Developing of a neural system, a Mamdani-type fuzzy inference system and a Sugeno-type fuzzy inference system for the identification of a volunteer position/activity
The project requires the analysis of a set of medical data. The application context is dermatology and, in particular, the compression therapy by means of bandages in the treatment of venous ulcers of the leg. In dermatology, venous ulcers are very frequent lesion of the skin of the legs, due to a compromised functioning of the blood circulation. To repair these ulcers, the blood circulation must be enhanced by exploiting the calf muscle pump which allows the blood to be moved back to the heart. The basic way to re-establish the blood circulation is the compression therapy. In the compression therapy, the doctor applies a compression bandage on the leg of the patient providing a given pressure of the ulcer area. The pressure applied by the bandage allows to repair the ulcer, typically, in a few months. The pressure applied by the bandage, and thus the efficiency of the therapy, depends on several factors:
- The type of bandage (depending, e.g., on the elasticity);
- The correct application of the bandage: an appropriate pressure should be applied in several parts of the leg according to the position of the ulcer, and the pressure should remain constant as much as possible between bandage applications;
- The activities performed by the patient during the therapy (e.g., walking, standing, etc.).
The data refer to 10 volunteers. A compression bandage was applied to their calf. Three sensors were applied in three different position to measure sub-bandage pressure. Each volunteer wears the bandage for 12 minutes. During this time, the volunteer perform different activities or maintains their positions. Each activity/position is performed/maintained for about 3 minutes. The sensors measure the pressure with the sampling time of about 82 ms. The positions/activities taken into account are the following:
- Supine position
- Dorsiflexion standing
- Walking
- Stair climbing
The data are organized as follows. One folder for each volunteer is provided. In each folder there are 4 files, one for each position/activity performed. Each file contains the measurements of the three sensors (in the first three columns), and the corresponding sampling time (in the last column). Please, note that the sensor measurements are electrical resistance and are measured in Ohms.
The objective of the project is twofold: on the one hand, we want to identify the position/activity of the volunteer by analyzing the pressure of the bandage, i.e., to distinguish among supine position, dorsiflexion standing, walking and stair climbing; on the other hand, we want to find out the least temporal interval that is necessary and sufficient to recognize the position/activity.