Vehicle Sensor Calibration Using Stationary Object Poses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle sensor calibration methods require specific targets and straight-line or stationary vehicle movements, limiting their applicability and efficiency in real-world scenarios.
Innovation Solution
A method that calibrates vehicle sensors using data from stationary objects while the vehicle changes pose, allowing calibration during regular use and eliminating the need for specific calibration targets, by determining calibration parameters or vehicle relative poses based on sensor data from different vehicle orientations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing calibration methods use specific calibration targets and require straight-line or stationary vehicle movement, then calibration accuracy can be maintained, but the applicability and efficiency in real-world scenarios deteriorates
Solution Approach 1:
The system uses stationary objects already present in the real-world environment (such as buildings, trees, or other fixed structures) as calibration references, eliminating the need for dedicated calibration targets. The vehicle's own sensor suite and motion data are used to perform calibration autonomously during normal operation, making the calibration process self-sufficient and adaptable to real-world conditions without requiring special equipment or controlled environments
Solution Approach 2:
The calibration method transitions from requiring static or constrained vehicle movement (straight-line or stationary) to utilizing dynamic, arbitrary vehicle motion patterns. By processing sensor data collected during natural driving maneuvers with changing poses and orientations, the system achieves calibration that is both accurate and applicable to real-world operational conditions where vehicles move dynamically
2Measurement precision
If existing calibration methods require specific calibration targets and controlled movement patterns, then calibration can be performed with sufficient precision, but the time and operational disruption increases
Solution Approach 1:
The calibration process is integrated into normal vehicle operation, allowing calibration to occur continuously during regular driving rather than requiring separate, dedicated calibration sessions. The system processes sensor data from stationary objects encountered during normal routes, enabling calibration to proceed without interrupting the vehicle's useful operational activities
Solution Approach 2:
The system continuously collects and processes sensor data from stationary objects during normal operation, preparing calibration information in advance. This allows calibration parameters to be determined proactively during routine driving, so that when calibration is needed, the process can be completed quickly using pre-processed data rather than requiring time-consuming controlled calibration maneuvers
Data Source
AI summary
A computer is programmed to receive first sensor data from a sensor of a vehicle indicating a first relative position of a stationary object, the first relative position detected while the vehicle is in a first vehicle pose; receive second sensor data from the sensor indicating a second relative position of the stationary object, the second relative position detected while the vehicle is in a second vehicle pose having a different orientation than the first vehicle pose; and determine one of a calibration parameter or a vehicle relative pose based on the first relative position, the second relative position, and the other of the calibration parameter or the vehicle relative pose. The calibration parameter defines a sensor pose of the sensor relative to the vehicle. The vehicle relative pose defines a transformation of the vehicle from the first vehicle pose to the second vehicle pose.