Dr.Dry: Neural-Network-Based Clothes Drying Time Prediction System

Project Overview

Dr.Dry is an innovative intelligent system designed to predict clothes drying time using advanced neural network technology. By integrating environmental sensors and machine learning, the system provides accurate drying time estimates, helping users efficiently manage their laundry process.

Key Features

• Real-time Drying Time Prediction: Estimates clothes drying time within 10 minutes of hanging
• Weather Integration: Combines real-time and forecasted weather data
• Multi-platform Support: Available on Web and Mobile (Android/iOS)
• Modular Design: 3D-printed sliding rails and storage box for easy installation

System Architecture

Component	Technology	Function
Hardware	ESP32	Central control unit
Sensors	DHT22, HX711, Soil Moisture	Environmental and clothing data collection
Frontend	React, Flutter	Web and mobile interfaces
Backend	Next.js, Firebase, Nginx	Server-side processing and authentication
Database	PostgreSQL (Prisma ORM)	Data storage and management
ML Model	Keras, TensorFlow, scikit-learn	Neural network for prediction

Prediction Model Details

Model Specifications

• Architecture: 7-layer Neural Network
• Activation Functions:
  • Hidden Layers: ReLU
  • Output Layer: Linear
• Regularization: Dropout layers to prevent overfitting

Performance Metrics

• Maximum Time Error: 47 minutes
• Accuracy within 20 minutes: 94.07%

Training Data

• Dataset Size: Approximately 35,000 records
• Split: 90% training, 10% validation
• Test Dataset: 1,000 records

Prediction Input Features

• 10-minute continuous environmental data
• Humidity and temperature (environment and clothing)
• Clothing weight and initial weight
• Total 52-dimensional feature vector

Target Prediction

Remaining time to complete drying + 30-minute offset

Drying Completion Criteria

1. Weight curve slope stabilization
2. Soil moisture sensor conductivity approaching zero

Target Users

Individual Users

• Busy professionals
• Apartment and small home residents
• Families with sensitive skin or young children

Institutional Users

• Universities (dormitories)
• Hotels
• Fitness centers

Future Development Roadmap

Phase	Key Objectives
Short-term (1 year)	- Model optimization - UI/UX improvements - Weather API integration - Initial market promotion
Medium-term (1-3 years)	- Smart device integration - Expanded application scenarios - International market expansion
Long-term (3-5 years)	- Smart home ecosystem integration - Cloud-based personalized services - Sustainable product development

Environmental Adaptability

Suitable for various settings including:

• Dormitories
• Balconies
• Indoor and outdoor spaces
• Diverse climate conditions

Sustainability

• Reduces dependency on energy-intensive dryers
• Promotes natural drying
• Potential for ESG (Environmental, Social, Governance) alignment

Technologies Used

• Neural Networks
• IoT Sensors
• Cloud Computing
• Cross-platform Mobile Development