Control interfaces for active trunk support
Stergios Verros is a PhD student in the research group Biomechanical Engineering (BE). His supervisors are prof.dr.ir. H.F.J.M. Koopman and prof.dr.ir. G.J. Verkerke from the Faculty of Engineering Technology (ET).
People with Duchenne muscular dystrophy (DMD) lose the ability to move due to severe muscular weakness hindering their activities of daily living (ADL). As a consequence, they have difficulties with remaining independent and have to depend on caregivers. Medication cannot prevent or cure DMD but it can increase the life expectancy of patients. Notwithstanding the increase in life expectancy, people with DMD have a lower Health Related Quality of Life (HRQoL) compared to people without DMD. A possible improvement could be achieved with assistive devices to perform ADL and, as a result, to depend less on caregivers.
Symbionics (2.1) has been focusing on developing dynamic trunk and head supportive devices for people with neuromuscular disorders to assist them when performing daily activities. Three sub-projects were defined; they investigated user involvement, passive trunk support and active trunk support. User involvement entailed the interaction between the trunk and arm when accomplishing daily tasks. A passive trunk support was designed and tested in an experimental environment by people without and with an early stage of DMD. As the DMD progresses, more assistance is needed which could possibly be provided by an active trunk support. Thus, an active trunk support (which is the focus of this thesis) concentrates on the actuation and control of a passive trunk support.
Operating and controlling an active assistive device requires a control interface. The control interface is responsible for converting the intended movement of the user into a device movement. Several control interfaces have been proposed for the control of assistive devices, the most common ones being a joystick, force sensors and sEMG (surface electromyography). We evaluated their performance by building an experimental user-controllable trunk support.
The goal of this dissertation, therefore, is to evaluate control interfaces for active trunk support.
To this end, several research questions were formulated and investigated:
I. Is there an alternative to the intuitive trunk control interface to steer trunk muscles?
Current research on the control of orthotic devices that use sEMG signals as control inputs, focuses mainly on muscles that are directly linked to the movement being performed (intuitive control). However, in some cases, it is hard to detect a proper sEMG signal (e.g., when there is a significant amount of fat) or specifically for EMG from trunk muscles, respiratory muscles are located in the trunk as well and can easily disturb the control signal, which can result in poor control performance. A way to overcome this problem might be the introduction of other, non-intuitive forms of control. We performed an explorative, comparative study on the learning behaviour of two different control interfaces, one with trunk muscle sEMGs (intuitive) and one with leg muscle sEMGs (non-intuitive) that can be potentially used for an active trunk support. Six healthy subjects undertook a 2-D Fitts’ law style task. They were asked to steer a cursor towards targets that were radially distributed symmetrically in five directions.
II. Which control interface aids an active trunk support better?
A feasibility study evaluated control interface performance with a novel trunk support assistive device (Trunk Drive) for adult men with Duchenne Muscular Dystrophy (DMD) namely, joystick, force on sternum, force on feet and sEMG (electromyography). This was done as a discrete position tracking task. We built a one degree of freedom active trunk support device that was tested on 10 healthy men. An experiment, based on Fitts’ law, was conducted for the evaluation.
III. Which assistive admittance controller performs best in a 1-D Fitts’ law task?
This study was dedicated to the development and assessment of three different admittance control algorithms for a trunk supporting robot; a law with constant parameters, a law with added feedforward force, and a law with variable parameters. A Fitts’-like experiment with 12 healthy subjects was performed to compare the control laws.
IV. Do people with DMD generate satisfactory signals which can potentially drive an active trunk support?
In a previous study, we showed that healthy people were able to control an active trunk support using four different control interfaces (based on joystick, force on feet, force on sternum and sEMG). All four control interfaces had different advantages and disadvantages. The aim of this study was to explore which of the four inputs could be detectably used by people with DMD to control an active trunk support. Three subjects with DMD participated in two experiments: an active experiment with an active trunk support assistive device and a static experiment without the active trunk support. The challenge in both experiments was to steer the cursor into a target of a graphical user interface using the signals from the different control interfaces.
We concluded that, although the non-intuitive force on feet control is one of the best interfaces for people with DMD to control an active trunk support some DMD patients find it easier to use the EMG from their leg muscles. The joystick is the only usable intuitive control interface but, the function of one hand has to be sacrificed. The decision, as to which control interface works best, must be made per individual.