Object-Level Sensor Fusion for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous vehicle systems rely on human intervention for various operations despite having automated functions like lane assist and pre-collision systems, and they struggle to effectively integrate and fuse data from heterogeneous sensors to provide comprehensive situational awareness.
Innovation Solution
A method and system that convert sensor data from multiple sensors into a common coordinate frame, predict object positions and velocities, associate detected objects with existing tracks based on kinematic, geometric, and classification information, and update these tracks to provide a fused object list, enhancing situational awareness in autonomous vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data from multiple heterogeneous sensors is integrated to improve situational awareness, then the comprehensiveness of object detection is improved, but the complexity of data processing and fusion increases
Solution Approach 1:
The system segments the complex sensor fusion task into distinct functional modules: a coordinate transformation module that converts sensor data to a common reference frame, a track prediction module that forecasts object positions, and a data association module that matches detections to tracks. This modular segmentation reduces processing complexity by handling each aspect separately rather than simultaneously
Solution Approach 2:
The patent introduces intermediate representations including predicted position/velocity/orientation as mediators between raw sensor data and final object tracks. These intermediaries simplify the association process by providing standardized comparison metrics (kinematic, geometric, and classification information) that bridge heterogeneous sensor formats
2Reliability
If automated systems automate more vehicle functions, then the reduction of human error is improved, but the extent of automation increases system complexity
Solution Approach 1:
The system implements self-service through autonomous track management where the data association module automatically determines whether to create new tracks, update existing tracks, or delete tracks based on detection data. The system serves itself by autonomously maintaining the object track database without human intervention, thereby reducing errors while managing automation complexity internally
Data Source
AI summary
Previous self-driving car systems can detect objects separately with either vision systems, RADAR systems or LIDAR systems. In an embodiment of the present invention, an object fusion module normalizes sensor output from vision, RADAR, and LIDAR systems into a common format. Then, the system fuses the object-level sensor data across all systems by associating all objects detected and predicting tracks for all objects. The present system improves over previous systems by using the data from all sensors combined to develop a single set of knowledge about the objects around the self-driving car, instead of each sensor operating separately.