Voice Activated Medicine Reminder Box with IoT Health Monitoring for Old People and Hospital

Our innovative IoT enabled Smart Medicine Box, designed to revolutionize medication management and health monitoring. With features including real time medication reminders, personalized dosage schedules, and mealtime alerts, our system ensures safety, efficacy, and security. This research comprises a combination of components such as the Arduino Nano, Wi Fi Module, and various sensors including Temperature, Pulse oximeter and Accelerometer. Our research offers seamless integration and accurate data collection. By wirelessly transmitting sensor data to the cloud, it enables continuous health monitoring and early detection of emergencies. It can show the time in real time on the LCD and also display patient’s health data. Whenever it’s time for medication, it plays a voice through the speaker to notify the doctor or nurses of the remaining medication. Additionally, the system indicates whether the medication should be taken before or after a meal. By using the HTTPS communication protocol, data is transmitted to the server from the Wi-Fi module ESP8266. The SIM800L ensures that data is available in the station section for power crises or any major issues. The system continuously monitors the patient’s body vibrations across three axes, generating real-time graphical data. When the health monitor detects measurements that exceed or fall below predefined thresholds, the cloud server automatically sends email and SMS alerts. Additionally, it promotes proactive healthcare by enabling physicians to securely access patient data remotely via a dedicated website and mobile applications. This smart system enhances medication adherence and safety, and improves medication and patient health monitoring.

Introduction

In Bangladesh, patient safety is a top priority, backed by constitutional provisions and the National Health Policy of 2011. With WHO’s support, initiatives like the Health Care Quality Strategy and the Patient Safety Action Plan are enhancing healthcare delivery across 510+ institutions. Real-time data dashboards drive improvements, ensuring universal access and better-quality care for all [1]. Islam et al. [2] demonstrate that a study conducted in Bangladesh identified 9% of households incurring catastrophic health expenditures, 7% resorting to distress financing, and 6% experiencing impoverishing health payments.

In the realm of health monitoring, the majority of research efforts have primarily concentrated on alert mechanisms utilizing buzzers, mobile applications, or web-based monitoring platforms. Several prior research studies have made significant contributions in this area. A smart medicine box was developed by Fauzan and Paramytha [3], utilizing an ESP32 microcontroller to integrate various components: ultrasonic sensors for automatic opening and closing, load cells for monitoring drug quantity, RTC sensors for scheduling, sound sensors for locating the box via clapping, and a buzzer for alerts. The ESP32 communicates drug consumption schedules to the user via Telegram. In [4], the authors designed an IoT-based Medicine Box system using NodeMCU ESP8266, an OLED display, a buzzer, and a pulse sensor to provide medication reminders and user-friendly support for elderly patients. Azlan et al. [5] proposed a smart medicine box equipped with visual and audible reminders to assist patients in taking their medication. The system is integrated with the Blynk app, which notifies care- givers when medication is taken. A weight sensor monitors the remaining medication, triggering alerts when it falls below 50% capacity, while Blynk provides weekly updates to caregivers. The medicine reminder system utilizes AT-mega328p and ESP32 for connectivity. A mobile app enables schedule input via WLAN, triggering alarms based on scheduled data, with GSM backup. Health monitoring features include LM35, MAX30100, DHT11, and MQ-2 sensors, displayed on an LCD screen by the ATmega328P microcontroller chip [6]. Chinchanikar et al. [7] designed a medicine box monitoring device featuring real-time tracking capabilities. The system incorporates an ESP module, an AC motor, a PIR sensor, and a buzzer, with IoT functionality enabled through the Blynk software, allowing for remote control and real-time sensor data visualization. This research presents a smart medicine box with IoT tools including Arduino, GSM module, and buzzer. Whenever the patient misses medication, the GSM module sends an SMS to the registered number after three buzzer alerts. Nurses can also monitor medication intake using this system [8]. Deepthi and Kumar [9] developed a medicine box system that incorporates a real-time clock (RTC) module, an Arduino, an LCD display, and a customized buzzer system. This system is designed to precisely remind patients to take their prescribed doses at scheduled times, ensuring timely medication adherence. Minaam and ELfattah [10] introduced a programmable medicine box with nine sub-boxes for managing different pills. It allows users to set daily dosage and timing, alerts with sound and light, and sends notifications via an Android app. Unlike traditional boxes, it automatically dispenses pills on schedule, preventing missed doses. Mukund and Srinath [11] proposed a smart medicine box with a microcontroller, keypad, LED display, motor controller, alarm system, and multiple pill containers. Users can schedule dispensing times, and the system provides visual and auditory alerts. A button dispenses pills and resets the alarm, while a secondary alarm signals low pill levels for refills. Philip et al. [12] designed a smart medicine box using an Arduino UNO, NodeMCU, RTC module, servo motor, and centrifugal pump. User inputs via an app are sent through NodeMCU and Firebase. The RTC triggers the servo for pills or the pump for liquids at scheduled times. An ultrasonic sensor detects if medication is taken and alerts the caretaker if missed. This IoT system combines an Arduino device with an Android app. Patients receive medication reminders through notifications, alleviating daily concerns. The app organizes medicine information and schedules. With an IR sensor, the Arduino device tracks medication intake, facilitating proper healthcare management [13]. Ayshwarya and Velmurugan [14] developed a medicine box with six sub-boxes and an Android app for reminders. It features a bio-sensor for temperature and heartbeat monitoring, along with authentication for safety. The system ensures precise timing, dosage, and stock indication, simplifying pill sorting during refills. Moise et al. [15] introduced a smart medicine box controlled via a phone app without subscription fees. Users can set pill distribution range and intervals through the app or on-board keys. Alerts notify if pills are left in the tray. A Portable Medicine Box was proposed by Chaudhari et al. [16], which enables real-time tracking using health sensors and Wi-Fi, linked to a Blynk app. It sends timely reminders to patients and provides real-time updates to caregivers, ensuring medications are taken on schedule through alerts and alarms. Chu [17] designed a Medicine Box, hospital platform, and smart bracelet. It reminds users to take medication, monitors their condition, and shares dosage and timing details with the platform for timely feedback and support. A Smart Medicine Dispenser system was introduced by Pathiraja and Wijesinghe [18], combining a physical device, mobile app, and cloud server. It dispenses medications on schedule, monitored via an HMI display, with the app enabling scheduling and the cloud storing and syncing data. Paul et al. [19] proposed a system that monitors heart rate and includes a fall detection sensor to prevent injuries and fatalities. It also uses a servo motor, controlled by a Real Time Clock (RTC), to open and close the tablet box at scheduled times, ensuring timely medication intake and reminders. Some of the earlier studies focused on medicine reminder boxes, while others detected patient’s body temperature, heart pulse rate, and blood oxygen levels separately. Previous research has not addressed issues such as onsite data availability during power crises, effective notification systems, and robust data security measures. Our developed system seamlessly integrates medication management and patient health monitoring. It wirelessly transmits sensor data to the cloud for continuous monitoring and early emergency detection. The LCD shows real time data and medication reminders can be scheduled for the morning, afternoon, and evening. and a speaker alerts caregivers about medication schedules with a prerecorded voice. Utilizing the HTTPS communication protocol, the ESP8266 Wi Fi module transmits data to the server, and the SIM800L module ensures data availability during power crises or major issues. The system tracks patient body vibrations along three axes, producing real-time graphical data. If the health monitor identifies measurements that go above or below the set thresholds, the cloud server instantly triggers email and SMS notifications. Moreover, it supports proactive healthcare by allowing physicians to securely access patient information remotely through a dedicated website and mobile apps.

Overview of the System

In Fig. 1, a system block diagram is represented. The system uses several sensors to perform functions including monitoring a patient’s temperature, heart rate, oxygen levels, and potential falls from the bed. The memory card will be inside the card reader, and when it is time for medicine, the voice will be amplified and played over the speaker. The speaker and amplifier will make the voice extremely audible. The LCD displays real time data, and medication reminders can be scheduled for the morning, afternoon, and evening. The Wi Fi module transmits data to the server using the HTTPS communication protocol, and the GSM module ensures on site data availability. The patient’s condition is shown on the server, allowing patient authorities to monitor remotely.

Fig. 1. Block diagram of the project.

Methodology

This project involves the development of a Voice Activated Medicine Reminder Box and an emergency notification system. In this section, the circuit diagram, patient accident detection, data transmission and Website Data Storage are explained in more details.

Circuit Diagram of the System

In Fig. 2, the Controlling Circuit of the Voice Message Section circuit diagram is demonstrated. The Arduino Nano is connected to the SD Card Reader via digital pins D10, D11, D12, and D13. The SD card reader stores voice commands to remind users of their medicine schedule. The signal pin of the amplifier is linked to digital pin D6 on the Arduino Nano. The amplifier output pins are connected to a speaker to amplify the voice commands. In Fig. 3, the Controlling Circuit of the Medicine Reminder Section is illustrated. The medication reminder part includes an Arduino Nano, push switch, LCD, buzzer, and real time clock module. The LCD’s digital pins D2, D3, D4, D5, D11, and D12 are connected to the Arduino Nano’s digital pins. The LCD displays data serially, and the RTC module’s analog pins A4 and A5 are connected to the Arduino Nano. Additionally, the push button’s digital pins D8, D9, and D10 are connected to the Arduino Nano. The buzzer is linked to digital pin D13 on the Arduino Nano.

Fig. 2. Voice message transmitter circuit (Slave Section).

Fig. 3. Medicine reminder transmitter circuit (Master Section).

In Fig. 4, the Controlling Circuit of the IoT Health Monitoring Section is represented. This section includes the Arduino Nano, MAX30100, MPU6050, LCD, and Wi- Fi module. The microcontroller connects to the MPU6050 accelerometer via analog pins A4 and A5. The LCD’s digital pins are connected to the Arduino Nano’s pins D2, D3, D4, D5, D11, and D12, respectively. The pulse oximeter MAX30100’s analog pins A4 and A5 are connected to the Arduino Nano’s analog pins. The temperature monitoring sensor’s analog pin A7 is also connected to the Arduino Nano. The Wi-Fi module ESP8266’s TX and RX pins are linked to the Arduino Nano. The receiver circuit diagram, illustrated in Fig. 5, includes an ESP32 microcontroller and a ZigBee module. The digital pins D2, D15, D18, D19, D22, and D23 of the ESP32 microcontroller are interfaced with the corresponding pins of the ZigBee module.

Fig. 4. IoT health monitoring transmitter circuit (Slave Section).

Fig. 5. System receiver circuit section.

Data Transmission and Website Data Storage

For secure data transfer, the HTTPS communication protocol is used to send sensor data from the monitor to the server. It’s a secure communication protocol that ensures better control and error free transmission. The reliability of the Voice Activated Medicine Reminder Box system is ensured through its use of Next.js and a Node.js server, which safely store data in a MySQL database. The Arduino Nano gathers sensor data via the ESP8266 Wi-Fi module from the transmitter and sends it to the Node.js server. A user friendly Next.js website interface makes the Medicine Reminder Box and patients’ health data accessible from this secure repository. Scalability is ensured by MySQL, which handles large volumes of data effectively and provides a solid foundation for real time medicine and patient health data management.

Patients Accident Detection

In Fig. 6, the patient is shown lying in a normal position on the bed. The MPU6050 3-axis sensor graph displays no significant variations, indicating stable conditions. In Fig. 7, the graph illustrates patient accident detection. The MPU6050 accelerometer monitors the patient’s body vibrations along three axes, producing real-time graphical data. If abnormal vibrations are detected, an alarm is triggered, alerting hospital staff through a notification beep.

Fig. 6. Patients normal stage.

Fig. 7. Patients accident detection.

Onsite Mobile Network Connectivity

Power crises and data issues threaten connectivity, but the GSM SIM800L module ensures continuous data avail- ability with minimal power, making it essential for keeping critical services online during disruptions.

Programming, Implementation and Results

The project flowchart, full implementation, outcomes, and system performance analysis will all be addressed in this section.

Flowchart

In Fig. 8, the system’s transmitter workflow is illustrated. The hospital authority pre-configures the medication schedule by setting the time and date using a real-time clock. The system validates these settings, triggers an alarm, and displays a medication reminder on the LCD when the scheduled time arrives. Patient health metrics, including pulse rate (BPM), oxygen saturation (SpO2), and body motion, are continuously monitored by sensors. The system processes all sensor data, including body vibrations along three axes, and transmits it to a receiver via the HTTPS communication protocol. This enables real-time alert notifications to be sent to the hospital authority for immediate action. In Fig. 9, the flowchart of the receiver section is depicted. The ESP microcontroller, integrated with a ZigBee module, is utilized to receive data. The received data is then transmitted to the server and stored in a MySQL database for further processing and analysis.

Fig. 8. Flowchart of the project (Transmitter).

Fig. 9. Flowchart of the project (Receiver).

In Fig. 10, the flowchart illustrates the system’s operation. Users can log in through the website or mobile app. The server stored data, checks the received data, and compares it with predefined threshold values. The results are displayed on the LCD, and the system continuously loops to monitor and verify data in real-time.

Fig. 10. System operation flowchart.

Result and Analysis

In Fig. 11, depicts the implementation of the system transmitter prototype, including its medication reminder and health monitoring components. The system features an Arduino Nano, Wi-fi module ESP8266, MPU6050, MAX30100, SIM800L GSM module, Temperature Sensor, SD Card Module, RTC Module, buzzer, battery, and a 12IC LCD module. It operates based on the patient’s balanced axis movement, medication box reminders, and health monitoring. The HTTPS communication protocol is utilized to transfer data to the receiver. The gyro sensor detects whenever the patient lies down on the bed, generating graphs to help doctors identify the patient’s condition. The SIM800L ensures that data is available in the station section for power crises or any major issues. On time medication is ensured from the medicine box, and a recorded voice alert is triggered by the authority simultaneously. Additionally, the system can measure patients’ health data on time and provide alerts via email notification.

Fig. 11. Project overview (Transmitter).

In Fig. 12, the medicine reminder box and all its sections are displayed. There are three different sections for morning, afternoon, and evening medicine. LCDs display information for on-site monitoring purposes. In Fig. 13, the Medicine Reminder Time Setting Section is shown. Using push switches, medication reminders can be scheduled for the morning, afternoon, and evening. The yellow switch activates the time-setting function, the red switch adds time, and the white switch advances or decreases the time and confirms the time setting.

Fig. 12. Medicine reminder box.

Fig. 13. Medicine reminder time setting section.

In Fig. 14, the Receiver sections of the medicine reminder box are illustrated. The system incorporates an Arduino Nano, an ESP32 Wi-Fi module, an antenna, and a ZigBee module. Data is transmitted to the cloud server using the HTTPS communication protocol. In Fig. 15, the website demonstrates real-time visualization of all monitoring parameters related to patient health. It can be remotely accessed by the patient’s guardians or hospital authorities for monitoring. This assists doctors or nurses in providing timely and proper treatment. Additionally, it facilitates the monitoring of a large number of patients’ data simultaneously, contributing to advanced medical facilities and ensuring proper treatment.

Fig. 14. Project overview (Receiver).

Fig. 15. Cloud monitor data.

In Fig. 16, the LCD output demonstrates the medicine reminder times for the morning, noon, and night. The LCD output helps doctors or nurses provide medicine on time. In Fig. 17, the LCD output demonstrates the patient’s health measurement data. The graphs illustrate that BPM, blood oxygen level, and body temperature vary with respect to time. The LCD output helps doctors or nurses monitor the patient’s health condition both onsite and online.

Fig. 16. Medicine reminder time display on LCD.

Fig. 17. Patient health data display on LCD.

In Fig. 18, the pulse oximeter MAX30100 can measure both the heart pulse rate and the oxygen level in the blood of patients in real time. The graph shows the variation of this data with respect to time. At the same time, all monitoring data is displayed on the LCD and stored simultaneously in the cloud.

Fig. 18. Patients BPM and oxygen level measurement.

During the measurement process, data is transmitted through the serial port, and the system generates trace waveforms from the collected pulse data. The dynamic pulse wave curve is depicted in Fig. 19. As shown in Figs. 20–22, the cloud server sends email and SMS notifications whenever the measured value exceeds or falls below the threshold. Upon receiving these notifications, the hospital authority can take immediate action to address the issue. The integrated alert systems ensure the smooth and efficient operation of the overall monitoring system.

Fig. 19. Recorded pulse data.

Fig. 20. Mobile visualization.

Fig. 21. SMS notification.

Fig. 22. Email notification.

In Fig. 23, the graph illustrates the system’s operation time versus elapsed time. It demonstrates that the actual operation time is slightly less than the required time. The cumulative delay time indicates that the system responds with minimal response time.

Fig. 23. System operation action timing.

In Table I, a patient’s medical records for a day are presented. The reports show that the patient’s heart rate sometimes exceeds the normal range and becomes abnormal. At the same time, it measures the patient’s blood oxygen saturation, which can vary with respect to time. The patient’s body temperature indicates either a normal condition or fever. By monitoring all data simultaneously, the doctor can take precautionary measures rapidly.

Table I. Patient Health Records Observation

Standard value	Measured data	Comments
60–100 (BPM)	49 BPM	Heart rate is normal
60–100 (BPM)	72 BPM	Heart rate is normal
60–100 (BPM)	104 BPM	Heart rate is abnormal
60–100 (BPM)	110 BPM	Heart rate is abnormal
98.6°F/37°C	98°F/36.67°C	normal
98.6°F/37°C	99°F/37.22°C	Fever
98.6°F/37°C	98°F/36.67°C	normal
98.6°F/37°C	100°F/37.78°C	Fever
95%–100%	94%	O2 Saturation is Below average
95%–100%	95%	O2 Saturation is Normal
95%–100%	93%	O2 Saturation is Below average
95%–100%	95%	O2 Saturation is Normal

In Fig. 24 displays body temperature measurements for participants “A” to “E” using a temperature sensor via the axillary method. Measurements were taken for 30–90 seconds, with proper sensor placement, arm positioning, and no clothing interference to ensure accuracy.

Fig. 24. Patient’s body temperature in degrees Celsius.

In Figs. 25 and 26, the MAX30100 sensor was employed to measure heart rate (HR) and blood oxygen saturation (SpO2) for participants labeled “A” to “E.” This sensor integrates high-intensity LEDs and a photodetector, enabling it to operate in dual modes to simultaneously capture HR and SpO2 data. The sampling rate of the sensor is adjustable, ranging from 50 to 100 samples per second, allowing for flexibility in data collection based on specific requirements. To ensure the accuracy and reliability of the measurements, stringent protocols were followed. These included thoroughly cleaning the participant’s finger to eliminate any contaminants, such as dirt or oils, that could interfere with the sensor’s readings. Additionally, careful attention was given to the precise placement of the finger on the sensor to ensure optimal contact and signal detection. These steps were critical to minimizing errors and maintaining the integrity of the collected data. The MAX30100’s ability to concurrently measure HR and SpO2, combined with the implemented protocols, ensured that the data obtained was both reliable and valid.

Fig. 25. Patient’s heart rate in bpm.

Fig. 26. Patient’s blood oxygen level shown as a line chart.

In Table II, the system’s operation is compared with that of previous research.

Table II. Comparison with Other Content

Paper ID	Fall Detect	Alert	Website	Mobile	SMS	Email	LCD
[3]	No	Buzzer	No	No	No	No	No
[4]	No	Buzzer	No	No	No	No	Yes
[5]	No	No	Yes	No	No	No	No
[6]	No	Buzzer	No	No	Yes	No	Yes
[7]	No	Buzzer	Yes	No	No	No	Yes
[8]	No	Buzzer	No	No	Yes	No	No
[9]	No	Buzzer	No	No	No	No	Yes
[10]	No	Buzzer	No	Yes	No	No	No
[11]	No	Buzzer	No	No	No	No	Yes
[12]	No	No	No	Yes	No	No	No
[13]	No	No	No	Yes	No	Yes	No
[14]	No	No	No	Yes	No	No	No
[15]	No	Buzzer	No	Yes	No	No	No
[16]	No	Buzzer	No	Yes	No	No	No
[17]	No	No	Yes	No	No	No	No
[18]	No	No	Yes	Yes	No	No	No
[19]	Yes	No	No	No	No	No	No
Propose	Yes	Voice	Yes	Yes	Yes	Yes	Yes

Examination of the System

The ZigBee network in this system operates on the 2.4 GHz band, with a coordinator node establishing the network and acquisition/routing nodes joining through it. Due to 2.4 GHz’s limited diffraction and through-wall capabilities, TI’s Z-Network software monitors and debugs the topology for stable transmission. AES-128 encryption ensures data integrity, enabling efficient and error-free communication [20]. During the system’s data transmission process, the packet loss rate calculated as the ratio of lost packets to the total packets sent was tested. The evaluation was conducted in an open grassy field using an acquisition node, a routing node, and a coordinator node. The acquisition node sent 1,000 packets to the coordinator node through the routing node. Testing was conducted at 20 m, 40 m, and 60 m, as a 20 m range is sufficient for patient activity in home or disease prevention scenarios. The communication range can be expanded by adding more routing nodes. In Table III, show a packet loss rate of about 5% at 60 m, with almost no loss at 20 m for 32-byte packets. This meets the system’s wireless data transmission requirements, making it practical for real-world use.

Table III. Test Results of ZigBee Network Communication

Distance	Packet	Packets sent	Packets Received	Loss Rate
20 m	32/byte	1000	1000	0%
20 m	64/byte	1000	996	0.4%
40 m	32/byte	1000	988	1.4%
40 m	64/byte	1000	984	1.6%
60 m	32/byte	1000	948	5.2%
60 m	64/byte	1000	944	5.6%