Multi-Camera Pose Measurement for GNSS-Denied Workspaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to accurately determine the position and orientation of objects indoors or in areas with blocked satellite signals, as they rely on GNSS or require targets for measurement, which are inefficient and unsuitable for working volumes.
Innovation Solution
A system using a set of mobile cameras and installed or projected targets to measure angles and calculate the pose of an object with high accuracy and update rate, employing weighted least squares estimation and inertial sensors for error correction, suitable for both indoor and outdoor applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GNSS is used to determine position and orientation, then measurement accuracy is improved (1 cm position, sub-degree orientation), but the system fails indoors or in areas with blocked satellite signals
Solution Approach 1:
The system divides the measurement task into multiple components: multiple cameras capture images from different perspectives, targets are distributed throughout the workspace at known locations, and computational algorithms process the data. This segmentation allows the system to achieve GNSS-level accuracy without requiring satellite signals, enabling operation in indoor environments.
Solution Approach 2:
The patent introduces targets as intermediary objects placed at known locations in the workspace. These targets serve as reference points that the cameras observe to calculate position and orientation. This intermediary approach replaces the direct satellite-to-receiver signaling of GNSS with a camera-to-target observation system, enabling operation in signal-blocked environments while maintaining measurement precision.
2Measurement precision
If laser range finder with azimuth and elevation encoders is used, then position and orientation can be determined, but the system is inefficient
Solution Approach 1:
The patent merges multiple measurement functions into a single camera-based system. Instead of using separate laser range finders, azimuth encoders, and elevation encoders, the system combines position and orientation determination into one integrated camera system that observes multiple targets simultaneously. This merging eliminates the inefficiencies of the separate-component approach while maintaining measurement capability.
Solution Approach 2:
The system uses periodic image capture at high frame rates to continuously track position and orientation. Rather than relying on continuous laser scanning with mechanical encoders, the camera system periodically captures images and computes pose information, achieving higher update rates and efficiency while maintaining precision.
3Measurement precision
If targets are used to determine position, then measurement can be performed, but the system is not suitable for determining position in the working volume
Solution Approach 1:
The patent transitions from using targets solely as position references to using them as three-dimensional spatial markers throughout the working volume. By distributing targets in 3D space and using multiple cameras to observe them from different angles, the system creates a volumetric measurement capability rather than just planar positioning. This dimensional approach enables accurate position determination anywhere within the working volume, not just at target locations.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables precise and rapid determination of object pose in various dynamic applications, both indoors and outdoors, with high accuracy and update rates, using camera-based measurements and inertial sensors to account for errors and interference.
Implementation Method 1
image data from the plurality of cameras
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and a method for determining a position and an orientation of an object relative to a defined reference frame is disclosed. A plurality of targets are disposed at known locations relative to the defined reference frame. A head assembly is disposed on the object, where the head assembly includes a plurality of cameras. The head assembly determines the position and the orientation of the object from location data associated with the plurality of targets and image data from the plurality of cameras.