|
|
d10c385bdc
|
feat: Add Issue #429 - Emotion engine with context-aware face expression selection
- ROS2 node subscribing to orchestrator state, battery, balance, person tracker, voice commands, health
- State-to-emotion mapping: navigation → excited, social → happy/curious, low battery → concerned, etc.
- Smooth emotion transitions (0.3–1.2s) with confidence tracking
- Idle behaviors: blink (~3s), look-around (~8s), breathing (sine wave)
- Social memory: familiarity-based warmth modifier (0.3–1.0) for known people
- Personality-aware responses: extroversion, playfulness, responsiveness, anxiety (0.0–1.0 configurable)
- Publishes /saltybot/emotion_state (JSON): emotion, intensity, confidence, expression name, context, idle_flags
- Configurable via emotion_engine.yaml: personality traits, battery thresholds, update rate
- Launch file: emotion_engine.launch.py
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
|
2026-03-05 08:53:42 -05:00 |
|
|
|
1878b09eee
|
feat: Add audio direction estimator (Issue #430)
Implements GCC-PHAT beamforming for sound source localization via Jabra mic.
- GCC-PHAT cross-correlation for direction of arrival (DoA) estimation
- Voice activity detection (VAD) using RMS energy + smoothing
- Stereo/quadrophonic channel support (left/right/front/rear estimation)
- ROS2 publishers: /saltybot/audio_direction (Float32 bearing), /saltybot/audio_activity (Bool VAD)
- Configurable parameters: sample_rate, chunk_size, publish_hz, vad_threshold, gcc_phat_max_lag
- Integration-ready for multi-person tracker speaker tracking
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
|
2026-03-05 08:53:25 -05:00 |
|
