Vehicle-Mounted Sensor Calibration Using Natural Vehicle Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current extrinsic calibration methods for vehicle-mounted sensors, such as LiDAR and 3D cameras, are inaccurate due to manufacturing tolerances and require external calibration targets, which are impractical for high-accuracy tasks in unconstrained 3D environments.
Innovation Solution
Utilize natural vehicle features, like outriggers and the floor, to perform extrinsic calibration by segmenting 3D models from point cloud data, correcting translation and rotation, and determining the sensor's actual position and orientation without external targets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If external calibration targets are used for sensor calibration, then calibration accuracy can be improved, but the complexity of the calibration process and the need for special equipment increases
Solution Approach 1:
The calibration system uses the vehicle's own natural features (outriggers, floor, body panels) as calibration references, eliminating the need for external calibration targets and special equipment. The vehicle calibrates itself by detecting its own structural features through the sensor, thereby resolving the contradiction between calibration accuracy and process complexity.
Solution Approach 2:
The patent introduces a calibration system that uses naturally occurring vehicle features as intermediaries between the sensor and the calibration process. These natural features serve as the mediating reference objects that enable accurate calibration without requiring external targets, thus reducing complexity while maintaining precision.
2Measurement precision
If external calibration targets are required, then calibration precision can be maintained, but the adaptability to different environments and vehicles is reduced
Solution Approach 1:
The calibration system is designed to work with universal natural features that exist on all vehicles (outriggers, floor, body panels), making the calibration process adaptable to different vehicle types and environments without requiring vehicle-specific or environment-specific calibration targets. This universal approach maintains precision while enhancing versatility.
Solution Approach 2:
Each vehicle uses its own unique natural features for calibration, making the system inherently adaptable to different vehicles and environments. The self-service nature of using vehicle-specific features ensures that calibration precision is maintained across diverse applications without requiring external standardized targets.
3Ease of manufacture
If manufacturing tolerances are considered in sensor mounting, then calibration accuracy deteriorates, but the ease of installation and manufacturing is improved
Solution Approach 1:
The calibration system uses feedback from detecting the actual positions of natural vehicle features to compensate for manufacturing tolerances in sensor mounting. By measuring the actual geometry of features like outriggers and floor panels, the system calculates correction transforms that account for installation variations, thereby maintaining high calibration accuracy despite ease of manufacture.
Solution Approach 2:
The system changes the calibration parameters by using actual measured positions of natural features rather than nominal design positions. This allows the calibration to adapt to actual manufacturing variations and installation tolerances, maintaining accuracy while allowing for easier manufacturing and installation processes.
Data Source
AI summary
In accordance with one aspect of the inventive concepts, provided is an autonomous mobile robot, comprising: at least one processor in communication with at least one computer memory device: at least one vehicle-mounted sensor configured to acquire measurement data of at least one feature of the robot; and an extrinsic calibration system comprising computer program code executable by the at least one processor to register the measurement data to a common coordinate frame to perform extrinsic calibration of the at least one sensor. The measurement data can comprise point cloud data.