基于单光子线性APD的信息提取算法研究

Research on information extraction algorithm based on single photon linear APD

针对新型单光子线性APD器件获得信息维度多、灵敏度高的特点,提出了一种基于单光子线性APD的微弱回波信号距离信息提取算法。首先利用微弱信号的泊松分布特性对回波信号进行建模,随后对每纳秒内的回波光子数进行统计,利用与脉宽等宽的距离窗进行滑窗求和,将最大值位置作为信号脉冲接收时刻,进而反算距离信息。设计了跟踪前的目标搜索策略和稳定跟踪后的距离门匹配算法,有效降低了背景光干扰和计算资源消耗,在回波光子数分别为200、10、5个时分别获得了100、9993、9658的计算正确率。对单光子量级信号设计了脉冲积累算法,单个光子信号回波的计算正确率为995。同时对静止目标通过算法迭代实现了信噪比为0dB甚至负dB的信息提取,对003个光子信号回波探测的计算正确率为961。仿真结果表明该算法可以综合应用单光子线性APD的强度信息以及单光子灵敏度,实现了在极限灵敏域下的距离信息提取。

Aiming at the characteristics of multi dimension and high sensitivity obtained by new single photon linear APD devices,an algorithm based on single photon linear APD is proposed to extract the distance information of weak echo signal.Firstly,the echo signal is modeled using the Poisson distribution property of the weak signal,followed by counting the number of echo photons per nanosecond,performing a sliding window summation using a distance window of equal width to the pulse width.The position of the maximum value is used as the moment of reception of the signal pulse,which in turn back calculates the distance information.In addition,the target search strategy before tracking and the range gate matching algorithm after stable tracking are designed to effectively reduce the background light interference and computational resource consumption.And the calculation accuracy rates of 100%,99.93%,and 96.58%are obtained when the number of echo photons is 200,10 and 5,respectively.The pulse accumulation algorithm is designed for the single photon signal,and the computational correctness of a single photon signal echo is 99.5%.Besides,the information extraction with a signal to noise ratio of 0 dB or even negative dB is realized through algorithm iteration for stationary targets,and the calculation accuracy of 0.03 photon signal echo detection is 96.1%.The simulation results show that the proposed algorithm can comprehensively apply the intensity information of single photon linear APD and single photon sensitivity to realize the distance information extraction under the limit sensitivity domain.