Vehicle Sensor Fusion Using Camera-Referenced Track Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle control systems struggle to accurately identify the location and distance of external objects using sensor fusion, particularly in driving assistance and autonomous driving modes, due to inconsistencies and inaccuracies in data from various sensors like LiDAR, camera, and RADAR.
Innovation Solution
A vehicle control apparatus and method that utilizes a camera as a reference sensor to set a reference angle range and distance, associating track information from RADAR and LiDAR to enhance accuracy in identifying external objects by forming a group of tracks, including visual objects, RADAR plots, and LiDAR point clouds or virtual boxes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If sensor fusion is used to identify external objects, then the coverage and detection capability are improved, but the measurement precision and reliability deteriorate due to inconsistencies from various sensors
Solution Approach 1:
The patent segments the sensor system into a reference sensor (camera) and target sensors (RADAR, LiDAR). Each sensor type processes and outputs track information independently, which is then associated through a structured framework. This segmentation allows each sensor to operate in its optimal performance range while maintaining overall system precision through the reference-based association mechanism.
Solution Approach 2:
The patent introduces a reference sensor (camera) as an intermediary that provides reference track information to mediate the association between target sensors. The reference track information acts as a common basis for associating RADAR plots and LiDAR point clouds, resolving inconsistencies by providing a unified reference frame that all sensors can align to, thereby improving measurement precision while maintaining multi-sensor coverage.
2Measurement precision
If multiple sensors are integrated for track association, then the identification accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent divides the complex sensor fusion system into distinct modules: reference sensor processing, target sensor processing, and association processing. Each module handles specific sensor types independently, reducing the complexity of data integration. The reference sensor generates reference track information that serves as a standardized input for association algorithms, simplifying the overall integration process.
Solution Approach 2:
The patent creates a universal association framework that can handle multiple sensor types (camera, RADAR, LiDAR) through a common reference-based approach. The association processing unit uses the same reference track information to associate different types of target sensor data, providing a multi-functional solution that reduces system complexity compared to separate processing pipelines for each sensor type.
3Measurement precision
If reference-based association is implemented, then the location and distance determination is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary processing by generating reference track information from the reference sensor before associating target sensor data. This reference track information is prepared in advance and serves as a ready-to-use basis for rapid association operations. By pre-processing the reference data and establishing a common reference frame beforehand, the actual association process can proceed more quickly with reduced computational overhead.
Solution Approach 2:
The patent applies local quality optimization by focusing the association processing on specific regions defined by the reference track information. Instead of processing all sensor data uniformly, the system associates target sensor plots and point clouds only within the relevant angular and distance ranges defined by the reference tracks. This localized approach reduces the overall processing time while maintaining precision in the areas that matter most for safety-critical decisions.
Data Source
AI summary
An apparatus for controlling a vehicle is introduced. The apparatus may comprise at least three different type of sensors. The first sensor may obtain visual track information about an external object within an image. The second sensor may provide track information in the form of plots corresponding to the external object. The third sensor may collect track information in clusters of points or virtual boxes corresponding to the external object. The apparatus may set a reference angle range and distance based on the first and second track information. The apparatus then may associate these with the third track information and output a signal for vehicle control.