ML-based fall detection system using Raspberry Pi (wearable system + supervised activity classification)
Other
3D Sensor
Action Recognition
3D SensorAction RecognitionBuilt an AI-based fall detection solution using wearable sensor motion data. Trained a machine learning model to classify human activities from accelerometer and gyroscope readings and generate immediate alerts for detected falls. This involved preparing labeled activity/fall outcomes for supervised model training and validating classification behavior against expected activities.• Used sensor-derived time-series features from an MPU6050/embedded setup to support supervised learning.• Defined labeling targets for activities (fall vs. non-fall) to train the classifier.• Performed model testing to confirm correct activity classification behavior.• Implemented the end-to-end pipeline to produce real-time alert outputs from predicted classes.
2024 - 2025