AI-BASED SMART NUTRITION AND HEALTH ASSISTANT

(Mobile Application for Healthy Nutrition (NutriGuide AI))



1. INTRODUCTION AND PROBLEM STATEMENT

In today's world, irregular diet, sedentary lifestyle, and the rise in

processed food consumption trigger chronic diseases such as obesity, type 2

diabetes, hypertension, and cardiovascular diseases. When individuals want

to eat according to their health status, they often encounter two major

obstacles:

# Not knowing how to translate their instant biometric data (blood sugar,

blood pressure, etc.) into their daily diet.

# Struggling to create healthy meal combinations that fit their immediate

health needs using only the available and limited ingredients at home.

Project Purpose

The purpose of this project is to develop an AI-powered mobile

application that analyzes instant health data obtained from wearable

technologies and user inputs to determine the user's immediate

macro/micro-nutrient requirements, and provides optimized, personalized

recipes using the ingredients already available at home.

2. SYSTEM ARCHITECTURE AND WORKING PRINCIPLE

The system consists of three main layers: data collection, dynamic needs

analysis, and recipe optimization.



+-------------------------------------------------+ | 1. DATA COLLECTION LAYER | | - Smartwatch (Blood Pressure, Heart Rate, Act.)| | - Manual Input / CGM (Blood Sugar, Triglyceride)| +------------------------+------------------------+ | v +-------------------------------------------------+ | 2. NEEDS ANALYSIS (ALGORITHM ENGINE) | | - Biometric Data Analysis | | - Macro/Micro Nutrient Requirement Mapping | +------------------------+------------------------+ | v +-------------------------------------------------+ | 3. RECOMMENDATION AND RECIPE LAYER | | - Selection of Raw Ingredients at Home | | - Database Matching and Recipe Generation | +-------------------------------------------------+


2.1. Data Collection Layer (Wearable Technology and User Input)

  • Wearable Device Integration: Blood pressure, heart rate, and daily activity

  • data are monitored instantly via the Spovan Electro 2X (or equivalent)

  • smartwatch integrated with the user.

  • Laboratory and Sensor Data: Sensitive data that devices cannot measure

  • directly, such as blood sugar (glucose) and triglyceride concentration, are

  • transferred to the system through Continuous Glucose Monitor (CGM)

  • integration or by the user manually entering their latest laboratory results

  • into the mobile interface.

2.2. Needs Analysis Engine

The application analyzes the incoming metabolic data and passes the food

items the user should consume or avoid through an algorithm filter:

  • Blood Sugar Analysis: Refers to the amount of glucose in a unit of blood.

  • In cases of high blood sugar, the system blocks simple carbohydrates with a

  • high glycemic index (floury foods, starchy foods, sugars) and increases the

  • limits for fiber and protein-rich foods.

  • Blood Pressure Analysis: This is the pressure exerted by the blood pumped

  • from the heart on the vessel walls. When high blood pressure (hypertension)

  • is detected, the system restricts daily sodium (salt) intake and highlights

  • foods rich in potassium/magnesium (such as green leafy vegetables).

  • Triglyceride Analysis: Shows the lipid (fat) level in the blood. If

  • triglyceride levels are high, the system completely eliminates trans fats and

  • saturated fats, while encouraging the use of omega-3 and unsaturated plant

  • oils (olive oil, avocado, etc.).

2.3. Smart Meal Recommendation and Recipe Optimization

  • Inventory Management: Using checkboxes on the user-friendly interface of the

  • application, the user selects the unprocessed raw foods currently available

  • at home (e.g., spinach, chicken breast, olive oil, bulgur, tomatoes).

  • Nutritional Value Database: A relational/non-relational database (SQL/NoSQL)

  • operating behind the application stores the exact ratios of carbohydrates,

  • fats, proteins, minerals, and vitamins per 100 grams of each unprocessed food.

  • Recipe Generation Algorithm: The system cross-references the ingredients at

  • home with the user's health needs. For example, if a user with high blood

  • sugar and triglycerides has "potatoes, chicken, and olive oil" at home, the

  • system filters out the potatoes or keeps them at a minimum. It then displays

  • a proportional recipe for "Baked Olive Oil Chicken" to the user, complete

  • with weight measurements and cooking instructions.

3. INNOVATIVE ASPECTS OF THE PROJECT

  • Dynamic Menu Management: Unlike standard diet applications, it does not

  • offer generic, fixed lists; it generates instant, dynamic recipes according

  • to the user's immediate physiological state and the ingredients available at

  • home.

  • Zero Waste / Sustainability: It encourages utilizing ingredients already

  • available in the kitchen, thereby preventing food waste.

  • Preventive Health Care: It helps individuals at risk of chronic diseases

  • maintain their parameters (blood pressure, sugar) in balance through daily

  • nutrition before illness occurs.


4. FUTURE DEVELOPMENTS (ROADMAP)

  1. Large Language Model (LLM) Integration: Powering the recipe generation

  2. engine with ChatGPT/Gemini APIs or other custom LLM APIs to offer users more

  3. creative and gourmet recipes.

  4. Advanced Sensor Integrations: Directly incorporating future smartwatch APIs

  5. capable of non-invasive (needle-free) blood sugar measurement as soon as

  6. they hit the market.

  7. Barcode Scanning System: Enabling users to scan packaged or whole raw foods

  8. via the smartphone camera to log them into the home inventory within seconds.

Author: Emre Pelit