暗物质粒子探测卫星粒子径迹重建算法研究

The Dark Matter Particle Explorer (DAMPE) was successfully launched on Dec. 17, 2015 to be the first Chinese space observatory dedicated to indirect study of the dark matter via searching the possible annihilation signal of dark matter particles. Being substantially a high-energy cosmic ray/Gamma-Ray detector, DAMPE is able to measure the charge, energy and direction of an incident particle by incorporating the detection of four sub-detectors: the Plastic Scintillator Detector (PSD), the Silicon Tungsten Tracker (STK), the BGO calorimeter and the Neutron Detector (NUD). On-obit tests showed that the satellite platform and all the payloads perform as or even beyond expectation. And naturally the follow-up tusk of the project will mainly focus on the data analysis and procession, among which track reconstruction plays a critical role. The direction measurement provided by track reconstruction is not only of significant scientific importance but also set the foundation of charge measurement and helps to improve the accuracy of energy measurement and particle distinguishing. The track reconstruction algorithm determines the accuracy, efficiency and speed of the reconstruction. Given our increasing understanding of the instruments and experiences accumulated in processing the data, and also motivated by the giant improvement embodied in the Pass 8 data release of Fermi, we realize the necessity and possibility in upgrading our current track reconstruction algorithm. In this project we aim to improve the reconstruction algorithm, mainly in the following aspects: To improve the BGO track reconstruction accuracy; To innovate the track pattern recognition algorithm based on the tree-based tracking method to improve the reconstruction efficiency; To apply the Non-Gaussian multi-scattering in the Kalman filter algorithm to improve the STK reconstruction accuracy; To parallelize the algorithm and introduce GPU to accelerate the reconstruction speed.

我国首颗天文卫星，暗物质粒子探测卫星已于2015年12月17日发射升空。在轨测试表明有效载荷各项指标都达到甚至超过预期。在此基础上，数据分析处理成为卫星在轨期间的任务重点。径迹重建做为数据处理的重要内容不仅直接给出入射粒子的方向，还是电荷测量的基础，并为能量测量和粒子鉴别提供辅助。径迹重建算法直接影响了我们最终获得这些物理量的精度、效率和速度。随着我们对探测器性能的熟悉，数据处理经验的积累以及对国际同类探测器的广泛调研，我们越来越感觉到当前采用的径迹重建算法存在着很大的改进的必要性和可能性。在此项目中，我们结合卫星在轨数据和模拟，尝试改进径迹重建算法。预期主要突破点包括：借鉴Fermi卫星经验，创新径迹识别算法，提高径迹重建，尤其是复杂重粒子径迹的识别和重建效率；将非高斯散射引入卡尔曼滤波算法中，提高径迹重建精度；通过算法并行化和引入GPU处理器提高径迹重建的速度。

暗物质粒子探测卫星（DAMPE）是我国首个空间宇宙线探测器，自2015年发射以来一直正常在轨运行至今。径迹重建是DAMPE利用其观测得到入射粒子方向的具体手段。除了获得粒子方向信息本身，径迹重建还是DAMPE电荷测量，以及在此基础上进行粒子鉴别的前题。DAMPE的方向测量依赖两个子探测器，硅微条探测器（STK）和BGO量能器（BGO）。其中STK是DAMPE方向探测的主要设备。STK径迹重建的优点是方向测量精度高，缺点是容易收到次级粒子的影响，如果判选错误，有可能带来较大误差。BGO径迹重建的特点正好相反，虽然测量精度低，但基于整个簇射包络的不易受干扰，在精度范围内可靠性高。本项目的工作可分为三个方面。首先，我们改进了BGO径迹重建。实现了BGO cluster径迹重建算法，在小幅度提高方向测量分辨率的同时，将径迹重建的效率提高了50%以上；其次，我们提出并实现新STK径迹重建算法。实现了两种基于线性拟合的STK径迹重建算法，分别为rebin算法和遍历算法。这两种方法的精度略逊于卡曼滤波算法，但重建效率有所提高，计算资源占用量大幅度降低；最后，我们结合BGO和STK的径迹重建，提出了适用于DAMPE的高精度、高效率、低计算消耗的整体径迹重建算法以测量粒子方向。