VDJ_NNdataset

Data collected at "Luigi Divieti" Posture and Movement Analysis Laboratory (Department of Electronics, Information and Bioengineering, Politecnico di Milano, Italy).
Data were simultaneously recorded from 3 commercial IMUs and an optoelectronic system during the execution of 112 Vertical Drop Jumps performed by 11 healthy participants. The data collected by both systems were time-synchronized. The 3-axes signals of right knee moment and right ground reaction forces were cut in order to isolate their values during the first landing phase of the VDJ using the event detection information obtained via Visual3D. IMU data were filtered to remove noise and to preserve the meaningful components of the signals by applying standard fourth order low-pass Butterworth filter with a 32 Hz cut frequency. After the application of the filter, the aligned signals of accelerometer and gyroscope (and the quaternions) of the three IMUs were cut in the same way as the signals recorded by the optoelectronic system. Finally, a table containing the matched signals of both systems in the first landing phase of each drop jump, as well as a reference column with the indexes of the frames of each jump, was obtained. Table is sorted according to Subjects' ID and the number of performed VDJs.

The data were used to develop a Neural Network able to predict the waverforms of knee moments and GRFs starting from the 3D accelerations and 3D angualr velocities recorded by the IMUs. Quaternions are inlcuded in the table but they were not used to develop the network.