161, Feb. 2023, Art. no. 107346, doi: 10.1016/j.optlaseng.2022.107346.

  8. T. Zhou, S. C. Popescu, K. Krause, R. D. Sheridan, and E. Putman, “Gold—一种用于波形 LiDAR 处理的优化去卷积算法，” ISPRS J. Photogramm. Remote Sens., vol. 129, pp. 131–150, Jul. 2017, doi: 10.1016/j.isprsjprs.2017.04.021.

  9. X. Gao, “基于高速采样的改进脉冲激光测距算法，” 在国际光电技术应用研讨会上展示, Beijing, China, B. Lu and H. Wang, Eds., Beijing, China : SPIE, Oct. 2016, Art. no. 1015316, doi: 10.1117/12.2247342.

  10. D. Li, M. Liu, R. Ma, and Z. Zhu, “基于TDC的8通道LiDAR接收机，具有多间隔检测和实时现场校准，” IEEE Trans. Instrum. Meas., vol. 69, no. 7, pp. 5081–5090, Jul. 2020, doi: 10.1109/TIM.2019.2954173.

  11. J. Brown, C. Hughes, and L. DeBrunner, “实时硬件设计用于提升激光检测和测距精度，” in Proc. Conf. Rec. 46th Asilomar Conf. Signals, Syst. Comput. (ASILOMAR), Pacific Grove, CA, USA, Nov. 2012, pp. 1115–1119, doi: 10.1109/ACSSC.2012.6489193.

  12. X. Xie, L. Xu, Z. Wang, and X. Li, “基于在线回波波形拟合的实时现场激光测距,” IEEE Sensors J., vol. 19, no. 20, pp. 9255–9262, Oct. 2019, doi: 10.1109/JSEN.2019.2924706.

  13. X. Xie, L. Xu, X. Li, and D. Li, “基于在线高斯-牛顿的并行流水线实时现场激光测距方法,” IEEE Sensors J., vol. 20, no. 13, pp. 7087–7096, Jul. 2020, doi: 10.1109/JSEN.2020.2978120.

  14. X. Xie, X. Li, D. Li, and L. Xu, “通过FPGA上的反向传播神经网络实现的实时现场激光测距,” IEEE Sensors J., vol. 21, no. 4, pp. 4664–4673, Feb. 2021, doi: 10.1109/JSEN.2020.3030030.

  15. Y. Jiang, Q. Liu, H. Cao, and Y. Song, “基于FPGA、使用准牛顿优化方法的飞秒激光测距峰值检测,” IEEE Access, vol. 8, pp. 47776–47786, 2020, doi: 10.1109/ACCESS.2020.2979268.

  16. X. Xu, Y. Chen, K. Zhu, J. Yang, Z. Tan, and M. Luo, “基于深度学习的FPGA脉冲激光测距方法研究,” IEEE Trans. Instrum. Meas., vol. 70, 2021, Art. no. 2511611, doi: 10.1109/TIM.2021.3096281.

  17. T. Bi, X. Li, W. Chen, and T. Zhang, “基于激光雷达饱和波形补偿的高斯-牛顿在线测距方法,” IEEE Trans. Instrum. Meas., vol. 72, 2023, Art. no. 8503512, doi: 10.1109/TIM.2023.3277948.

  18. X. Li, T. Bi, and L. Xu, “基于模块级流水线的高频全波形激光雷达高斯-牛顿在线测距方法,” Measurement, vol. 229, Apr. 2024, Art. no. 114351, doi: 10.1016/j.measurement.2024.114351.

  19. G. Mountrakis and Y. Li, “用于波形激光雷达处理的线性近似迭代高斯分解方法,” ISPRS J. Photogramm. Remote Sens., vol. 129, pp. 200–211, Jul. 2017, doi: 10.1016/j.isprsjprs.2017.05.009.

  20. Y. Lu, H. Ma, E. Smart, and H. Yu, “面向实时性能的自动驾驶车辆定位技术综述,” IEEE Trans. Intell. Transp. Syst., vol. 23, no. 7, pp. 6082–6100, Jul. 2022, doi: 10.1109/TITS.2021.3077800.

  21. Z. Wang, L. Xu, D. Li, Z. Zhang, and X. Li, “基于 FPGA 的波形分解在线多目标激光测距,” IEEE Sensors J., vol. 21, no. 9, pp. 10879–10889, May 2021, doi: 10.1109/JSEN.2021.3060158.

  22. G. Zhou, X. Zhou, J. Chen, G. Jia, and Q. Zhu, “基于 FPGA 的 LiDAR 回波高斯分解算法,” Sensors, vol. 22, no. 12, p. 4628, Jun. 2022, doi: 10.3390/s22124628.

  23. X. Li, Z. Zhang, X. Xie, L. Xu, and D. Li, “基于矩阵稀疏化和无除法高斯–约旦求解器的多目标在线测距方法,” Meas. Sci. Technol., vol. 32, no. 9, Sep. 2021, Art. no. 095207.

  24. X. Li, T. Bi, Z. Wang, L. Xu, and Y. He, “基于阻尼高斯–牛顿的在线测距，用于低信噪比和高重叠波形的点提取,” Measurement, vol. 199, Aug. 2022, Art. no. 111479, doi: 10.1016/j.measurement.2022.111479.

  25. X. Xie, L. Xu, X. Li, and D. Li, “基于时分复用的在线高斯–牛顿多回波分解方法，用于实时现场激光测距,” IEEE Sensors J., vol. 22, no. 5, pp. 4152–4163, Mar. 2022, doi: 10.1109/JSEN.2022.3144522.

  26. U. Soederman, A. Persson, J. Toepel, and S. Ahlberg, “关于全波形航空激光扫描仪数据的分析与可视化,” Proc. SPIE, vol. 5791, p. 184, May 2005, doi: 10.1117/12.604655.

  27. X. Li and P. Luo, “利用多脉冲相干平均提升 LiDAR 的测距性能,” IEEE Sensors J., vol. 19, no. 15, pp. 6270–6278, Aug. 2019, doi: 10.1109/JSEN.2019.2910561.