This repository contains a Python-based signal processing pipeline designed to estimate the footprint angle of an agricultural tire moving at 10 km/h using raw inertial sensor data (accelerometer and gyroscope).
- Data Ingestion & Cleaning: Handled raw data parsing issues, including correcting European CSV formatting (comma decimals) and fixing mis-indexed columns that distorted the data structure.
- Signal Filtering: Implemented a 4th-order Butterworth Low-Pass Filter (50Hz cutoff) via
scipy.signalto eliminate high-frequency road texture noise from the Radial Acceleration signal. - Event Detection: Built a dynamic thresholding algorithm (set to 50% of the baseline centrifugal force) to robustly detect the exact entry and exit points of the tire's contact patch.
- Kinematic Calculation: Calculated the final footprint angle by integrating the Angular Velocity over the isolated contact duration.
TireDataAnalysis.ipynb: The core Jupyter Notebook containing all data cleaning, filtering, and calculation logic.TireDataAnalysis.html: A static export of the notebook for quick viewing of the code and interactive matplotlib visualizations.PreprocessedTireDataN.csv: A 200-row sample of the preprocessed dataset used for the final calculations (truncated for confidentiality).
The algorithm successfully processed 397 individual tire rotations, filtering out micro-vibrations to calculate a physically accurate average footprint angle of 15.30°.