Loading repository data…
Loading repository data…
sainimal1ba-hue / repository
This is is an AI-powered hearing-aid companion app that predicts sensory overload in autistic children before meltdowns occur. It analyzes early stress signals, movement, noise, and location to adapt noise cancellation and trigger calming interventions in real time.
A transparent discovery signal based on current public GitHub metadata.
This score does not audit code, security, maintainers, documentation quality, or suitability. Verify the repository and its current documentation before adoption.
This project is a predictive assistive system designed to help autistic children manage sensory overload before meltdowns occur.
Instead of reacting after stress escalates, the system analyzes physiological and environmental data in real time to predict rising stress and intervene with calming support.
This prototype runs entirely on pre-defined, simulated sensor values. No hardware (wearables, heart rate monitors, accelerometers) is required to run the demo.
data/sample_audio/sample_audio.wav or generated tones when pre-meltdown is detectedYou can observe the full pipeline (CALM → ELEVATED → PRE_MELTDOWN, intervention, soothing music) without any real sensors.
Autistic children often experience sensory overload due to background noise, movement irregularities, and stress triggers. These can lead to meltdowns that impact concentration, emotional regulation, and social interaction.
This system uses real-time signals like:
to compute a continuous stress score, detect hidden patterns, and predict pre-meltdown states.
When a high-risk pattern is detected, the app automatically plays soothing audio and adjusts noise behavior to help regulate stress before escalation.
logs/events.jsonl)git clone https://github.com/sainimal1ba-hue/AI-Powered-Sensory-Overload-Prevention-for-Autism.git
cd AI-Powered-Sensory-Overload-Prevention-for-Autism
pip install -r requirements.txt
python run.py --simulate
When pre-meltdown is detected, soothing music plays for ~25 seconds. Logs appear in the console and in logs/app_main.log.
python run.py --simulate --driver
python -m src.app.main --simulate
python -m src.app.driver_mode --simulate
├── config/ # config.yaml, logging.yaml
├── data/
│ ├── datasets/ # calm_state.csv, pre_meltdown.csv, etc.
│ └── sample_audio/ # Calming audio for intervention
├── logs/ # app_main.log, driver_mode.log, events.jsonl
├── scripts/ # run_local.py, preprocess_audio.py
├── src/
│ ├── app/ # main.py, driver_mode.py
│ ├── audio/ # noise_cancellation, calming_audio
│ ├── prediction/ # meltdown_predictor
│ ├── sensors/ # heart_rate, movement
│ └── utils/ # logger, event_logger
├── tests/
└── run.py # Entry point
The demo uses pre-defined simulated data that mimics a stress escalation:
Demo video coming soon.
Unlike existing tools that detect meltdowns after visible signs appear, this system predicts hidden stress signatures before escalation and triggers early calming support. The safety rule (reducing noise cancellation when moving at high speed) keeps users aware of their environment in vehicles or high-motion situations.
Contributions are welcome!
Submit issues & pull requests for signal processing, performance optimization, and prototype UI.
MIT License