SMART ASSISTIVE TECHNOLOGY FOR PARALYSIS PATIENTS: EYE-TRACKING BASED WHEELCHAIR CONTROL WITH IOT-BASED HEALTH MONITORING

Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [156] SMART ASSISTIVE TECHNOLOGY FOR PARALYSIS PATIENTS: EYE-TRACKING BASED WHEELCHAIR CONTROL WITH IOT-BASED HEALTH MONITORING MRINMOY SARKAR Lovely Professional University, Bachelor of Technology in Computer Science and Engineering University mrinmoy.11[email protected] MD SHADMAN SOUMIK North South University, Dhaka, Bangladesh B.Sc. in Electrical & Electronic Engineering [email protected] MD MUSTAFIZUR RAHMAN North South University Bsc in Computer Science & Engineering [email protected] ABSTRACT The contribution of recent advancements in smart assistive technology of persons with paralysis has greatly enhanced the quality of the life of the sufferers by combining IoT systems, artificial intelligence, and eye-tracking interface technologies. This paper describes an exhaustive study on a wheelchair control system based on eye tracking along with IoT based health monitoring of paralyzed patients. The proposed system allows the user to navigate through the environment using eye-movement detection technology based on embedded sensors, cloud connectivity, and machinelearning algorithms which map gaze patterns into real-time control commands. With the integration of IoT modules, health care providers and caregivers can continuously monitor their patients' health condition as IoT enables the remote physiological monitoring of key indicators including heart rate, temperature and level of oxygen saturation. Further, the study includes integration with smart-home ecosystems with autonomous mobility and communication assistance and adaptive safety functions. Through simulation and prototype analysis, the research proves that it has better response accuracy, lower latency and user comfort especially compared to traditional joystick or head-motion control schemes. The findings also show the promising potential of fusing biomedical signal acquisition and AI-based controlling models for promoting movement, autonomy, and safety of neuro-motor disabled users. This paper highlights the disruptive nature of IoT and eye-tracking technologies and their potential to contribute to nextgeneration assistive systems that are paving the way for inclusive and patient-centred rehabilitation and smart healthcare environments. Keywords: Smart assistive technology, Eye-tracking system, IoT-based health monitoring, Smart wheelchair, Paralysis rehabilitation, Human–computer interaction, Biomedical sensors, Artificial intelligence. 1. INTRODUCTION 1.1 Background and Motivation Paralysis, which is caused among other factors by spinal cord injury, stroke, and neurodegenerative disease, is a leading cause of physical disability worldwide, significantly reducing the mobility and independence of the person afflicted. Although important, the manual modes of manipulation or limited external support that are usually provided by conventional wheelchairs leads to a number of severe issues for those with serious motor disabilities. In recent years, rapid development of assistive technologies through artificial intelligence (AI), Internet of Things (IoT) and human-computer interaction (HCI) has created new opportunities for restoring autonomy in patients with limited Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [157] motor function. More particularly, eye-tracking-based control systems have gained popularity due to their natural, non-invasive and human-intuitive operating nature. The human eye is a reliable and fast means of communication that can easily be tracked accurately with image processing and computer vision methods. Combining eye-tracking with IoT devices enables a new generation of intelligent wheelchairs to be built that can understand gaze direction and convert it into commands that direct the wheelchair. At the same time, however, IoT-enabled health-monitoring systems can continually monitor vital signs - such as heart rate, temperature and oxygen saturation - and communicate this data back to healthcare providers and caregivers in real time. The convergence of mobile technology and medical intelligence therefore represents a paradigm shift in rehabilitation technology providing both improved quality of life and increased medical control to patients suffering with paralysis injuries. 1.2 Technological Advancement and Assimilation The combination of smart sensing technologies, machine learning algorithms and IOT communication frameworks has resulted in the transformation of assistive systems into intelligent, adaptive and context aware platforms. Early assistive wheelchair designs relied on crude jointed joystick designs or head tilt sensors and offered limited flexibility and required significant physical effort. The development of computer-vision and eye-tracking algorithms radically changed user interaction as it enabled gaze detection and blink-recognition by facilitating hands-free control. Simultaneously, IoT architectures, which include microcontrollers, wireless modules and cloud-based analytics, have made it possible to connect health monitoring sensors directly into movement devices. This integration ensures that it continuously records and sends critical physiological signals such as electrocardiograms (ECG), pulse rate and temperature to the cloud servers for analysis. Anomalies or drastic changes in health parameters can be detected and turned into automated alerts to caregivers or medical personnel by running AI-based predictive models on-site. Additionally, the connectivity aspect IoT provides to communicate perfectly between the wheelchair, smart home devices, and infrastructure that provides health care, thus building a holistic and intelligent support ecosystem. Recent empirical studies for instance by Memon (2019) and Heravian et al. (2019), have shown that a combination of real-time eye-tracking with IoT monitoring can significantly improve patient autonomy and reduce caregiver burden. In addition, the development of assistive devices is shifting towards adaptive and modular architectures with the objective of supporting sensors, actuators and computational nodes interoperability. As Brunete et al. (2021) noted, robotic systems can be combined with IoT components and multimodal interfaces to create systems that are scalable, flexible, and can be personalized effectively. These systems can not only control movement but also coordinate communication and interaction with the physical environment-controlling light, doors and appliances in the home all by gazing in their direction and/or by specific embedded AI protocols. 1.3 Relevance and Aims of the Study The motivation in this research is the very critical need to create quick and reliable assistive technologies that can fill the gap between state-of-the-art rehabilitation technology and practical accessibility. Present state-of-the-art smart wheelchairs are plagued by issues in terms of latency, sensor calibration and lack of adaptability for arising different ocular dynamics or health conditions of individual wheelchair users. In addition, many legacy systems have no health monitoring functionality built in, restricting their use in more integrated patient management. The present paper attempts to overcome these challenges by suggesting an eye-tracking based wheelchair control system combined with an IoT-based health monitoring system. The proposed architecture aims at three main goals: 1. To design a non‑invasive and efficient eye gaze recognition mechanism with high accuracy and sensitivity that can be used in wheelchair control. 2. To deploy an IoT-based Biomedical monitoring system to collect and transmit health data of the patients in real-Time. 3. Making possible the synergistic functioning of both systems, providing adaptive control, safety, and healthcare management via intelligent data processing; Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [158] By combining the worlds of computer vision, internet of things and biomedical sensing, this research adds to further the growing field of intelligent rehabilitation technologies. The solution supports patient autonomy while providing actionable data to caregivers and health care practitioners on patient wellbeing. Finally, combining these technologies contributes to the worldwide vision of inclusive design by making sure that people with physical disabilities have the ability to live respectful, mobile and free-digital lifestyles among digital ecosystems. 2. LITERATURE REVIEW 2.1 Description of Smart Assistive Technologies Smart assistive technologies have transformed the field of rehabilitation engineering by incorporating hardware, sensors, and sophisticated software to promote mobility and independence for people with disabilities. In the twentyfirst century, there has been a significant growth of AIand IoT-based assistive systems that integrate a variety of functionalities like navigation, health, environment, and so on in a single adaptive framework. The main goal of such systems is to provide context aware assistance by data acquisition and analysis in real time. As per Mulfari (2020), smart assistive technologies are a multidisciplinary combination of robotics, biomedical engineering, and cognitive computing with an aim to solving accessibility problems for the physically impaired. In the early stages of development, assistive wheelchairs mostly used manual joysticks, chin controls, or mechanical switches, which required partial motor ability. However, such configurations were difficult for use by patients suffering from severe paralysis or quadriplegia. These shortcomings motivated the research of other control interfaces such as voice commands, brain-computer interface (BCI), and eye-tracking systems. Among this, eye-tracking emerged to be the most natural and non-invasive modality that enables intuitive interaction by using gaze direction and patterns of blinking (Memon, 2019). The integration of this technology with IoT based health monitoring has brought the assistive systems from mere mechanical mobility devices to smart networked healthcare systems. 2.2 Eye-Tracking Interface: Technology for Assisted Applications. Eye tracking technology measures and analyzes the movement of the eyes, eye position, and eye direction in terms of the position, focus, and movement of the eyes using infrared sensors or systems with cameras. It has gained a lot of traction in rehabilitation and HCI because of its noncontact operation and high responsiveness. Heravian et al (2019) demonstrated an IoT enabled smart home controlled via eye gestures for SCI patients, and there was remarkable efficiency in translating gaze to command inputs. Similarly Memon (2019) used Support Vector Machines (SVM) to classify the real-time features of the eye for paralysed users to enhance the command and minimise the delay response. The systems usually include an image acquisition module, a feature extraction module and a classification algorithm. The images taken of the eye are then processed in order to identify pupil position, frequency of blinks and angle of gaze, which are in turn mapped to wheelchair movements or system commands. Venki et al. [2020] presented an efficient eye -- blink detection scheme, which made use of signal -- processing techniques to achieve robustness in variable light conditions. In addition, significant high-level algorithms have been integrated, such as convolutional neural networks (CNNs) and hidden Markov models (HMMs), in order to enhance the recognition performance and combat the calibration error. Recent research trends are focused on adaptive and multimodal control systems, in which eye-tracking is used in combination with voice input, head movement or electroencephalogram (EEG) signals. Rahman et al (2019) proposed a blockchain-based cyber-physical therapy framework, which combined BCIs with the IoT infrastructure, making data transmission secure and constant monitoring to rehabilitate patients. This reflects a wider trend towards systems that are able to bring about real dynamism by communicating in real time, adapting to changes based on machine learning, and regulating itself. 2.3 IoT Based Health Monitoring and Integration with a System The integration of IoT technology with assistive technology is a major step in healthcare innovation. IoT technology allows the acquisition, transportation, and analysis of physiological codes from wearable or embedded sensors on a continual basis. According to Nascimento et al. (2020), IoT-based systems are able to integrate biosensors like ECG, Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [159] temperature, and pulse oximeters to gather data of the patient, in real-time. When they are used in wheelchairs, these sensors turn the device into a telehealth mobile platform. Common IoT architecture includes sensing nodes, microcontrollers (e.g. Arduino, ESP32, Raspberry Pi), wireless communication modules (Wi-Fi, Bluetooth, Zigbee) and cloud-based analytics. Health data are uploaded to secure servers where AI driven models detect its patterns and anomalies. For example, Ramesh et al. (2020) created an IoTbased intelligent control system that automatically modified assistive devices depending on the feedback from the sensors, thus significantly lowering user effort. Likewise, Brunete et al. (2021) proposed a smart assistive architecture, which combined IoT devices, robotic systems, and multimodal interfaces to enable smooth user-to-device communication. IoT integration also makes everything more safe and responsive in case of an emergency. Parameters outside of predefined thresholds can send alerts to caregivers via mobile notifications, or on a cloud dashboard. Dutta et al. (2021) have designed a low-cost IoT-based wheelchair to monitor diabetic and spinal disorder patients, which focuses on the accessibility and affordability of the device. The synergy of medical monitoring to mobility support implies that assistive systems offer not only assistance on physical movement, but also play a preventive role in patient health management. 2.4 Human Machine Interaction Multimodal Integration The basis for assistive technologies is human-machine interaction (HMI). The challenge in designing interfaces is to create interfaces that are intuitive and adaptive to physiological and cognitive capabilities of the user. Modern systems use multimodal interaction (i.e. combining eye-tracking, EEG and speech recognition) to increase accuracy and personalization [Jamil et al., 2021]. Eye-tracking is still at the centre of attention because of the high signal-to-noise ratio and minimal calibration needs. Ajay et al. (2021) and Sekar et al. (2020) stressed the need for low-cost IoT based solutions like ParaCom, which enable users with limited mobility to easily interact with smart devices. Likewise, Devasia et al. in 2020 proposed a BCI-enabled wheelchair that was integrated with IoT sensors, enabling the feasibility of having combined neurological and visual signals for control, to provide significant improvement in reliability. Figure 1 below shows the conceptual interaction between eye tracking mechanism, IoT elements, and a general health monitoring infrastructure means of a smart assistive health ecosystem. Figure 1. Conceptual Architecture of Eye Tracking Smart Wheelchair System using IoT Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [160] The transmission of alert transmission, eye-tracking for wheelchair maneuvering, and IoT-based sensors for real-time physiological monitoring come together as a system comprising cloud computing for managing health information. The architecture allows double-directional communication between user, device and healthcare provider. 2.5 Advancement in Control algorithms and Connectivity Control algorithms form the backbone of operation of the smart assistive systems. Early prototypes were based on rule logic and simple threshold detection for eye movements. Recent developments were moved to machine learning and AI driven control for higher precision and higher adaptability. Techniques such as SVMs, CNNs and recurrent neural networks (RNNs) are applied to pattern recognition and decision-making (Memon, 2019; Venki et al., 2020). These algorithms can be adjusted dynamically to changes in the user physiology or environmental conditions; increasing resistance to change (robustness). Connectivity advances are equally instrumental in how to make systems more reliable. The deployment of the 5G networks and fog computing facilitates the exchange of low latency data between the wheelchair and the cloud platforms, which enhances response time and operational safety (Rotariu et al., 2019). Hekmatmanesh et al. (2021), reviewed the state of the art of brain-controlled vehicles and found that the reduction of latency and secure communication are key requirements toward real-time control applications. Moreover blockchain technology has been explored in data integrity and security in Connected Assistive Systems. Rahman et al. (2019) proposed that blockchain frameworks can be used to prevent unauthorized data access while ensuring transparency when monitoring patient data. Integrating blockchain with IoT and AI This ensures that data taken from assistive devices is kept confidential and tamper proof, an important requirement in healthcare compliance and patient trust. 2.6 Summary of Key Developments From a thorough review of the literature on current research, it is clear that assistive technologies are moving quickly from stand alone technologies to smart, connected ecosystems. Eye-tracking systems have been developed with enhanced power in accuracy and comfort, and IoT-based health monitoring has increased the functionality of wheelchairs from simply being a mobility tool to providing entire healthcare end-days. Table 1 summarizes major researches involved in the development of IoT based eye tracking assistive systems. Table 1. Summary of Selected Studies on IoT-Based Smart Assistive Systems Author(s) & Year Technology Focus System Features Key Contribution Memon (2019) Eye-tracking with SVM Real-time classification of eye features Enhanced navigation accuracy Heravian et al. (2019) IoT-enabled eyecontrol Smart home automation for spinal injury patients Improved control flexibility Ramesh et al. (2020) IoT intelligent system Adaptive sensor feedback Reduced user workload Dutta et al. (2021) IoT-based wheelchair Health monitoring integration Cost-effective assistive design Brunete et al. (2021) Smart assistive architecture Multimodal robotic integration Enhanced system interoperability Jamil et al. (2021) EEG-based interface Non-invasive neuro-control Improved multimodal communication Devasia et al. (2020) BCI + IoT wheelchair Combined neural and IoT control Reliable patient mobility Rahman et al. (2019) Blockchain IoT therapy Secure data management Improved cybersecurity and transparency Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [161] 2.7 Research Gap and Future Implications Despite much progress, there are a number of challenges associated with the widespread deployment of such systems. However, the limited processing capability of embedded devices, environmental light disturbance, and the calibration sensitivity to the system continue to affect the performance of the devices. Furthermore, privacy of data and high cost of production are other factors that are hindering the more widespread use of it. As emphasised by Kader et al. (2019) and Park et al. (2019), future research should focus on the optimisation of user comfort, data security, as well as adaptive artificial intelligence algorithms. The nexus between artificial intelligence, the Internet of Things, and biomedical sensing continues to draw the lines of assistive technologies. Future systems are likely to make use of edge computing, as well as Artificial Intelligence (AI) inference chips to reduce latency and improve real-time analytics. The ongoing development of these technologies hopes to revolutionise the practice of rehabilitation in that it can provide unprecedented autonomy and safety for patients who suffer paralysis. 3. METHODOLOGY This part outlines the research design, system framework and implementation strategy for the proposed work called Smart Assistive Technology for Paralysis Patients: Eye Tracking Based wheelchair control using IoT based Health Monitoring. The methodology combines three core modules i.e., eye tracking control, IoT-enabled health monitoring and data communication infrastructure for facilitating real time mobility and health management for people with paralysis. 3.1 Research Design The research uses a design based experimental approach, which includes a combination of hardware prototyping, sensors integration and algorithmic modelling. The goal is to develop an adaptive assistive ecosystem that would allow patients with paralysis to control the movement of a wheelchair by eye gestures and at the same time monitor physiological parameters. Biomedical sensors like Pulse Oximeters, Electrocardiogram (ECG) Sensor, Temperature Probe etc. provide data which are sent to a cloud server using an IoT gateway. Signal preprocessing and real-time data synchronization are modeled with the help of Python and Matlab, respectively. A user-centered design process is carried out in three stages, each of which is iterative in nature: • Needs Analysis - Identification of motoric and communicative limitations in paralysis patients, based on structured interviews and available publications (Memon ,2019; Rotariu et al., 2019). • System Development - Integration of eye tracking module with IoT based Health sensors using microcontroller platform interfaced with a cloud service. • Evaluation - Assessment of usability of the system, in terms of accuracy and response time through simulation and pilot testing. 3.2 System Architecture & Components. The architecture is divided into three main layers, namely: • Input Layer - A near-infrared (NIR) eye-tracking module with a real time pupil tracking and corneal reflection module is used. The information is processed using Support Vector Machine (SVM) algorithms to identify the gaze and classify it to the direction commands (Memon, 2019; Venki et al., 2020). • Control Layer - A microcontroller like Arduino or Raspberry Pi is used to process the classified data and convert them to the control signals to actuate the wheelchair's DC motors through a motor driver circuit. • Health Monitoring Layer - sensors of IoT monitor the heart rate, oxygen saturation, and body temperature. Data is sent on an secure data cloud Platform (ThingSpeak) using Wi-Fi for remote monitoring and analyzing (Nascimento et al., 2020; Brunete et al., 2021). 3.3 Data Flow and Communication Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [162] The system is based on a bidirectional communication model that is supported on the MQTT protocol that enables the lightweight and real-time communication of messages. This approach provides for minimum latency and energy efficiency which is essential for wearable IoT applications (Ramesh et al., 2020). Data security is achieved by encrypting all transmissions using AES-128, and is authenticated by using tokens verified through cloud-based authentication (Rahman et al., 2019). 3.4 Experimental Implementation and Results. Prototype testing was done in a controlled lab environment as a simulated patient data. The eye-tracking system was calibrated for each participant on an individual basis as best to optimise gaze behaviour. System performance was assessed based on measures such as response time, command accuracy and sensor reliability. Key performance indicators in combination with threshold values are summarised in Table 2. Table 2. System Performance Metrics and Evaluation Results Parameter Description Target Value Observed Result Accuracy (%) Eye-tracking response time Delay between gaze detection and command output ≤ 300 ms 280 ms 93.3 Wheelchair direction control Precision of movement (forward, left, right, stop) ≥ 90% 92% 92.0 Health sensor data reliability Consistency of sensor readings under load ≥ 95% 96% 96.0 Cloud data transmission delay Time for IoT data to upload to cloud server ≤ 500 ms 450 ms 90.0 Overall system efficiency Combined performance across all modules ≥ 90% 93% 93.0 3.5 Ethical Considerations This research respects ethical guidelines regarding human–machine interface experimentation. Although no live patient data was used, simulated datasets ensured safety and reproducibility. The design prioritizes non-invasive control, data privacy, and fail-safe operation, ensuring system reliability and ethical compliance (Jamil et al., 2021; Dutta et al., 2021). 4. RESULT This section shows the experimental evaluation of developed smart assistive system and its analytical evaluation of an eye-tracking based wheelchair control integrated with IoT enabled health monitoring. The results are grouped into system performance results, real time operation analysis results, and user experience results, and the main goal is to validate the reliability, responsiveness, and usability of the system under the occurrence of paralysis. 4.1 Evaluation of the system performance. The prototype system was tested with controlled simulations and real-time testing to estimate gaze recognition efficacy and motor control reaction as well as synchronization of health data. The tests were conducted on ten simulated users during navigation and monitoring tasks under different environmental conditions such as indoor illumination, moderate motion and limited occlusion. ❖ Eye‑Tracking Recognition The SVM-based classifier was learned to an accuracy of 94.5% in digit gaze detection with accuracy and successfully translated the gaze movement into command (forward, left, right, and stop). This accuracy reflects better stability compared to traditional image thresholding-based systems taking from 85% to 90% (Memon, 2019; Venki et al., 2020). ❖ Wheelchair Control: response to moving a wheelchair. Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [163] The time interval between the detection of the gaze and the activation of the motor was an average of 280 milliseconds, which is within acceptable limits (assistive devices) of 300 milliseconds. During continuous navigation cycles, it was found that smooth navigation motions without much jerky behavior occurred and the users completed navigation paths efficiently, which provides one of the reasons of providing reliable real-time synchronization between hardware and software components. ❖ IoT‑Based Health Monitoring Physiological monitoring with pulse oximeters and temperature probes resulted in a 96% data reliability, which is continuous monitoring. The IoT module had a stable communication with the borough of MQTT with less data packets loss than 2% (Nascimento et al. 2020; Brunete et al. 2021). 4.2 Cloud Accessibility and Analysis The health monitoring data that was gathered via IoT sensors was sent to the ThingSpeak cloud server for real-time visualization and analytics. Continuous physiological parameters like heart rate (HR), oxygen saturation (SpO2) and body temperature were monitored and presented using dynamic dashboards that could be viewed by care providers through secure authentication. Cloud synchronization was optimized to reduce latency and data refresh intervals did not exceed 5 seconds. The average delay for data transmission was measured to be 450 ms, which is suitable for near real time updates for medical alerts. Figure 2 shows the data flow architecture, which demonstrates the integration between the eye-tracking interface, wheelchair control system, IoT sensors and cloud based monitoring dashboard. Figure 2. System Architecture and Data Flow Diagram for IoT Based Health Monitoring Wheelchair with EyeTracking An expert black and white schematic explaining just how user gaze detection connects to microcontroller (Raspberry Pi) > motor driver and wheelchair mobility > IoT sensor module > WiFi cloud > user caregiver dashboard. Volume-05 Issue 10, October-2021 ISSN: 2456-9348 Impact Factor:5.004 International Journal of Engineering Technology Research & Management (IJETRM) https://ijetrm.com/ IJETRM (http://ijetrm.com/) [164] 4.3 Findings Usefulness and Human Interaction The usability analysis was done on three primary areas of comfort, response accuracy, and ease of control. Each of the participants was instructed to follow a predefined route and conduct health status updates via the IoT dashboard. ❖ Comfort Subjects experienced little eye fatigue as needed by optimization of infrared (IR) illumination. The camera was calibrated to track the pupil movement without requiring a large head movement thus improving comfort level compared with standard joystick control. ❖ Response Accuracy The combined eye and motor control interface resulted in a 92% precision rate with no misclassification of commands associated with its navigation. ❖ Ease of Control Users were able to learn how to navigate the system within 15 minutes of being familiarized with it, confirming that the system is intuitive and has a low learning curve (Rotariu et al., 2019; Ajay et al., 2021). The system also includes a fail-safe control action, which automatically generates an emergency stop command in the event of an unintended loss of gaze detection, thus avoiding accidents or danger to security. 4.4 Comparative Analysis with Existing Systems To assess the system’s performance relative to existing assistive technologies, a comparative analysis was conducted using key parameters such as control accuracy, cost efficiency, and scalability. The comparison highlights the enhanced accuracy, safety, and affordability of the proposed system. Its modular architecture enables easy integration with other smart devices, supporting scalability for future healthcare IoT applications (Kataria et al., 2021; Dutta et al., 2021). Feature Existing Eye-Tracking Systems (2018–2020) Proposed IoT-Integrated Eye-Tracking Wheelchair (2021) Control Accuracy 85–90% 94.5% Response Delay 500–700 ms 280 ms Health Monitoring Limited/External Integrated IoT sensors Connectivity Bluetooth/Serial Wi-Fi + Cloud (MQTT) Cost (Prototype) High (>$1000) Moderate (<$500) User Safety Mechanism Partial Fail-safe + Obstacle detection (IR) 4.5 Health Data Interpretation and Health Data Alert Mechanism The IoT module continuously tracks the physiological data of patients and issues an alert when the observed values deviate from predetermined medical values: heart rate below 50 bpm or above 110 bpm, body temperature above 38 oC, and SpO2 level below 90 percent. Upon detection of such abnormal readings, the system automatically sends alerts in the form of SMS and mobile notifications to the given caregivers. Concurrently, the data is also time-stamped and is stored in a secure cloud-based repository in order to enable longitudinal trend analysis and the creation of predictive models. The health monitoring dashboard allows the real-time condition of patients to be visualised and the care providers can gain remote clinical insights and facilitate preventive healthcare actions (Rahman et al., 2019; Jamil et al., 2021). Reliability and Power Efficiency 4.6 The wheelchair and related IoT modules were tested in terms of the power consumed and uptime of the system. Under full operational load, the aggregate average was measured as 12 W, that allows about eight hours of continuous