Accepted for/Published in: JMIR Formative Research
Date Submitted: Jan 16, 2025
Date Accepted: Dec 2, 2025
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Adaptive Feeding Robot for Patients with Motor Impairments: A Time Series Forecasting Approach
ABSTRACT
Background:
Eating is a fundamental daily activity essential for maintaining independence and quality of life. However, individuals with neuromuscular impairments often face significant challenges in achieving autonomous eating due to the limitations of existing assistive devices, which are largely passive and lack adaptive capabilities.
Objective:
This study introduces an advanced feeding robot designed to address these challenges by integrating time series decomposition, autoregressive integrated moving average (ARIMA), and feedforward neural networks (FFNN). The robot is aimed at enhancing feeding precision, efficiency, and personalization, ultimately promoting autonomy for users.
Methods:
The proposed feeding robot employs a combination of sensors and actuators to collect real-time data, including facial landmarks, mouth status (open/closed), fork-to-mouth and plate distances, and the force and angle required for food handling. ARIMA and FFNN algorithms analyze this data to adapt to users' unique eating patterns and preferences. Predictive modelling with ARIMA anticipates changes in user behavior, enabling dynamic adjustments to meal portions and feeding habits. A strain gauge sensor regulates pressure for stabbing or scooping, while ultrasonic sensors ensure precise positioning. Facial landmark detection and object recognition algorithms enhance safety by verifying open-mouth conditions and monitoring plate contents.
Results:
The integration of ARIMA and FFNN significantly improved feeding accuracy and personalization. The combined model achieved an〖 R〗^2 of 94%, outperforming standalone ARIMA (85%) and FFNN (88%). This fusion of algorithms enabled the robot to model temporal user interaction dynamics, leading to enhanced precision. The results demonstrated an adaptive response, with the robot achieving a high success rate and minimizing delays. The predictive system also optimized portion sizes and timing, improving user satisfaction. The response time decreased by 28% across 150 iterations, while the success rate increased from 75% to 90%, demonstrating the robot's learning and adaptability.
Conclusions:
By leveraging time series decomposition, ARIMA, and FFNN, the feeding robot achieved substantial improvements in accuracy, responsiveness, and personalization. The integration of predictive algorithms and adaptive learning mechanisms allowed the system to model both long-term and short-term user interaction dynamics effectively. These advancements highlight the robot's potential to deliver a highly precise, safe, and personalized feeding experience for individuals with motor impairments. Clinical Trial: N/A
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.