A Neck Mounted Interface for Sensing of Swallowing Activity based on the Swallowing Sound
Due to various reasons significant number of people suffer from dysphagia and due to the aging of the population, the number are continuously going up. Although some dysphagia testing methodologies such as videofluoroscopy and manometry exist, they are both expensive and inconvenient to use. As a result, we are developing a neck mounted interface which could be easily used by both physician, therapists and the patients.
Development of Robot Mask to support facial expressiveness
The Robot Mask is a wearable device developed to assist expressiveness of the face. By using mechanical means and electrical stimulations to manipulate the skin, it generates morphological changes on the face, which are similar to natural facial expressions. The Robot Mask has been designed based on anatomical and biomechanical characteristics of facial expressions and at present it assists the oral group of facial muscles. Furthermore, it enables seamless control of assisted expressions through the use of embedded biorobotic control (controlling based on EMG and other bioelectrical signals) based brain interactions. The device is portable, lightweight, standalone, can be worn unaided and can be used without any technical assistance. Primarily designed to assist facial paralysed patient, the device could also be used for health care management, expression modelling and expression sharing.
The research was taken up at a time where no robotic technology was available to assist facial paralysed patients and I succeeded in developing a device to support expressiveness of the face.
A Study on Facial Morphology
Analysing facial morphology is important to understand features of natural facial expressions of healthy people. Furthermore, it provides guidelines for creating artificial expressions. In this regard, a geometric morphometric technique can provide a structured approach to the analysis of landmarks for facial shape variations. Although, there are studies on facial expressions for graphical and emotional modelling, little research has been done on facial expressions from a robotics perspective.
The research was conducted by using marker based motion tracking and 3D volumetric analysis on the face. From this research I obtained the dynamic characteristics of facial expressions: motion profiles and direction and rate of motion.
Shape Memory Alloy (SMA) based Actuators
Developed as actuators for the Robot Mask, this work has produced linear actuators with position controllability that can operate without any mechanical noise. Actuators, which are used to assist facial expressions, need to operate silently. Furthermore, they need to be safe and fail-safe. Shape Memory Alloys (SMA) are silent during operation and provides a high power to weight ratio compared to traditional motors. However, due to metallurgical characteristics such as hysteresis demonstrated by SMA, controlling them is difficult. In this research, I developed SAM-based unidirectional and bidirectional linear actuators with feedback position control to safely manoeuvre the facial skin. Actuators are equipped with embedded microcontrollers to control them and synchronize them with other actuators when used in a multi-actuator system.
Video: Here you can see the 3 actuators in operation at their normal speed
Selective Electrical Nerve Stimulation on the Face
When used on the face for rehabilitation, to eliminate abnormal patterns of muscle innervations, localizing the targeted nerve is important.
Although the use of a single pair of electrodes has been the standard practice for stimulation, positioning them over the targeted nerve branch is difficult. However, array electrodes have many advantages in selective innervation of closely distributed bunch of nerves. In this study I designed a 28-field partially flexible Platinum electrode array, which is suitable for intra-cranial stimulations, to achieve a higher selectivity while being unobtrusive.
After placing the electrode array over the parotid gland for stimulation, I derived a mapping between the stimulation location and the resulted muscle contraction. I also found the stimulation parameters that could be used to control muscle contraction. Furthermore, I derived a methodology to estimate the effect of simultaneous stimulation of multiple electrodes, which is important in assisting expressiveness of the face. These finding are useful for both assisting facial expressiveness and targeted physiotherapy.