Interactive Bio-Feedback for Gait Training Robot

BSc. BFH in Mechanical Engineering, Matthias Schindelholz, Noser Engineering AG

"The Interactive Bio-feedback allows patients to control the speed of LYRA® by their own effort. They are actively involved and see instant results of their activity. This invention may lead to faster and more effective gait rehabilitation and reduced medical costs for neurologic impaired patients."

- BSc. BFH in Mechanical Engineering, Matthias Schindelholz, Noser Engineering AG

The Challenge:

Neurologically impaired patients need to recover the cardiovascular system and relearn lost skills such as floor walking. The LYRA® is a robotic device that assists the patient and therapist in this difficult tasks. Until now, the device speed was manually controlled by the therapist. An interactive speed control mode has to be developed in order to enable the patient to control the training speed by his own effort.

The Solution:

A new software module has been developed within the existing LabVIEW code. It monitors and analyses the motor torque which is influenced by the patients' behaviour on the device. It can distinguish between passive, neutral and active gait which leads to decrease, maintain or increase speed. An auto-calibration routine is developed to adapt the sensitivity levels of the module according to the patients' capabilities.


BSc. BFH in Mechanical Engineering, Matthias Schindelholz - Noser Engineering AG

NI Products used in this Project: LabVIEW, CompactRIO, Real-Time, FPGA, SoftMotion



Stroke, spinal cord injured and similar patients suffer under various neurological impairments after the first emergency phase. Two important disabilities are the loss of walking ability and the massive reduction of cardiovascular fitness after just a few days of inactivity. Neurological impairments in the brain or the spinal cord lead to muscular control reduction and therefore to a restricted ability to keep the cardiovascular system active. The key question is: Which type of training is most effective in gait rehabilitation if you don’t have enough control of your muscles? A proper solution is the Gait Trainer (GT) LYRA®. In modern rehabilitation canters, robotic assisting GTs can train patients while their feet are attached onto two end effector plates and their weight is compensated by a Body Weight Support (BWS) system. LYRA® GT provides a physiological gait trajectory to simulate floor walking making it possible to efficiently relearn walking and regaining cardiovascular fitness.


Initial position and goal

Until now, the GT has been controlled manually by the therapists. This means setting up the step length, body weight unloading, gait speed, training duration etc. In other words, the therapist had to decide the intensity and the training conditions for the patient. The basic idea of the project was to introduce interactivity between the GT and the patient. A new function has been developed to allow the patient to control the intensity of the training interactively by adjusting the training speed.


This might improve the gait relearning process and cardiovascular fitness level since self-paced activity of the patient is crucial in rehabilitation. The new function is called "Interactive Bio-Feedback" (IBF) and it allows the patient to control the gait speed based on his own effort while walking on the GT. One obvious effect, is that the patient would be highly motivated since he will be able to control his own walking speed by changing walking effort. If you think about the fact, that such patients are not able to walk independently anymore, this is a far-reaching and relevant experience for them. The intensity of the training will be well adapted to the patient automatically, influenced by its remaining abilities and increased motivation. However, the therapist can still control the device by enabling or disabling the feature in real time and also monitoring the patients' condition while giving some guidance and motivation as usual.


The Interactive Bio-Feedback feature has been invented by the innovation team of ABILITY Switzerland. With the collaboration of the National Instruments engineers from Noser Engineering AG, a prototype algorithm has been developed within the existing software of the GT in order to prove the concept of the invented feature. ABILITY Switzerland AG has now deployed the algorithm fulfilling all medical requirements that are specified within the IEC-62304 standard. The LabVIEW code is based on a Real-Time / FPGA Architecture and CompactRIO technology in conjunction with SoftMotion.


Methods and Realization

An important requirement of the IBF was to realize it without additional sensors in order to avoid the medical recertification of the device. And as we all know, certification of medical devices is not the fastest, cheapest and easiest process in the world.


The basic functionality proposed by the innovation team of ABILITY Switzerland AG is to monitor the patients' actual effort by detecting torque fluctuations of the left and right mechanisms of the GT. An algorithm has been developed by Noser Engineering AG to capture the torque signals, process them and control the training speed by updating the velocity commands to the motors. After some preliminary tests, it was apparent that only a specific part of the gait trajectory is sensitive enough to see a substantial difference in torque and thus allowing to distinguish between active, normal and passive gait within the LYRA GT®.


A special algorithm processes the torque readings in this specific section for both left and right side to decide if the gait speed has to be increased / decreased or stay constant. The speed is updated after 2 or more footsteps since the patient needs some reaction time to get adapted to it. The speed delta is always a constant value. It is important for the patient to recognize the different speed levels but it should be equally considerable that the speed should not change too high as that might cause insecurity.


When the patient walks without special effort (normal gait) the algorithm detects that the torque is between an upper and lower threshold. The speed is maintained. Passive behaviour leads to increased motor torque since the device has to do more work to maintain the gait trajectory. The torque overshoots the upper threshold, and as a result the speed is decreased. On the other hand, active gait causes the torque to undershoot the lower threshold and consequently the device to increase speed. The IBF concept is kind of self-regulating since the patient may get to his personal speed limit and cannot keep up the needed effort at a faster speed due to personal coordination limitations. For the patient, it is really motivating to get to a higher speed level, introduced by its own effort. For safety reasons, a maximum and minimum speed limit was also implemented to keep the training speed within the permissible speed range.

Since LYRA® is used by various affected patients with different fitness status, the torque thresholds are hard to define as fixed values. In order to adjust the sensitivity of the feature to suitable levels for each patient an auto-calibration routine is implemented. While the auto-calibration routine is activated, the therapist has to instruct the patient to walk on the GT while focusing on staying as neutral as possible. The IBF records the patient’s torque fluctuations and defines a patient specific torque threshold, which then is used when the auto-calibration routine is over and IBF is active. With this function, there is no need for the therapist to find these values manually and changes of the patients' daily conditions are considered. After the short auto-calibration routine, the IBF is ready to be activated.



The Interactive Bio-feedback allows patients to control the speed of LYRA® by their own effort. They are actively involved and see instant results of their activity. This invention may lead to faster and more effective gait rehabilitation and reduced medical costs for neurologic impaired patients.


Author Information:

BSc. BFH in Mechanical Engineering, Matthias Schindelholz
Noser Engineering AG
National Instruments, Platz 4
Root D4, 6039
Tel: +41 41 455 66 06