Kinematic algorithm to determine the energy cost of running with changes of direction

Changes of direction (CoDs) have a high metabolic and mechanical impact in field and court team sports, but the estimation of the associated workload is still inaccurate. This study aims at validating an algorithm based on kinematic data to estimate the energy cost of running with frequent 180°-CoDs. Twenty-six physically active male subjects (22.4 ± 3.2 years) participated in two sessions: (1) maximum oxygen uptake (V̇O_2,max) and maximal aerobic speed (MAS) test; (2) 5-m continuous shuttle run (two 5-min trials at 50% and 75% MAS, 6-min recovery). In (2), full-body 3D-kinematics and V̇O₂ were simultaneously recorded. Actual cost of shuttle running (C_meas) was obtained from the aerobic, anaerobic alactic and lactic components. The proposed algorithm detects "braking phases", periods of mostly negative (eccentric) work occurring at concurrent knee flexion and ground contact, and estimates energy cost (C_est) considering negative mechanical work in braking phases, and positive elsewhere. At the speed of, respectively, 1.54 ± 0.17 and 1.90 ± 0.15 m s^-1 (rate of perceived exertion: 9.1 ± 1.8 and 15.8 ± 1.9), C_meas was 8.06 ± 0.49 and 9.04 ± 0.73 J kg^-1 m^-1. C_est was more accurate than regression models found in literature (p < 0.01), and not significantly different from C_meas (p > 0.05; average error: 8.3%, root-mean-square error: 0.86 J kg^-1 m^-1). The proposed algorithm improved existing techniques based on CoM kinematics, integrating data of ground contacts and joint angles that allowed to separate propulsive from braking phases. This work constitutes the basis to extend the model from the laboratory to the field, providing a reliable measure of training and matches workload.