**4. Discussion**

After analyzing the results obtained at the end of the experimental tests, the torquemonitoring system was validated. The system performed accordingly, and no events were reported during the experimental runs. The system was able to record the torques during the entire rehabilitation procedure and a special section was implemented in the user interface to show torque variation during the rehabilitation using a code of colors (green for normal run and yellow for when the torque is not normal, but it is in allowable limits, and red for when the torque in the system exceeds a certain imposed value). The allowable torque for each patient is established by the physiotherapist after the first general evaluation of the patient using a series of dynamometers to test the spasticity of the patient and the ability to perform a certain motion.

Using the Biometrics sensors, the rehabilitation system may be also used as an evaluation device. The robotic system is able to perform "blind" rehabilitation. Using this feature, after the general health assessment of the patient by the physiotherapist, ASPIRE may be used to measure the spasticity level of the patient. The goniometers are placed on the patient body in the previously identified areas (Section 2.4). After calibrating the sensors, the patient is placed in the rehabilitation system and prepared for the rehabilitation procedure. In the user interface maximum amplitudes for rehabilitation motion are introduced and the torque monitor system is started. The speed of the rehabilitation motion is set to lowest, and the number of repetitions is set to 1. After the robotic system starts to perform the rehabilitation motion, the sensor system records the real-time amplitude of the motion performed, and the torque is monitored within the allowable limits. If the system succeeds

in performing the entire motion without triggering the torque monitor, then the patient has no spasticity, but if the torque monitor is triggered then the spasticity degree of the patient is computed using the value given by the sensor measuring system. This value can be used as a starting point for the future rehabilitation sessions and after each session this value can be tested to see if there are some improvements.

Torque monitor can be a useful tool to test some of the effects of some neurological diseases. For example, if after rehabilitation, graphical results show regular variation that can be interpreted as noise, further analysis may reveal that the noise is caused by the patient's tremor.

Based on the success of the experimental tests, and on the feedback received during the experimental tests from both the subject and from the robotic system, the functionality of the improved control system was validated. The system is able to accurately monitor the torques from the system using only the encoder of the servomotors without using any additional sensors for torque monitoring. The obtained values were checked against the values that resulted from the virtual simulation of the system and the differences were within an allowable range, the differences caused by the structure of the material, density, and even the manufacturing of some components.

Future work targets the implementation of interactive user interface using human– robot interaction modalities in order to prepare the robotic system for another session of clinical trials for the fall of 2021.
