Machine learning can improve accurate monitoring of head impacts in american football

Head impact monitors can track athlete impacts and estimate their spatial and temporal aspects. And to monitor athletes in real-time, a monitoring system must report data accurately and precisely, and avoid reporting too many spurious, false events. For decades this has proven very labor intensive and difficult to accomplish. In this paper we present a means to greatly enhance impact monitoring in real-time using spatial and temporal aspects of impact kinematics applied to a machine learning model based on an impact monitoring mouthguard (IMM) system (Figure 1) deployed in American football.