Instructions DB3

Acquisition Protocol
The third Ninapro database is thoroughly described in the paper: "Manfredo Atzori, Arjan Gijsberts, Claudio Castellini, Barbara Caputo, Anne-Gabrielle Mittaz Hager, Simone Elsig, Giorgio Giatsidis, Franco Bassetto & Henning Müller. Electromyography data for non-invasive naturally-controlled robotic hand prostheses. Scientific Data, 2014" (http://www.nature.com/articles/sdata201453).
Please, cite this paper for any work related to the Ninapro database.

Please, consider that several clinical parameters related to the amputation can significantly influence the capability to classify hand movements using sEMG, as described in Atzori et al., Journal of Rehabilitation research and Development, 2016.
Please, cite this paper for any work related to Ninapro dataset 3.

The experiment is divided in three exercises:
1. Basic movements of the fingers and of the wrist
2. Grasping and functional movements
3. Force patterns
In the first two exercises, the subjects naturally try to repeat several movements represented by movies that are shown on the screen of a laptop.
In the last exercise the subjects have to think to press combinations of fingers with an iccreasing force on a custom made device called Finger Force Linear Sensor (Castellini 2012).
The muscular activity is gathered using 12 active double–differential wireless electrodes from a Delsys Trigno Wireless EMG system. The electrodes are positioned as shown in the figure: eight electrodes are equally spaced around the forearm in correspondence to the radio humeral joint; two electrodes are placed on the main activity spots of the flexor digitorum and of the extensor digitorum as described in; two electrodes are placed on the main activity spots of the biceps and of the triceps. The described locations have been chosen in order to combine a dense sampling approach with a precise anatomical positioning strategy.
The two electrodes placed on the main activity spots of the flexor digitorum and of the extensor digitorum were not used for subject 6 and 7 due to space reasons.

The electrodes were fixed on the forearm using their standard adhesive bands. Moreover, a hypoallergenic elastic latex–free band was placed around the electrodes to keep them fixed during the acquisition.
The sEMG signals are sampled at a rate of 2 kHz.
During the acquisition, the subjects were asked to repeat the movements with the right hand. Each movement repetition lasted 5 seconds and was followed by 3 seconds of rest. The protocol includes 6 repetitions of 49 different movements (plus rest) performed by 40 intact subjects. The movements were selected from the hand taxonomy as well as from hand robotics literature.
The number of movements is reduced in subjects 1, 3, and 10 due to experimental causes, e.g. stress.
The number of repetitions in subject 1 was originally 10 and has been reduced to 6 during post-processing in order to avoid classification biases.
This subject was the first subject being recorded with the acquisition protocol and determined the choice of reducing the number of repetitions to 6.

Data Sets

For each exercise, for each subject, the database contains one matlab file with synchronized variables.
The variables included in the matlab files are:
- subject: subject number
- exercise: exercise number
- acc (36 columns): three-axes accelerometers of the 12 electrodes
- emg (12 columns): sEMG signal of the 12 electrodes
- glove (22 columns): uncalibrated signal from the 22 sensors of the cyberglove
- inclin (2 columns): signal from the 2 axes inclinometer positioned on the wrist
- stimulus (1 column): the movement repeated by the subject.
- restimulus (1 column): again the movement repeated by the subject. In this case the duration of the movement label is refined a-posteriori in order to correspond to the real movement
- repetition (1 column): repetition of the stimulus
- rerepetition (1 column): repetition of restimulus
- force (6 columns): force recorded during the third exercise
- forcecal (2 x 6 values): the force sensors calibration values, corresponding to the minimal and the maximal force.

The cyberglove signal corresponds to raw data from the cyberglove sensors located as shown in the following pictures.
The raw data are declared to be proportional to the angles at the joints in the CyberGlove manual.