This project focuses on designing a self-adaptive differential amplifier circuit that compensates for environmental variations like voltage and temperature changes. Using a combination of BJT-based differential amplifier design and machine learning integrated with an Arduino board, the system dynamically adjusts resistor values to ensure consistent gain and performance.
- Introduction
- Explains the role and importance of differential amplifiers in signal processing and noise rejection.
- Applications
- Includes usage in audio systems, biomedical devices (e.g., ECG), and noise-canceling headphones.
- Challenges Addressed
- Mitigating temperature and voltage variations using ML predictions.
- Proposed Solution
- Implementation of a machine learning-driven system using a linear regression model to predict optimal resistor values based on voltage and temperature conditions.
- Core: Differential amplifier with two BC547B transistors.
- Components:
- Q1 and Q2: Transistors for differential operation.
- Biasing Resistors: Ensure proper operating points.
- Voltage and Temperature Sensors: Measure environmental parameters.
- Software Tools:
- LTSpice: Circuit simulation.
- Python: Training the linear regression model.
- Arduino IDE: Implementing ML predictions on hardware.
- Hardware Components:
- Arduino Uno, DHT11 temperature sensor, and voltage sensor.
- Trained a linear regression model on 12,221 data points generated via LTSpice.
- Model Equation:
Gain = 4.1623 * DC + 0.0058 * Temp - 14.6694 - Result: Accurate gain predictions based on environmental conditions.
- Amplifier Performance:
- Gain measurement aligns with ML predictions.
- Noise rejection meets expected standards.
- Arduino Integration:
- Successfully implemented real-time ML-based gain adjustments.
| Team Member | Contribution |
|---|---|
| Kashik P S | ML model development, data generation in LTSpice. |
| Sravani | Circuit design, simulation, hardware testing. |
| Ved Prakash | Sensor calibration, simulation, hardware testing. |
| Rohan Kumar | ML integration, Arduino coding, hardware testing. |
Find project code, videos, and additional resources:
Project Resources Link
- Temperature Impact: Gain prediction at different temperatures (e.g., 27°C, 40°C).
- Voltage Sensor Calibration: Limited by DC-only capability.
- Data Generation: Time-intensive simulation for ML training.
This README provides an overview of the ML-Driven Self-Adaptive Differential Amplifier project, highlighting its goals, methods, and results. For further details, refer to the linked resources or contact the contributors.