Ordinal UNLOC Indoor Localization via Ordinal Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Indoor localization is challenging due to unreliable distance measures caused by physical obstacles and dynamically changing conditions, making it difficult to achieve accurate location estimation in indoor environments without existing infrastructure like GPS systems.
Innovation Solution
The Ordinal UNLOC framework uses ordinal comparison data from pairwise distance measurements between sensors to estimate target locations, transforming dissimilarities into distances through machine learning and optimization techniques, eliminating the need for direct distance measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If direct distance measurements are used for indoor localization, then location estimation can be performed, but the measurements become unreliable due to physical obstacles and dynamically changing conditions
Solution Approach 1:
Instead of directly measuring distances and using them for localization, the patent inverts the approach by measuring ordinal relationships (which sensor is closer) and then inferring distance information from these ordinal comparisons. This inversion allows the system to avoid the unreliability of direct distance measurements while still obtaining useful distance-related information for localization.
Solution Approach 2:
The patent introduces ordinal comparison data as an intermediary between the raw sensor measurements and the final distance estimates. Rather than directly using potentially erroneous distance measurements, the system uses ordinal relationships as a mediator that is more robust to environmental disturbances, then transforms these ordinal data into distance information through statistical methods.
2Ease of operation
If wireless sensor networks with anchors are deployed for indoor localization, then location estimation becomes possible, but the distance measures inferred from signal proxies remain unreliable in indoor environments
Solution Approach 1:
The patent replaces the conventional approach of using signal strength or time-of-flight proxies for distance measurement with a statistical ranking system. Instead of relying on physical signal propagation characteristics that are distorted by indoor environments, the system uses ordinal comparisons and statistical inference to estimate distances, substituting a computational approach for physical measurement methods.
3Loss of information
If conventional localization methods using noisy distance measures are applied, then target location can be estimated, but the presence of walls, furniture, and people makes distance measurement unreliable
Solution Approach 1:
The patent performs preliminary ordinal comparisons between all sensor pairs before attempting to estimate distances or locations. By first establishing the ordinal relationships (which sensor is closer to which target) and then using these pre-established rankings to inform distance estimation, the system prepares robust comparative data that is less susceptible to the effects of indoor obstacles before the actual localization computation occurs.
Data Source
AI summary
A method for determining location of a target within an indoor environment, including the steps of: classifying a set of anchors having known locations within the indoor environment and a set of targets having unknown locations within the indoor environment, wherein each of the anchors and targets comprise hardware having sensors and wireless communication capabilities; creating a set of ordinal pair data sets comprising relative distances between each target and all anchors; ranking and aggregating the ordinal pair data sets to produce a set of dissimilarities that approximate distances; transforming the dissimilarities into estimated distances between each anchor and target using the known distances between the anchors as calibration; and inferring location of targets by formulating and solving a multidimensional unfolding optimization.