The u
se of mu
sculo
skeletal modeling a
s a tool for analy
sing performance
sport i
s increa
sing. Thi
s typically involve
s simulating an athlete'
s motion a
s well a
s the external load
s they experience and a
sse
ssing mu
scle activitie
s with re
spect to the given kinematic
s. The finding
s of any analy
si
s are therefore dependent on the accuracy of the kinematic
s. Furthermore, for
sport
s where the environment directly influence
s the kinematic
s, for example
swimming, it i
s preferable to capture the athlete'
s motion in thi
s environment. For the example of
swimming, u
sing typical optical ba
sed
sy
stem
s i
s challenging due to marker occlu
sion
s and the reflection of the water, while the frequently u
sed manual digiti
sation of video data i
s laboriou
s and time inten
sive. Inertial mea
surement unit
s (IMU), however, have been
shown to be
suitable alternative for capturing gait kinematic
s.
spar0010">This paper compares results from a musculoskeletal model of human underwater flykick where the kinematics have been determined from two sources; manual digitisation and IMUs. The model simulates the anatomy of the trunk and lower limbs while motion is prescribed for; pelvic-pitch and pelvis-thorax, hip, and knee flexion-extension. It is found that the knee, hip and pelvic-pitch angles derived from the IMU exhibit close agreement to the manual digitisation process and captures the swimmer's motion well as compared to the respective video frames. The musculoskeletal model was executed for both input types and the observed maximum muscle activities were similar in both trend and mean. It is therefore suggested the multiple IMUs can be reliably employed in determining joint kinematics of underwater flykick.