Indoor Wayfinding Using Visual Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Indoor wayfinding applications face challenges in accurately determining a user's starting position and orientation due to signal changes from WIFI access points and Bluetooth beacons, which can be affected by environmental factors and beacon placement, leading to faulty navigation.
Innovation Solution
A wayfinding system that identifies objects at known locations using camera image information to determine device pose and orientation, generating pathway instructions based on spatial relationships between enclosed spaces without relying on indoor positioning systems or floor plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If indoor positioning systems using WIFI access points and Bluetooth beacons are used to determine user position, then wayfinding capability is enabled, but positioning accuracy deteriorates due to signal changes from environmental factors and beacon placement
Solution Approach 1:
The patent replaces the wireless signal-based positioning system (WIFI/Bluetooth) with a visual recognition system using camera images and object databases. Instead of relying on radio signal strength and triangulation, the system uses image processing to identify objects and determine device pose, fundamentally substituting the positioning mechanism to eliminate signal interference issues.
Solution Approach 2:
The patent introduces object image information as an intermediary between the user device and the positioning system. Rather than directly measuring position through signals, the system uses captured images as intermediaries to query a database of object locations and orientations, indirectly determining device position through visual feature matching.
2Reliability
If indoor positioning systems are deployed to enable navigation, then wayfinding functionality is provided, but system complexity increases due to infrastructure requirements
Solution Approach 1:
The patent enables the mobile device to perform positioning autonomously using its own camera and processing capabilities. The device captures images, processes them locally or with minimal server assistance, and determines its position without requiring external positioning infrastructure like WIFI access points or Bluetooth beacons to be installed and maintained.
Solution Approach 2:
The patent extracts the positioning functionality from the building infrastructure and relocates it to the user's mobile device. Instead of requiring the building to be equipped with positioning infrastructure, the positioning capability is extracted and embedded in the portable device, eliminating the need for complex installed systems.
3Measurement precision
If floor plans and interior mapping are used to generate navigation instructions, then route guidance accuracy is improved, but system complexity and data requirements increase
Solution Approach 1:
The patent segments the interior space representation into discrete object entities with known positions and orientations, rather than requiring complete floor plan mappings. Instead of storing and processing entire building layouts, the system uses a database of individual object image information and locations, breaking down the navigation problem into object recognition and matching tasks.
Solution Approach 2:
The patent uses copied object image information from a database to represent physical objects in the environment. Rather than requiring direct access to physical floor plans or detailed interior mappings, the system uses digital copies of object images and their associated position data to create a virtual representation sufficient for navigation purposes.
Data Source
AI summary
An indoor wayfinding application operates to generate pathway instructions based on knowledge about a mobile device user orientation, a starting location, a destination location and based upon knowledge of spatial inter-relationships and dimensions of enclosed spaces in a building. The user orientation is determined by an image comparison algorithm operating on a query image captured by a camera and a reference image, and the starting location is determined by comparing the query image to a plurality of reference images, each one of which is at a known location, for a best match.