A Soft 3-dimensional Force Sensor For Measuring Head Impacts In Football Helmets

Abstract

Three-dimensional (3D) force data from head impacts in football is important for understanding traumatic brain injuries (TBIs). Football helmets have not yet been instrumented with true 3D force sensors compelling the development of soft 3D force sensors that could be seamlessly integrated into existing helmet padding. PURPOSE: To develop and determine the accuracy and reliability of an in-helmet soft 3D force sensor in comparison to force plate data. METHODS: A force sensor was fabricated by embedding magnetic powder within a cylindrical rubber cushion which was affixed to a magnetometer. Prior to integration in a helmet, known compressive forces were slowly applied to the sensor to assess performance in a controlled setting. The sensor was then inserted and fixed into a cavity created on the forehead pad of a football helmet. The sensor was controlled by a microcontroller, which communicated data wirelessly to a nearby computer at a sampling rate of 29 Hz. Thirteen trials were performed in which the helmet, while being worn, impacted a vertically mounted 3D force plate from different angles. The reliability of the helmet sensor in relation to the force plate was assessed by the interclass correlation coefficient (ICC) of the peak force measured in the primary axis of the impact (i.e. z-axis for a straight impact, x-axis for an angled impact) from the helmet sensor and the force plate. RESULTS: The magnetometer’s measurements of magnetic field strength exhibited a strong second order correlation to the force applied to the sensor (r2 = 0.995). For the helmet-force plate impacts, the force vectors were clearly observable for all trials by analyzing the force-time curves of each of the three axes of the helmet sensor. The helmet sensor displayed excellent agreement with the force plate for peak force (ICC = 0.981). In addition, peak forces for 8 of 13 impacts exhibited moderately low error (1-13%) when compared to the force plate, but the rest of the impacts had larger errors (25-51%). CONCLUSIONS: The sensor performed excellent in controlled studies and had high agreement with the force plate according to the ICC. Most impacts were accurately measured, but a higher sampling rate is needed for consistent and accurate quantification of impacts that contain higher frequency components.