Signal Sensei
ML-driven signal processing for Forward Error Correction code identification.
PythonTensorFlowSignal ProcessingRNN
Overview
Signal Sensei is a machine learning system developed for the Smart India Hackathon 2023 that identifies Forward Error Correction (FEC) codes from blind signal data. The system uses recurrent neural networks to classify signal patterns without prior knowledge of the encoding scheme, enabling signal intelligence applications.
The pipeline ingests raw signal data, applies preprocessing transforms, and feeds features into an RNN classifier that has been trained on synthetic and real-world FEC-encoded signals. The project reached the finalist stage at SIH 2023.
Key Achievements
- SIH 2023 Finalist
- RNN-based FEC code classification from blind signals
- End-to-end signal processing pipeline from raw data to classification
Tech Stack
- Python -- primary language for ML pipeline development
- TensorFlow -- deep learning framework for RNN training and inference
- Signal Processing -- domain-specific transforms for feature extraction
- RNN -- recurrent architecture suited for sequential signal pattern recognition