Real-Time Signal Prediction System

Requirements

• Python 3.9
• Install dependencies: pip install -r requirements.txt

Predict.py (TCN Model)

Model: TCN pt10.04

For 500Hz sampling frequency, recommended parameters: delay=20000, refresh=0.2

Usage:

python Predict.py <filename> <delay> <refresh> <offset>

Arguments:

Argument	Description
filename	Input CSV file path (file should update continuously)
delay	Warmup time in samples. 20000 = 40 seconds at 500Hz. Single wait at startup only
refresh	Output refresh interval in seconds. Can be set to 0 for minimal latency
offset	Calibration offset. Should give ~-0.302V at zero force

Output: CSV file with processed signals, updated in real-time

Example:

python Predict.py test0105.csv 20000 0.2 -0.302

---

Predict-KF.py (DLKF Model)

Model: DLKF pt0.795

For 500Hz sampling frequency, recommended parameters: delay=550, refresh=0.2

Usage:

python Predict-KF.py <filename> <delay> <refresh> <offset>

Arguments:

Argument	Description
filename	Input CSV file path (file should update continuously)
delay	Warmup time in samples. 550 = ~1.1 seconds at 500Hz
refresh	Output refresh interval in seconds
offset	Calibration offset for environment and sensor variations

Output: CSV file with processed signals, updated in real-time

Example:

python Predict-KF.py test0105.csv 550 0.2 -0.302