7.7.16. PIR Detection

Detect presence and motion using PIR sensor data.

7.7.16.1. Overview

This example demonstrates presence and motion detection using Passive Infrared (PIR) sensor data. Unlike simple threshold-based detection, this ML approach can distinguish between different types of motion and reduce false positives.

Application: Security systems, occupancy sensing, smart lighting, IoT

Task Type: Time Series Classification

Data Type: Multivariate (PIR sensor readings)

7.7.16.2. Configuration

common:
  target_module: 'timeseries'
  task_type: 'generic_timeseries_classification'
  target_device: 'CC2755'  # Optimized for wireless PIR applications

dataset:
  dataset_name: 'pir_detection'

training:
  model_name: 'PIRDetection_model_1_t'
  training_epochs: 50
  batch_size: 32

testing: {}
compilation: {}

7.7.16.3. Running the Example

cd tinyml-modelzoo
./run_tinyml_modelzoo.sh examples/pir_detection/config.yaml

7.7.16.4. Dataset Details

Input Variables:

  • PIR sensor analog readings

  • Multiple PIR channels (if available)

Classes:

  • No presence

  • Presence detected

  • Motion detected (optional sub-classes)

7.7.16.6. See Also