LiftRight: Quantifying strength training performance using a wearable sensor

Abstract

Exercise is shown to affect the physical, psychological, and cognitive aspects of one's health. Despite the prevalence of fitness trackers for tracking heart-rate, steps, and other wellness metrics, very few tools exist to monitor strength training performance. This paper presents LiftRight which introduces a low-cost approach for quantifying an individual's performance during strength training sessions. Unique challenges include the lack of useful qualitative or quantitative metrics, and monitoring the complex motion of the exercises. We leverage an arm-mounted inertial sensor for tracking the user's arm and computing associated kinetics. We choose three upper-body strength training exercises, and segment the workout trace into constituent sets, repetitions, and phases. We also compute performance metrics such as velocity, range of motion, stability angle, and sticking points for eight candidates over a total period of 26 weeks. LiftRight achieves 96% accuracy in identifying sets and repetitions in a workout. In terms of lifting practices and its affects on performance, our quantitative and qualitative observations are found to be consistent with existing literature. We believe that the ability to accurately monitor weight training will lower the entrance barrier and help prevent injuries by helping users and trainers alike.