Assessment of head dynamics using a flexible self-powered sensor and machine learning, capable of predicting probability of brain injury

• Machine learning is used with a flexible, self-powered sensor to predict velocity. • A method is presented for estimating the probability of injury from sensor's output. • Sensor detects impact sounds, influencing angular velocity predictions and insights. This work presents the application of a flexible, self-powered sensor designed to predict angular velocity and acceleration during head kinematics associated with concussions. This paper-thin, flexible device, which exhibits piezoelectric-like properties, is strategically placed on the back of a human head substitute to capture stress and strain in this region during whiplash events. The mechanical energy generated by varying magnitudes of whiplash is converted into electrical pulses, which are then integrated with multiple machine learning models. These models were tested and compared, demonstrating their ability to accurately predict angular velocity and acceleration of the head. This predictive capability can be utilized to assess the probability of brain injury. The findings demonstrate that this system not only enhances the understanding of head impact dynamics, but also opens avenues for developing more effective injury risk assessment tools. By combining innovative sensor technology with advanced machine learning techniques, this study contributes to improved safety monitoring in high-risk environments, such as high-contact and automotive sports.