The algorithms used to track calories/energy in wearable devices can be broadly divided into a calorie consumption estimation algorithm and a calorie intake estimation algorithm. I’ll explain in detail how each algorithm works and its considerations.

1. Calorie burn estimation algorithm:

Calorie consumption is calculated as the sum of the basal and active metabolites. Wearable devices mainly focus on estimating activity metabolites and utilize the following factors.

1) Basic Metabolic Rate (BMR): The minimum amount of energy that is needed to sustain life. It uses formulas that make predictions based on age, gender, height, weight, and more. For example, the Harris-Benedict formula and the Mifflin-St Jeor formula are widely used. Wearable devices calculate BMR based on the personal information entered by the user.
2) Activity Energy Expenditure (AEE): The amount of energy consumed through physical activity. The wearable device estimates the AEE using the following sensor data.

• Accelerometer: It measures the intensity and frequency of a movement. It determines the number of steps, activity time, and activity type, and estimates energy consumption based on them. It increases accuracy by applying different calorie consumption coefficients depending on the type of activity. For example, walking and running have different coefficients because even activities during the same time period have different calorie consumption.
• Heart Rate Sensor: Changes in heart rate are closely related to the intensity of exercise. Using heart rate data, it is possible to estimate energy consumption during exercise more accurately. In particular, it calculates calorie consumption by taking advantage of the tendency that the heart rate increases as the intensity of exercise increases.
• Altimeter: You can take into account your energy consumption on an uphill or downhill road by measuring altitude changes. For example, since walking the same distance consumes more energy when walking uphill, it corrects your calorie consumption to reflect altitude changes.

3) The type of calorie consumption estimation algorithm:

• Machine learning-based algorithms: build models that predict calorie consumption by learning historical data and sensor data. For example, various machine learning algorithms such as regression analysis, support vector machines (SVMs), and random forests can be used. These algorithms have the advantage of being able to provide more accurate predictions by learning individual characteristics and activity patterns.
• Metabolic Equivalent (MET)-based algorithm: It expresses the intensity of an activity in units called MET, and calculates the calorie consumption according to the MET value. For example, walking is defined as 3 MET, running is defined as 8 MET, and the calorie consumption is calculated by applying the corresponding MET value according to the type of activity of the user.

2. Calorie intake estimation algorithm:

In the past, users had to enter their own food information, but recently, methods for estimating intake using the following technologies are being studied.

• Food Photo Analysis: Estimate the type and quantity of food by analyzing food photos taken with a smartphone camera. Using deep learning-based image recognition technology, it accurately identifies the types of food and calculates the intake by estimating the size and volume of food.
• Barcode Scan: Scan the bar code on the food packaging to get food information, and provide calorie information when you enter your intake.
• Link to food database: Link to online food database to search for food information, and provide calorie information when you enter your intake.

[Improved accuracy of calorie/energy tracking algorithms]:

The accuracy of a calorie-tracking algorithm may vary depending on an individual’s physical characteristics, activity patterns, sensor accuracy, and the like. The following factors should be considered to improve accuracy.

• Personalized settings: You need to accurately enter personal information such as age, gender, height, and weight to increase the accuracy of your basal metabolism estimation.
• Sensor Calibration: The sensor on the wearable device should be periodically calibrated to reduce measurement errors.
• Utilize a variety of sensors: We need to fuse information from a variety of sensors, such as accelerometers, heart rate sensors, altimeters, to make more accurate estimates.
• Utilizing AI Technology: Using AI technologies such as deep learning to learn individual activity patterns and perform more accurate estimation of calorie consumption and intake.

Through these efforts, the ability to track calories/energy of wearable devices will be able to provide more accurate and useful information.