feat: Multi-sensor fusion (Issue #490) #498

Merged
sl-jetson merged 3 commits from sl-perception/issue-490-sensor-fusion into main 2026-03-05 17:24:04 -05:00
Collaborator
No description provided.
sl-jetson added 3 commits 2026-03-05 17:23:54 -05:00
Complete over-the-air (OTA) firmware update system with:

Features:
- Downloads releases from Gitea (seb/saltylab-firmware)
- Automatic colcon build in staging directory
- Symlink-based atomic deployment
- ROS2 service restart via systemd
- Automatic rollback on build failure
- Version tracking in ~/.saltybot-data/versions.json
- Update history with timestamps

Safety:
- Blocks updates if robot velocity > 0.05 m/s
- Velocity monitoring via odometry subscription
- Backup before update for recovery

Triggers:
- MQTT /saltybot/ota_command: 'check', 'update:<version>', 'rollback'
- /saltybot/ota_status: JSON status updates
- Dashboard integration ready

Configuration:
- Gitea API base, repo info, directories
- Build timeout: 3600s (1 hour)
- Service restart automation
- Backup retention policy

ROS2 package structure complete with launch files and config.
Implement context-aware text-to-speech with emotion-driven expression for SaltyBot.

Features:
  ✓ Context-aware greetings (time of day, person names, emotion)
  ✓ Priority queue management (safety > social > idle)
  ✓ Emotion-based rate/pitch modulation (happy: faster+higher, sad: slower+lower)
  ✓ Integration with emotion engine (Issue #429) and TTS service (Issue #421)
  ✓ Configurable personality parameters
  ✓ Person recognition for personalized responses
  ✓ Queue management with 16-item buffer

Architecture:
  Node: tts_personality_node
    - Subscribes: /saltybot/tts_request, /saltybot/emotion_state, /saltybot/person_detected
    - Publishes: /saltybot/tts_command (formatted for TTS service), /saltybot/personality_state
    - Runs worker thread for asynchronous queue processing

Personality Parameters:
  - Name: "Luna" (default, configurable)
  - Speed modulation: happy=1.1x, sad=0.9x, neutral=1.0x
  - Pitch modulation: happy=1.15x, sad=0.85x, neutral=1.0x
  - Time-based greetings for 4 periods (morning, afternoon, evening, night)
  - Known people mapping for personalization

Queue Priority Levels:
  - SAFETY (3): Emergency/safety messages
  - SOCIAL (2): Greetings and interactions
  - IDLE (1): Commentary and chatter
  - NORMAL (0): Default messages

Files Created:
  - saltybot_tts_personality package with main personality node
  - config/tts_personality_params.yaml with configurable parameters
  - launch/tts_personality.launch.py for easy startup
  - Unit tests for personality context and emotion handling
  - Comprehensive README with usage examples

Integration Points:
  - Emotion engine (Issue #429): Listens to emotion updates
  - TTS service (Issue #421): Publishes formatted commands
  - Jabra SPEAK 810: Output audio device
  - Person tracking: Uses detected person names

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- saltybot_sensor_fusion: ROS2 node for LIDAR + depth sensor fusion
- Fuses RPLIDAR A1M8 (360° 2D) + RealSense D435i (front 87° 3D)
- Message filters for time-synchronized sensor inputs
- Smart blind spot handling: rear/sides LIDAR-only, front uses both
- Publishes /scan_fused (unified LaserScan) + PointCloud2 for voxel layer
- Configurable front sector angle (±45°), range multiplier, max range limit
- Parameters: depth_range_multiplier=0.9 (safety margin), max_range=5m

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
sl-jetson merged commit 5f6a13ccca into main 2026-03-05 17:24:04 -05:00
Sign in to join this conversation.
No Reviewers
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: seb/saltylab-firmware#498
No description provided.