Controlling physiology during robot automated treadmill training

Abstract

The Lokomat gait orthosis automates treadmill training after SCI or stroke. The patient’s engagement was shown to have major influence onto the outcome of the rehabilitation. In order to influence the engagement of a patient during training, we try to control the patient’s psycho-physiological state via measurable quantities. As an example, we control the heart rate by adapting the Lokomat walking speed. On the one side, this enables us to prevent patient overstress during training. On the other side, this is the first step towards influencing the patient’s psycho-physiological state, which is reflected by physiological quantities such as the heart rate. We implemented a real time processing algorithm that uses a combination of steep-slope-threshold and beat expectation for heart beat detection. A PI fuzzy logic controller was implemented, which controls the Lokomat walking speed in order to stabilize the heart rate at a desired value. Single case results show that the controller was able to stabilize the heart rate of subjects around a desired value. Future developments of such controllers will have take into account the level of active participation shown by the subject during walking.