Demonstration of adaptive machine learning-based distribution tracking on a compact accelerator: Towards enabling model-based 6D non-invasive beam diagnostics

Abstract: Intense charged particle beam dynamics are dominated by complex nonlinear collective effects such as space charge forces and coherent synchrotron radiation in a 6 dimensional (6D) phase space (x,y,z,px,py,pz). Non-invasive 6D phase space diagnostics would be of great benefit to all particle accelerators. Physics models can potentially serve as non-invasive beam diagnostics, however the main challenges they face are uncertain parameters and beam distributions to be used for initial conditions which drift with time requiring lengthy measurements that interrupt accelerator operations. We present a fast adaptive machine learning method to automatically track time varying distributions and quantum efficiency maps and demonstrate the potential to predict 6D phase space.