The recent growth in the use of inertial sensors and their applicability in areas such as health, sports, navigation systems and aircraft / aerospace engineering is a phenomenon that can have significant impacts on the use of this technology. In addition to the low cost, small size and low power consumption, the inertial sensors have the ability to measure the intensity of physical quantities of a moving object without the need for external references. In the health area, particularly in monitoring, this type of sensor, has been under investigation in situations of support for physical rehabilitation and monitoring of the mobility of Parkinson's and Alzheimer's patients. The effectiveness of these sensors depends on the ability of the algorithms of gathering information from each of them and estimating more accurate information by sensor fusion algorithms. The fusion of inertial sensors data with other types of sensors such as force censors, allows devices to determine patients movement intentions. Moreover, the physical rehabilitation of patients that suffered accidents or traumatic actions frequently requires the use of physical structures to guide and sustain the required rehabilitation exercises. In these cases accurate motion and force controls are mandatory to prevent further injuries and allow patient rehabilitation. The proposed system enables the control of limb movements based on the sensor data and the patient intentions. The type and intensity of the training may be programmed by the medical staff that receives feedback from the wireless sensors. The patients receive a number of visual and physical stimulus generated by a virtual reality installation.