How yudhik11/heart-rate-monitoring_ppg works
Core Components
• Data Input/Acquisition:
– Represent the raw input signals, which include PPG signals (from two pulse oximeters), ECG signals, and accelerometer data captured via the wristband.
– Mark that the data comes originally from .mat files (as seen under Problem_Statement/Data and Ouput directories).
• Preprocessing & Signal Conditioning:
– Include modules for filtering and conditioning the raw biosignals. Files like H_filter.m, BP.m (Bandpass Filter), and perhaps LMS_sig.m indicate filtering stages.
• Feature Extraction & Peak Detection:
– Show that after conditioning, the system detects key features. Scripts such as discoverPeak.m, findClosestPeak.m, and findPksInRange.m hint at algorithms for peak detection.
• Heart Rate Estimation Algorithms:
– Identify the core estimation methods (e.g., FOCUSS.m, MPinverse.m, Periodogram.m, SMART.m, SSR.m, SSR_JOSS.m, and SPT_JOSS.m). These modules likely implement different algorithmic approaches for estimating heart rate.
• Control/Integration Script:
– The script.m (and team_42.sh for batch execution) serve as the central controller that sequences the processing steps.
• Data Output & Reporting:
– The Ouput folder and report_team_42.pdf indicate reviewed and saved results. Show that results are written out (including BPM trace files) and possibly compared to ground truth.
Relationships and Data Flow
• From acquisition, route the signals to a “Preprocessing” block (filtering using BP.m, H_filter.m, LMS_sig.m, etc.) that cleans the signal.
• The filtered signal then flows into the “Feature Extraction” block, illustrated with peak detection routines (discoverPeak, findClosestPeak, findPksInRange).
• Results from these algorithms then aggregate—via the main control script—to produce heart rate estimations.
• Finally, direct the output to a “Results / Reporting” block, showing generation of BPM trace mat files and other documented outputs.
Architectural Patterns and Technologies
• That the design is modular: each MATLAB function (.m file) encapsulates a specific algorithmic step with a clear input/output pattern.
• That the architecture follows a pipeline (or layered) pattern—data moves from acquisition to preprocessing, then feature extraction and estimation, and finally to reporting.
• Denote that the implementation is in MATLAB with research-focused algorithms, often benchmarking multiple methods against each other.
• External documentation references (like the IEEE papers in the README and additional PDFs in resources) which provide context and are supportive of choosing different modules.
Implementation Details
• Then branch out to “Feature Extraction” (peak detectors) leading into multiple algorithmic blocks (FOCUSS, SMART, SSR, etc.).
• A central orchestration module (script.m and associated shell script) that ties these components together.
• Finally, the “Output” block should capture how the results (heart rate estimates; BPM trace files) are stored and reported.
• A section referencing “Documentation & Resources” (pointing to PDFs in Problem_Statement and resources) that validate the design choices.
Final Touches
• Every component has a clear label.
• Data handling (input/output operations),
• Configuration/Control scripts.