Multi-Sensor Calibration Using Lane Markers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-sensor calibration techniques for autonomous vehicles are time-consuming, require complex setup, and lack accuracy due to reliance on target calibration boards and pairwise calibration methods, which are cumbersome and prone to errors, especially when calibrating multiple cameras and LiDAR sensors with inertial measurement unit-global navigation satellite system (IMU-GNSS) sensors.
Innovation Solution
A method for simultaneous multi-sensor calibration using lane markers, where sensors like cameras and LiDARs obtain data on lane markers, calculate extrinsic parameters by comparing pixel and 3D world coordinates, and utilize a processor to estimate these parameters based on loss functions, correlating point cloud data with a point cloud map reference and IMU-GNSS measurements to achieve precise calibration without the need for multiple target calibration boards.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If target calibration boards and pairwise calibration methods are used, then calibration can be performed, but the process becomes time-consuming and complex
Solution Approach 1:
The patent extracts the calibration target from artificial boards to natural lane markers on the road. By using lane markers that already exist in the driving environment, the system eliminates the need for separate calibration boards and pairwise calibration procedures, significantly reducing calibration time while maintaining accuracy through direct multi-sensor observation of the same natural target
Solution Approach 2:
The patent makes the calibration system universal by using lane markers that serve both as road guidance indicators and as calibration targets. This multi-functional approach allows the same road infrastructure to provide navigation information and sensor calibration data, eliminating the need for dedicated calibration equipment and simplifying the overall calibration process
2Measurement precision
If target calibration boards are used, then calibration can be performed, but the setup becomes complex and cumbersome
Solution Approach 1:
The patent removes the complex calibration board setup and replaces it with naturally occurring lane markers. This extraction eliminates the need for physical calibration boards, mounting structures, and complex positioning equipment, reducing setup complexity while maintaining calibration functionality through direct observation of lane markers by multiple sensors
Solution Approach 2:
The patent enables the calibration system to use existing road infrastructure (lane markers) that already exists in the environment. The lane markers serve themselves as both road guidance elements and calibration targets, eliminating the need for external calibration equipment and simplifying the setup process to only require the vehicle's own multi-sensor system
3Productivity
If pairwise calibration methods are used for multiple sensors, then calibration can be performed, but errors increase and accuracy decreases
Solution Approach 1:
The patent merges multiple pairwise calibration operations into a single multi-sensor calibration process. By having all sensors (cameras, LiDAR, radar) observe the same lane markers simultaneously and perform joint optimization, the system eliminates error accumulation from sequential pairwise calibration while improving overall calibration efficiency and accuracy
Solution Approach 2:
The patent implements feedback mechanisms where the system continuously optimizes extrinsic parameters by comparing observations from multiple sensors with ground truth lane marker positions. This feedback loop allows the system to correct errors and refine calibration accuracy through iterative optimization, rather than relying on error-prone pairwise methods
Data Source
AI summary
Techniques for performing multi-sensor calibration on a vehicle are described. A method includes obtaining, from each of at least two sensors located on a vehicle, sensor data item of a road comprising a lane marker, extracting, from each sensor data item, a location information of the lane marker, and calculating extrinsic parameters of the at least two sensors based on determining a difference between the location information of the lane marker from each sensor data item and a previously stored location information of the lane marker.