The reconstruction of charged particle trajectories at the Large Hadron Collider and future colliders relies on energy depositions in sensors placed at distances ranging from a centimeter to a meter from the colliding beams. We propose a method of detecting charged particles that decay invisibly after traversing a short distance of about 25 cm inside the experimental apparatus. One of the decay products may constitute the dark matter known to be 84% of all matter at galactic and cosmological distance scales. Our method uses graph computing to cluster spacepoints recorded by two-dimensional silicon pixel sensors into mathematically-defined patterns. The algorithm may be implemented on silicon-based integrated circuits using field-programmable gate array technology to augment or replace traditional computing platforms.

大型强子对撞机和未来对撞机带电粒子轨迹的重建依赖于传感器中的能量沉积，这些传感器放置在距离碰撞光束一厘米到一米的范围内。我们提出了一种检测带电粒子的方法，这些带电粒子在穿过实验装置内约 25 厘米的短距离后不可见地衰减。其中一种衰变产物可能构成暗物质，在银河系和宇宙学距离尺度上，已知占所有物质的 84%。我们的方法使用图形计算将二维硅像素传感器记录的空间点聚类为数学定义的模式。该算法可以使用现场可编程门阵列技术在基于硅的集成电路上实现，以增强或取代传统的计算平台。