Vehicle Radar Calibration via 3D Coordinate Transformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing calibration methods for vehicle-mounted radar systems, which combine millimeter-wave and laser radars, fail to accurately convert 2D measurements from millimeter-wave radars to 3D measurements, leading to inaccuracies in obstacle detection due to the missing dimension in 3D measurement.
Innovation Solution
A method and apparatus that calibrate the vehicle-mounted radar system by detecting point cloud data and two-dimensional data sets using both radars, determining calibrated turning angle, displacement, and vertical coordinates to convert millimeter-wave radar data into laser radar coordinates, enabling accurate obstacle detection by integrating point cloud and converted two-dimensional data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing calibration methods are used to convert millimeter-wave radar data to laser radar coordinates, then the calibration process can be completed, but the measurement precision deteriorates due to the missing dimension in 2D to 3D conversion
Solution Approach 1:
The patent introduces a vertical dimension (z-axis) to transform the 2D millimeter-wave radar coordinate system into a 3D coordinate system compatible with the laser radar. By adding the vertical coordinate information and performing 3D coordinate transformation, the patent resolves the dimensionality mismatch between the two radar systems, enabling accurate obstacle detection while maintaining reasonable algorithmic complexity.
2Reliability
If millimeter-wave radar and laser radar are integrated for obstacle detection, then the reliability of detection is improved, but the device complexity increases due to coordinate system calibration requirements
Solution Approach 1:
The patent employs an optimization algorithm that uses feedback from the coordinate transformation process to iteratively adjust and refine the calibration parameters. By minimizing the deviation between transformed millimeter-wave radar data and laser radar data through repeated optimization cycles, the system achieves accurate calibration while managing the complexity through systematic feedback-driven parameter adjustment.
3Measurement precision
If 2D millimeter-wave radar data is directly used without dimension conversion, then the processing speed is maintained, but the measurement precision deteriorates due to inability to provide accurate 3D obstacle information
Solution Approach 1:
The patent performs preliminary coordinate transformation and vertical coordinate assignment to the millimeter-wave radar data before integration with laser radar data. By pre-processing the 2D data into 3D format using the calibrated transformation parameters, the system ensures that the data is ready for accurate 3D obstacle detection without requiring complex real-time conversion during obstacle detection, thus maintaining processing efficiency while achieving 3D accuracy.
Data Source
AI summary
An obstacle detection method and apparatus for a vehicle-mounted radar system. The method includes: detecting a first point cloud data set of calibration objects located at a plurality of preset positions and detecting a first two-dimensional data set of the calibration objects; calibrating a vehicle-mounted radar system based on preset mounting positions, the first point cloud data set, and the first two-dimensional data set, to obtain a calibrated turning angle difference, a calibrated displacement difference, and a calibrated vertical coordinate; detecting a second point cloud data set of an obstacle and detecting a second two-dimensional data set of the obstacle; converting the second two-dimensional data set into a laser radar coordinate system based on the calibrated turning angle difference, the calibrated displacement difference, and the calibrated vertical coordinate, to obtain converted two-dimensional coordinates; and integrating the second point cloud data set and the converted two-dimensional coordinates to determine a position of the obstacle.