Skip to main content

Architecture

System overview​

The proposed system architecture implements a comprehensive biomechanical monitoring solution supporting six distinct user interfaces, one web application, and one Raspberry Pi server deployment. The design follows a microservices pattern to ensure scalability, maintainability, and clear separation of concerns across functional domains.

image

Frontend Architecture​

The frontend architecture provides differentiated user experiences based on role-based access control, implementing two primary application modes:

Administrator Portal​

  • System Monitoring Interface: Real-time dashboard displaying system health metrics – such as sensor live battery, service status, and operational analytics
  • Data Export Interface: Tools for generating and exporting reports in PDF, CSV, and JSON formats with comparative analysis capabilities
  • User Management Interface: User account administration including creation, modification, deactivation, and role assignment

User Dashboard​

  • Real-time Monitoring Interface: Live data visualization and feedback during suit operation with immediate posture and movement analysis
  • Historical Analytics Interface: Longitudinal data review with trend analysis, performance metrics, and progress tracking over time
  • Profile Management Interface: Personal account configuration, and preference management

Backend Microservices Architecture​

Data Ingestion Service​

Purpose: Handles data acquisition dumped to the main application’s backend when the user connects to the raspberry’s server

  • Data Validation & Sanitization: Ensures data integrity and compliance with expected formats
  • Data storing: When the user connects to the raspberry’s server, the data on the time-series DB on the raspberry is dumped to the application’s backend running on the user’s machine, which is then saved to the application’s main time-series DB to ensure no data is lost

Processing Pipeline​

Purpose: Transforms raw sensor data into actionable biomechanical insights

  • Sensor Fusion Service: Correlates and synthesizes data from multiple sensor inputs
  • Posture Classification Service: Identifies and categorizes user postures (further down the line ML could be implemented here)
  • Fatigue Analysis Service: Detects fatigue patterns
  • Movement Pattern Recognition: Analysis and recognition of repetitive user movements

Alert Engine​

Purpose: Monitors user data streams and triggers appropriate notifications

  • Rule Manager: Configures and maintains alert criteria and business rules
  • Threshold Calculator: Dynamically adjusts alert thresholds based on individual user baselines (stored on the document-oriented DB)
  • Notification Dispatcher: Coordinates delivery of alerts across supported platforms

Analytics Service​

Purpose: Provides business intelligence and reporting capabilities

  • Trend Analysis: Identifies patterns and correlations in user data over time
  • Statistical Aggregation: Computes summary statistics and performance metrics
  • Report Generator: Creates structured reports for both individual and comparative analysis

User Management Service​

Purpose: Manages user identities, authentication, and authorization

  • Authentication Service: Handles user login, session management, and security tokens
  • Profile Manager: Maintains user preferences, historical data, and personal configurations

Administration Service​

Purpose: Provides system administration and operational oversight

  • User management: Administrative operations for user accounts with bulk processing capabilities
  • System management: Aggregates operational metrics, service health status, and usage statistics
  • Role-Based Access Control: Enforces permission policies and administrative privilege management
  • Administrative operations: System configuration, maintenance controls, and global settings management

Audit Service​

Purpose: Maintains comprehensive activity logs for security and compliance Activity Logger: Captures user actions, system events, and data access patterns Log Query Engine: Provides filtered search, aggregation, and retrieval of audit records Compliance Exporter: Generates audit trails for regulatory and security reviews Retention Policy Manager: Automates log lifecycle management and archival processes

Export Service​

Purpose: Handles data export and comparative reporting functionality

  • Report Generator: Creates customized reports for users, periods, and system metrics
  • Data Formatter: Transforms data into required formats (PDF, CSV, JSON) with appropriate styling
  • Comparison Engine: Performs statistical comparisons between users and time periods
  • Batch Export Processor: Manages asynchronous export jobs with progress tracking

Raspberry Backend​

The Raspberry Pi serves as the computing hub that interfaces directly with wearable sensors, providing real-time data acquisition and preprocessing. This edge layer ensures low-latency processing and operational resilience even during network interruptions.

Sensor data aggregation​

Purpose: Unifies heterogeneous sensor data streams into standardized, time-synchronized packets for efficient storage.

  • Buffer Management: Implements circular buffers for temporary data storage during network latency periods
  • Time-Synchronization Engine: Aligns timestamps across all sensor sources using hardware clock synchronization

Sensor management​

Purpose: Provides comprehensive lifecycle management and monitoring of all connected sensor hardware.

  • Sensor Discovery & Registration: Interface to add and authenticate new sensors
  • Configuration Manager: Pushes calibration parameters and operational settings to individual sensors

Data collection/processing​

Purpose: Executes essential data preprocessing and quality assurance to reduce application’s processing load and ensure data integrity.

  • Raw Data Acquisition: Interfaces with sensor drivers to collect raw biomechanical measurements
  • Signal Preprocessing: Applies noise filtering, outlier detection, and signal normalization algorithms
  • Data Quality Validation: Implements sanity checks and validation rules to flag corrupted or implausible readings
  • Local Data Caching: Stores processed data in a time-series DB to ensure no data is lost. This data is then dumped to the user when they connect to the raspberry’s server

Data marshalling and dispatching​

Purpose: Provides real-time data streaming and protocol transformation for direct client connections, enabling immediate feedback to users during suit operation.

  • Multi-Format Marshalling: Converts internal data structures to the correct data format ready to be sent to the client
  • Payload Optimization: Minimizes data payload size while preserving critical biomechanical information
  • Type Validation: Ensures data integrity during format transformation processes
  • WebSocket Server: Maintains persistent, full-duplex connections with connected clients
  • Session Authentication: Validates user credentials and establishes secure communication channels
  • Heartbeat Monitoring: Implements keep-alive mechanisms to detect and clean up stale connections
  • Quality-of-Service Tiers: Provides differentiated data streaming rates based on network conditions
  • Backpressure Handling: Manages data flow control when clients experience processing bottlenecks