Micro:bit & Vital Signs from https://pulsesensor.com/

[JanTadeuszEkiel]

Micro:bit & Vital Signs

{:style="max-width: 50%;"}

Project abstract: Using the micro:bit and the pulse sensor ( https://pulsesensor.com/ ) we show how you can measure your pulse by analyzing the sensor's analog signal and how to use Console and USB connection for transfer of pulse data from micro:bit to PC.

Level of research: Masters, Phd, R&D, Medicine science

Status: Complete

Who is involved: Jan Ekiel, Liki Mobile Solutions, Lodz University of Technology

What is the area of research: Computer Science, Block coding, Analog signal and data analyzing, Vital signs and pulse,

Output: Where has the research surfaced:

1. https://medium.com/liki-blog/micro-bit-vital-signs-b76e495f6a59
2. https://github.com/JanTadeuszEkiel/homelab

Contact: e-mail: jan.ekiel@likims.com; LinkedIn: https://www.linkedin.com/in/jan-ekiel-a1998a4/; www.likims.com

Content of the article:

Micro:bit & vital signs

1. About the project: "Heart rate monitoring and pulse measurement"

This project shows in practice how to use micro:bit to implement a pulse measurement device. In order to build the pulse monitor and the measurement device, I used the micro:bit board and an optical heart rate detector. (pulsosensor.com)

For today's DIY we will need:

1. One micro:bit chip - www.microbit.org 
2. One pulse sensor (Pulsesensor) - www.pulsesensor.com
3. Three jumper wires for tiles with a female ending or a one-sided ending with a hook.
4. One power supply with a micro-B USB plug (i.e. a power adapter used to charge multiple mobile phones)
5. Optionally, one adapter for micro:bit which makes it easier to connect the cables.

The optical detector (pulsosensor.com) can be attached to your finger. The device scans the fingertip with an intensive light source and monitors the amount of light that returns to the detector. The blood flowing through the blood vessels absorbs the light. Within the function of time, we can observe that at one moment more light returns to the detector and next time - less. In this way we can observe the course of heart rate. The detector has got 3 wires. Two of them are used for power supply and one of them is an analog output that displays an electrical signal, proportionally to the level of light in the detector. The analog output can be connected directly to the analog controllers' interfaces. I connected them to the micro:bit board, as shown in diagram 1.

Diagram 1: The connection of micro:bit board with the pulse sensor.

Because micro:bit can share its power with peripheral devices, I connected it to the wires powering the optical detector. I used analogue-digital interface of micro:bit no. 2 to read the analog signal that represents the heart rate.

Note: the detector (pulsosensor.com) is very sensitive to motion (as it generates distortion). It is also important to choose the place of measuring the pulse and the pressure while attaching the detector.

At the website: https://pulsosensor.com you will find the simplest micro:bit algorithm for reading the heart rate signal from the sensor. In this project I have shown the extended potential of the micro:bit circuit. I used a USB connection between micro:bit and a PC to present the pulse pathway. The micro:bit programming environment enables receiving data from the USB cable and presenting them in a console on a browser or a dedicated software in the form of numbers or graphs. It is also possible to save read values to CSV files for further analysis. In this case, in the graph presented in the console, you can observe the course of the measured pulse. The value of the pulse is presented on a LED display.

Figure 2.1 presents a block program, which I divided into 5 main blocks. The first one is performed when the application starts. The four remaining ones are the loops, which are executed in a parallel manner without any “pauses”. Figure 2.2 presents a list of block variables used in the program. Although the variables do not require any special form of declaration, it is good to consider their numbers and names beforehand. Figure 3 shows the same program in JavaScript.

Useful remark:

Programmers can quickly switch between a block programming view and a JavaScript view. Both modes can be combined. If you’re missing a block, it would be easier to get this feature into JavaScript. After switching to block programming mode, the added function or a line of code in JavaScript will be presented on a separate block, which will be included in the whole program sequentially.

Figure 2.1. Block algorithm

https://makecode.microbit.org/_iid8zbRi7Uq2

Figure 2.2. Variables used in a program.

Figure 3. Source code in JavaScript.

let time1 = 0
let delta_t = 0
let time2 = 0
let PulseDet = 0
let pulse_out = 0
let counter = 0
basic.showString("PULS__")


basic.forever(function () {
    basic.showNumber(pulse_out)
    basic.showIcon(IconNames.Heart)
})

basic.forever(function () {
    PulseDet = pins.analogReadPin(AnalogPin.P2)
})

basic.forever(function () {
    serial.writeValue("Pulse diagram", PulseDet)
})

basic.forever(function () {
    if (PulseDet > 870 && counter == 0) {
        time2 = input.runningTime()
        delta_t = time2 - time1
        time1 = time2
        counter = 1
        pulse_out = (60000 - 60000 % delta_t) / delta_t
    } else if (PulseDet <= 430 && counter == 1) {
        counter = 0
    }
})

3. Program’s description

Figure 4 shows a diagram of the analogue signal that is received from the sensor. The graph represents the course of the measured heart rate. The observed waveform is not as precise as shown on ECG monitors but it gives the possibility to observe and measure the pulse.

Figure 4. Analogue signal transmitted from the Pulsosensor which represents the pulse rate.

The electrical signal on the analogue micro:bit output can reach a value between 0 and 3 V. The variable representing the read signal can achieve an overall value between 0 and 1023. The range of these values in our project can be read on the analog-digital interface no. P2. For the course of the pulse shown in Figure 4 there is a minimum and a maximum value of the recorded signal. This signal decreased to the minimum value of 165 and reached the maximum value of 949. The trial and error method (non-scientific) has defined for this program two decision-making levels of the signal received from the pulsometer (Figure 5):

Figure 5. Two decision-making levels defined for the pulse estimation algorithm.

1. When the signal level from the pulse meter rises above the value of 870, the program saves the current time of this event, assuming that an impulse was detected. Since this moment the program will not search for the next signal increase above the level of 870 unless…

2. …the signal from the pulse meter drops below the level of 430. Then, the program will be waiting again until the signal level rises above 870 in order to record the time of next impulse and to calculate the time (delta) between the first detected impulse and the next one. The calculated delta will allow you to measure the pulse in the next step.

Figure 6. Impulse searching and the recording of the time when the impulse appears

We can often observe some interference in the pulsometer’s signal. After the signal rises above the level of 870, it can go slightly below this level in short time and then rise above it again repeatedly. This type of signal disturbance would result in an error of the heart rate measurement with a value of even several hundred percents. Therefore, it is necessary to determine next decision threshold of 430. That’s the value below which the signal must fall before waiting for the next rate impulse begins. (Figure 6).

The algorithm of the precise heart rate calculation can be certainly more accurate. The way presented in the project was chosen as a compromise between the quality of the pulse measurement and the complexity of the algorithm, including the number of parameters and conditions in the program.

The above algorithm allows to detect impulses and calculate the time lapse between two consecutive impulses (Figure 7).

delta_t = time2 — time1

The time lapse between impulses (delta_t) was calculated by subtracting “time1” from “time2”. The time saved in “time2” variable is the moment of the most recent impulse detection, and the time saved in “time1” variable is the moment of the previous impulse detection.

Figure 7: Calculation of the time lapse between impulses.

After calculating the time lapse between impulses, it is possible to calculate the pulse, which is defined as the number of heartbeats per minute. The “Running time” function used in the block program records the system time to a variable. This time is recorded right at the moment when this function is being called. The time value is recorded in milliseconds. Therefore, the time lapse calculated in the program is also saved in thousandths of a second. For the correct calculation of the pulse rate, we divide 60 000 (one minute equals 60 seconds and 60 000 milliseconds) by the elapsed time or “delta_t”.

In Figure 3 that presents the block algorithm you can see how to calculate the pulse based on the passage of time. The variables in the block programming environment of micro:bit can only assume integer values. Therefore, the result of the calculation includes a correction for the remainder of the division.

Find out more about the program in the micro:bit block environment:

1. “On Start” block:

The algorithm in this block is executed only once. It appears as the first one right after the program starts on the micro:bit device. In this block there is a function that shows the welcome message “Pulse____” on the LED display.

The remaining four program blocks are the loops, which are performed parallelly.

2. Loop №1 — reading of the signal value from the pulsometer.

The “Forever” loop is performed, as the name suggests, continuously. Loop no. 1 reads the value of the signal from the pulsometer. As figure no. 1 shows, the pulsometer is connected to the analog-digital interface (pin) P2 of the micro:bit circuit. The “analogue read pin P2” function enables reading the value of the analogue signal. The read value is saved in the “PulseDet” variable. The process of reading the signal value and writing it to the “PulseDet” variable is repeated and executed in an endless loop.

3. Loop №2 — pulse calculation

This loop is performed in one of three defined scenarios. These scenarios are executed according to the conditional function “if” and “if else”. The first scenario is carried out along with two conditions fulfilled at the same time:

1. the value of the “counter” variable equals 0,
2. and the value of the pulsometer’s signal kept in the “PulseDet” variable is higher than 870.

If both conditions are met, it can be assumed that a rising side of the impulse has been detected. The time of this incident is recorded, and then we can use the recorded time to calculate the pulse. The time of the latest pulse is recorded in “time2” variable. In the next step “delta_t”, the elapsed time between impulses, is calculated. The time of the previous impulse stored in “time1” variable is also used here.

It is worth noting that after launching the program, the first calculated pulse will be faulty.
When the “time1” variable is called for the first time, it stores the “Zero” value.

After calculating a lapse of time between impulses, “time2” (the time of the latest impulse occurrence) becomes “time1” (the time of the previous impulse occurrence). Copying the variables of the “time2” value to the “time1” variable is executed for the purpose of calculations, which will be done during the re-run of loop 2. In the next step the “counter” variable gets the value “1”. Because of that the time lapse between impulses and the pulse rate will not be calculated until the value of the pulsometer’s signal drops below 430. When signal will achieve value below 430 (P2<430), the value of “counter” variable changes to “0”. Then, the pulse is calculated. To the “pulse_out” variable there is saved a value representing the pulse: the number of impulses per minute. It is calculated as the quotient of 60000 milliseconds and “delta_t” variable. We can calculate exactly, how many elapses between two impulses fit in one minute. Variables in a block programming environment can only take integer values. However, the division operation may not result in an integer. That’s why the program includes a patch, which enables avoiding an error of writing a floating-point number (the result of the division) into the “pulse_out” variable, which can store only integral numbers. Getting rid of the “fractions” is possible thanks to “remainder of” function. First, the remainder of the division (60000/delta_t) is subtracted from 60000 milliseconds and then the result of this operation is divided by “delta_t” variable. In this way the result stored in “pulse_out” variable will always have an integer value. After calculating the pulse, the program finishes performing the condition for the scenario in which the pulse signal increases above 870 and “counter” value is 0. Next, loop 2 re-starts the performance.

The second possible scenario for loop no. 2 specified in the condition “else if” will be executed along with two conditions fulfilled simultaneously:

1. The value of “PulseDet” variable representing the level of pulsometer’s signal is less than or equal 430.
2. The value of ‘counter’ variable is 1.

The ‘counter’ value is set as 1 when the pulsometer’s signal rises above the value of 870. Since the program cannot allow any interference, the time of the next impulse (when signal rises above 870) will not be recorded until the signal value drops to 430. When this condition is met, the program changes the value of “couter” variable to 0 and from that moment, Loop №2 will again record the occurrence of impulse which level gets above 870.

The third scenario of Loop 2, as well as “if” and “else if” conditions will occur when none of the conditions described above is met. This means that when e.g. the signal value rises above 870 and the value of “counter” variable equals 1. In this scenario there is only one operation performed: the re-start of Loop no. 2 execution.

4. Loop №3 — pulse value displayed on micro:bit circuit LED display

Here, in the loop, the value of the “pulse_out” variable is shown on the display screen. The pulse value will be presented on the screen in turns with the “heart” icon.

5. Loop №4 — transmitting the pulse signal value to a PC or other device connected to the USB cable of the micro:bit chip.

Within the loop, “serial write value” function will send a text to the USB interface. The text includes the words: “Pulse diagram” and a number that is the value currently stored in “PulseDet” variable. As a result, in the development environment console, you can observe a graph presenting the variation of the pulsometer’s signal level in the function of time. In the console it is also possible to read the received signal values and save them as a file for further analysis — e.g. to develop even better pulse calculating algorithm. In a block programming environment, we can turn on the console view by clicking “Show console device” button (Figure 8). This button appears only if the program uses “serial write value” function and when micro:bit is connected to the computer.

Figure 8: View of the block programming environment with a console to monitor data driven from the USB cable.

[microbit-mark]

Thanks for the contribution! I'll look to include it on https://tech.microbit.org/labs/

[JanTadeuszEkiel]

Great :)
That would be an honor for me :)
It it will helpful I'm sharing my GitHub repo: https://github.com/JanTadeuszEkiel/homelab with the same content.