Data-Driven Indoor Positioning for Low-Cost Robotics Research
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing indoor positioning systems for robotics research and education are either too expensive or lack the necessary accuracy, making them unsuitable for resource-limited environments and entry-level applications.
Innovation Solution
A low-cost indoor positioning system utilizing multiple overhead cameras and data-driven modeling algorithms, such as polynomial regression, Kriging, and machine learning, to convert camera coordinates to world coordinates, achieving cm-level accuracy with a total system cost of $300-$500.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If commercial positioning systems are used, then positioning accuracy is improved, but system cost increases significantly
Solution Approach 1:
The patent replaces expensive commercial positioning systems with inexpensive overhead cameras and fiducial markers. The system uses multiple low-cost cameras positioned overhead to capture images of fiducial markers attached to mobile robots, achieving accurate positioning without the high cost of commercial solutions. The fiducial markers are simple printed patterns that can be easily manufactured and replaced.
Solution Approach 2:
The patent substitutes complex commercial positioning hardware with a vision-based system using overhead cameras and image processing. Instead of using specialized positioning transmitters and receivers, the system captures 2D images of fiducial markers and computationally reconstructs 3D positions, replacing mechanical/electronic positioning infrastructure with optical sensing and algorithmic processing.
2Device complexity
If PDR methods are used, then system cost is reduced, but positioning accuracy deteriorates and error accumulates
Solution Approach 1:
The patent introduces overhead cameras as an intermediary reference system that provides absolute position measurements. Instead of relying solely on incremental dead reckoning that accumulates error, the system periodically observes fiducial markers through overhead cameras to reset and correct position estimates, preventing error accumulation while maintaining low cost.
Solution Approach 2:
The system implements feedback by using overhead camera observations of fiducial markers to correct and update the robot's position estimate. The observed positions from the overhead view provide ground truth measurements that feed back into the positioning algorithm to correct drift and accumulate accuracy, eliminating the error accumulation problem of pure PDR methods.
3Measurement precision
If communication-based positioning is used, then positioning accuracy is improved, but system cost and infrastructure complexity increase
Solution Approach 1:
The patent creates a 2D copy of the physical space through overhead camera images. By capturing images of fiducial markers from overhead cameras and processing these 2D image copies, the system reconstructs 3D position information without requiring physical positioning infrastructure like UWB anchors or Bluetooth beacons. The visual copy of the environment serves as the positioning reference.
Solution Approach 2:
The overhead cameras serve multiple functions: they provide positioning references for mobile robots, enable visualization of robot locations, and can potentially be used for other monitoring tasks. The fiducial markers serve both as positioning references and as visual indicators of robot presence and orientation, reducing the need for separate positioning and identification systems.
Data Source
AI summary
The disclosure deals with system and method subject matter for a low-cost, accurate indoor positioning system that integrates image acquisition and processing and data-driven modeling algorithms for robotics research and education. Multiple overhead cameras are used to obtain normalized image coordinates of ArUco markers, and presently disclosed methodology converts them to the camera coordinate frame. Various data-driven models are disclosed to establish a mapping relationship between the camera and the world coordinates. A number of data pairs (for example, 150) in the camera and world coordinates are generated by measuring the ArUco marker at different locations and then used to train and test the data-driven models. With the model, the world coordinate values of the ArUco marker and its robot carrier can be determined in real time. A straightforward polynomial regression approach can achieve a positioning accuracy of about 1.5 cm.