用深度学习实现七维轨迹重建，提升大体积靶实验精度

The VAMOS++ magnetic spectrometer is characterized by a large angular and momentum acceptance and highly non-linear ion optics properties requiring the use of software ion trajectory reconstruction methods to measure the ion magnetic rigidity and the trajectory length between the beam interaction point and the focal plane of the spectrometer. Standard measurements, involving the use of a thin target and a narrow beam spot, allow the assumption of a point-like beam interaction volume for ion trajectory reconstruction. However, this represents a limitation for the case of large beam spot size or extended gaseous target volume. To overcome this restriction, a seven-dimensional reconstruction method incorporating the reaction position coordinates was developed, making use of artificial deep neural networks. The neural networks were trained on a theoretical dataset generated by standard magnetic ray-tracing code. Future application to a voluminous gas target, necessitating the explicit inclusion of the three-dimensional position of the beam interaction point within the target in the trajectory reconstruction method, is discussed. The performances of the new method are presented along with a comparison of mass resolution obtained with previously reported model for the case of thin-target experimental data.