name: sensor-fusion description: Multi-sensor fusion algorithms for perception in autonomous driving allowed-tools:
- Read
- Write
- Glob
- Grep
- Edit
- WebFetch
- WebSearch
- Bash
metadata:
version: "1.0"
category: automotive-engineering
tags:
- adas
- autonomous-driving
- perception
- sensor-fusion
Sensor Fusion Skill
Purpose
Enable multi-sensor fusion algorithm development for autonomous driving perception including object detection, tracking, and environmental modeling.
Capabilities
- Camera, radar, lidar data preprocessing
- Object detection fusion algorithms
- Tracking filter implementation (Kalman, EKF, UKF)
- Association algorithms (Hungarian, GNN, JPDA)
- Occupancy grid fusion
- Confidence estimation and sensor weighting
- Time synchronization handling
- Ground truth comparison and metrics
Usage Guidelines
- Preprocess sensor data for consistent coordinate frames
- Select appropriate tracking filters based on object dynamics
- Implement robust association for multi-target scenarios
- Fuse sensor confidence for reliable perception
- Handle time delays and synchronization issues
- Validate fusion against ground truth data
Dependencies
- ROS/ROS2
- TensorFlow
- PyTorch
- NVIDIA DriveWorks
Process Integration
- ADA-001: Perception System Development
- ADA-002: Path Planning and Motion Control
- ADA-003: ADAS Feature Development
- ADA-004: Simulation and Virtual Validation