Multi-Sensor Calibration Using a Virtual Overall Sensor Frame
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor systems on moving objects face challenges in accurately calibrating multiple sensors with overlapping detecting ranges, leading to inconsistent and error-prone surrounding area monitoring, particularly in dynamic environments.
Innovation Solution
A method that defines a virtual overall sensor by merging the detecting ranges of individual sensors and performs sensor-to-sensor calibration followed by sensor-to-object calibration, using a common coordinate system to combine measurement data and reduce calibration errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple individual sensors with overlapping detecting ranges are used to monitor the surrounding area, then the coverage and redundancy of the sensor system are improved, but the calibration complexity and error propagation increase
Solution Approach 1:
The patent merges multiple individual sensors with overlapping detecting ranges into a single virtual overall sensor. This consolidation reduces calibration complexity by treating multiple sensors as one unified sensing entity, while maintaining the extended coverage provided by the overlapping detecting ranges of individual sensors.
Solution Approach 2:
The patent introduces a common coordinate system as an intermediary framework to transform and integrate measurement data from multiple sensors. This intermediary coordinate system enables consistent data fusion across all sensors without requiring complex pairwise calibration between each sensor, thus reducing overall calibration complexity while preserving comprehensive coverage.
2Measurement precision
If traditional sensor-to-sensor calibration is performed for each individual sensor, then the installation parameters of each sensor can be determined, but inconsistent surrounding area representation and calibration errors occur
Solution Approach 1:
The patent combines all individual sensors into a virtual overall sensor and performs a single holistic calibration. This approach ensures consistent surrounding area representation by calibrating all sensors relative to the same reference frame simultaneously, eliminating inconsistencies that arise from separate sensor-to-sensor calibrations while maintaining accurate installation parameter determination.
3Adaptability or versatility
If sensors of different types with overlapping detecting ranges are integrated, then the versatility and robustness of the sensor system are improved, but the coordinate transformation and data fusion difficulty increase
Solution Approach 1:
The patent creates a universal common coordinate system that can accommodate multiple sensor types (video, radar, lidar, ultrasonic) with different detecting ranges and characteristics. This universal framework simplifies coordinate transformations by providing a single reference frame for all sensor types, enabling seamless data fusion while maintaining the versatility benefits of multi-type sensor integration.
Data Source
AI summary
A method and device for calibrating a sensor system of a moving object. The sensor system includes a plurality of individual sensors. Each individual sensor has a particular detecting range. Each of these sensors having a detecting range at least partially overlapping with at least one further sensor of the sensor system. The method includes: defining a virtual overall sensor based on a merger of the particular detecting ranges of each individual sensor; determining first coordinates of a plurality of external objects, as well as second coordinates of selected points of the moving object; and orienting the virtual overall sensor relative to the moving object, as a function of the first and second coordinates.