Multiaxis Sensor Calibration for Accurate 3D Vehicle Measurements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional sensor calibration methods for vehicles, such as autonomous cars, drones, and robots, are time-consuming and costly due to individual sensor calibration, and two-dimensional rotations fail to accurately capture three-dimensional data variations, especially for IMUs, leading to potential errors and inflexibility.
Innovation Solution
A multiaxis rotating platform capable of three-dimensional rotation around yaw, pitch, and roll axes, combined with fiducial targets and a scene tracking system, allows simultaneous calibration of multiple sensors by estimating calibration parameters like orientation, position, and synchronization offsets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensors are calibrated individually using conventional methods, then each sensor can be calibrated with focused attention, but the process becomes time-consuming and costly
Solution Approach 1:
The patent combines multiple sensor calibration processes into a single simultaneous calibration operation. The multiaxis rotating platform rotates the vehicle through three-dimensional space while fiducial targets are tracked by a camera system, allowing extrinsic parameter calibration for multiple sensors (cameras, lidar, radar, IMUs) to occur concurrently rather than sequentially, thereby reducing total calibration time while maintaining accuracy
Solution Approach 2:
The calibration system is designed to calibrate multiple types of sensors simultaneously using a universal approach. The same rotating platform and fiducial target system can calibrate cameras, lidar, radar, and IMUs in one operation, making the process more efficient without sacrificing the specific calibration needs of each sensor type
2Device complexity
If two-dimensional rotation on a turntable is used for calibration, then the setup is simple, but it fails to capture three-dimensional data variations and compromises calibration accuracy
Solution Approach 1:
The patent transitions from two-dimensional rotation (yaw only on a turntable) to three-dimensional rotation by adding pitch and roll axes. The multiaxis rotating platform enables the vehicle to be rotated around multiple axes simultaneously, creating three-dimensional motion that properly excites all sensor measurements and allows accurate calibration of extrinsic parameters for sensors like IMUs that require three-dimensional motion data
Solution Approach 2:
The system uses dynamic three-dimensional rotation rather than static positioning. The multiaxis platform continuously rotates the vehicle through various orientations, allowing the calibration system to capture sensor data across a full range of motions. This dynamic approach ensures that all sensor parameters can be accurately determined through motion-based calibration
3Ease of manufacture
If vehicle-only rotation is used for calibration, then the system is simple to implement, but it introduces inflexibility for large or heavy-duty vehicles with limited maneuverability
Solution Approach 1:
Instead of rotating the vehicle itself (which is difficult for large vehicles), the patent inverts the approach by rotating the calibration platform underneath the vehicle. The multiaxis rotating platform supports and rotates the entire vehicle assembly, allowing even heavy-duty vehicles with limited maneuverability to be calibrated. This inversion of the rotation mechanism solves the flexibility problem while maintaining system simplicity
Data Source
AI summary
The present disclosure relates to sensor calibration systems for sensor-equipped units, such as vehicles. The system includes a multiaxis rotating platform or an assembly of platforms capable of rotation and tilt in three-dimensional space around yaw, pitch, and roll axes. The system also features a plurality of fiducial targets positioned on and around a rotating platform, with some fiducials rotating with the platform, along with a scene tracking system to monitor and track the fiducial targets and the sensor-equipped unit. In one embodiment, the sensor-equipped unit rotates around a set of fiducial targets, while in another, a set of fiducial targets rotate around the sensor-equipped unit. A controller, coupled to the sensors, platforms, and scene tracking system, receives and processes data from these components. It analyzes the data and determines at least one calibration parameter from the processed data to calibrate one or more sensors.