Automotive Sensor Misalignment Detection Using Clutter Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automotive detection systems, such as radar and LiDAR, face challenges in accurately determining and correcting for sensor misalignment due to environmental influences and installation errors, which affect the precision of target bearing angle measurements.
Innovation Solution
A method and system that utilize multiple detections of clutter objects over a time interval to plot patterns on an x-y plane, determining the misalignment angle by analyzing the orientation of these patterns, which can be used to alert operators and apply calibration corrections, without relying on Doppler velocity information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vehicle dynamic information is used to verify trajectories and estimate bearing bias, then sensor misalignment can be monitored and corrected, but the technique depends strongly on the quality and availability of vehicle dynamic data
Solution Approach 1:
The patent extracts the misalignment detection function from dependence on vehicle dynamic information by using stationary clutter objects as reference points. The system plots detection patterns on an x-y plane using only sensor detection data, eliminating the need for vehicle speed, yaw rate, or steering angle information while still achieving misalignment monitoring and correction
Solution Approach 2:
The system uses stationary clutter objects in the environment as self-provided reference points for misalignment detection. These clutter objects naturally serve as calibration references without requiring external vehicle dynamic data inputs, allowing the system to self-calibrate using environmental features
2Measurement precision
If multiple sensors with overlapping fields of view are used to determine misalignment, then calibration can be performed, but the system complexity increases and requires multiple sensors
Solution Approach 1:
The patent segments the misalignment detection problem into independent analysis of detection patterns from stationary clutter objects. By analyzing the geometric patterns formed by multiple detections of the same clutter object on an x-y plane, the system achieves calibration using data from a single sensor without requiring multiple overlapping sensor fields
Solution Approach 2:
The system creates a geometric representation (copy) of the detection data by plotting patterns on an x-y plane. This copied representation allows analysis of misalignment through pattern orientation without requiring physical multiple sensors, using mathematical transformation instead of hardware multiplication
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In automotive detection systems, sensor alignment is important to proper operation of the system. A portion of automobile or vehicle 50 is equipped with detection system 10, including forward-looking detection sensor module 12C, illustrating misalignment of detection sensor module 12C. Vehicle 50 is moving with a vehicle velocity Vh. A target object 51, which is detected by detection sensor module 12C, can be present in the path of vehicle 50. However, due to misalignment error, target object 51 is detected as target object 53, which differs from the actual real target object 51. As a result, there exists a true target angle θref for the real target object 51 and a detected target angle θerr for the detected target object 53, where these angles are measured from a line 55 that extends from vehicle 50, in its forward direction, extending along the velocity Vh, which can be considered to extend along the detection boresight of the detection sensor module 12C. An automotive detection system includes a signal transmitter and a receiver, which generates receive signals indicative of reflected signals. A processor (i) receives the receive signals, (ii) processes the receive signals to generate detections of one or more objects in the region, each of the detections being associated with a position in a two-dimensional orthogonal coordinate system in a plane in which the sensor is moving, (iii) detects a pattern of detections in the two-dimensional orthogonal coordinate system by determining a quantity of detections having coordinates in a rectangular region within the two-dimensional orthogonal coordinate system, (iv) determines an angle of an axis of the rectangular region with respect to a reference direction in the two-dimensional orthogonal coordinate system, and (v) determines an angle of misalignment of the sensor from the angle of the axis of the rectangular region with respect to the reference direction. The detection system may be a radar system or a LiDAR system. Preferably, the processor filters the detections by limiting a quantity of detections in each of a plurality of two-dimensional grids within the two-dimensional orthogonal coordinate system.