Page 145 - Kaleidoscope Academic Conference Proceedings 2024

ARTIFICIAL INTELLIGENCE DRIVEN TILT SENSOR BASED SMART DRINKING

DEVICE FOR STROKE SURVIVORS

Preeta Sharan , Anup M Upadhyaya , R Vasanthan
1
2
1
1
The Oxford College of Engineering, Bengaluru, Karnataka, India
2 The Oxford College of Physiotherapy, Bengaluru, Karnataka, India

ABSTRACT

The proposed work introduces a tilt sensor device designed underscores the need for effective rehabilitation strategies
to monitor glass orientation during drinking activities in and assistive devices to improve independence and quality
stroke survivors. Phase 1 of the study assessed the device's of life post-stroke [2]. Research indicates that stroke patients
reliability in 96 normal individuals, achieving a correlation often encounter difficulties using everyday objects like cups,
coefficient (r) of 0.99. In Phase 2, 96 stroke survivors were glasses, forks, and pens. Specifically, they struggle with
divided into six subgroups based on specific tilt orientations maintaining the vertical orientation of handheld objects,
of the glass during activity. The device's concurrent validity, particularly during transferring them to or from a table.
measured by Pearson's correlation, was 0.78 compared to These challenges stem from factors such as muscle weakness,
motion analysis data from KINOVEA. Intraclass correlation coordination issues, and sensory deficits resulting from the
(ICC) analysis demonstrated high agreement of 0.99 stroke [3].
between the actual angle readings and the measurement
angle from each trial. Results indicated that the device Presently, there are no standardized scales available for
significantly reduced orientation range from 2.31 degrees evaluating drinking task performance. A review of existing
without feedback to 0.85 degrees with feedback, highlighting tools and smart objects that track the positioning of the glass
its effectiveness in providing real-time feedback during during drinking activity in stroke survivors has been
drinking tasks. Additionally, the test-retest reliability conducted as part of the literature review. However, no tool
(interclass correlation) was 0.99, supporting the device's or device offers specific feedback on the orientation of the
consistency over time. Further work will involve the path for glass during drinking. Previous efforts have been made to
development of an AI-driven app using SQL files from data develop a smart cup capable of monitoring the arm and hand
collected from stroke survivors, aiming to provide activity of stroke patients while they drink. This cup would
personalized rehabilitation strategies. The developed tilt detect the amount of liquid in the cup, the force used to grip
sensor device shows promise as a reliable tool for it, its orientation, and any tremors [4]. In other context Roby-
monitoring glass orientation during drinking activities in Brami et al., conducted a study to quantify hand orientation
stroke survivors, with potential implications for enhancing and arm joint rotations during unconstrained reaching
rehabilitation outcomes in this population. movements in healthy and hemiparetic individuals. Four
electromagnetic sensors were used to measure hand
Keywords – Tilt sensor, Smart Glass, Stroke, Kinovea, IoT, orientation and compute wrist, elbow and shoulder joint
Artificial Intelligence angles. Hand azimuth is correlated to arm movement
direction in both healthy and hemiparetic individuals [5].
1. INTRODUCTION Timmermans A A et al. conducted a study to find the skill
preference of person with stroke regarding arm-hand training
Rehabilitation after a stroke is crucial for restoring motor and and examines the relationship between the use of patients
cognitive functions. Post stroke, various muscle deficit affected arm and their training preferences. A list of 10 most
functions and muscle weakness, spasticity and different preferred skills were found and analyzed between a group of
problems are observed. However, stroke monitoring and sub-acute and chronic stroke population. They concluded
rehabilitation are expensive, requiring costly infrastructures that patient preferences related mostly to task involving
and extensive medical staff involvement over prolonged ‘manipulation in combination with positioning’ and
periods [1]. Upper extremity dysfunction is a common ‘manipulation’ which is more similar to the drinking task that
challenge for stroke survivors, affecting their ability to involves grasping and manipulating the cup to the mouth
perform daily tasks. Acute paresis of the hand or upper limb with maintaining the stability of the cup position to avoid
affects up to 87% of stroke patients, with 40-80% water spillage [6]. Nayeem et al., presented the “MAGIC
experiencing insufficient restoration of function. This TABLE,” an uniquely designed tool that records the

978-92-61-39091-4/CFP2268P @ITU 2024 – 101 – Kaleidoscope