AI Positioning Accuracy Verification Test Mechanism

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current 5G and NR wireless telecommunication systems lack defined test mechanisms to verify the positioning accuracy of User Equipment (UE) for Positioning Reference Signal (PRS) based measurements, and there is a challenge in extracting the ground truth of actual location coordinates.

Innovation Solution

The proposed method involves transmitting a request to measure location coordinates from testing equipment to a device, receiving a model inference comprising location coordinates from the device, and determining whether the difference between the model inference and known ground truth location coordinates is within a threshold value for a defined number of samples.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If AI/ML based positioning models are deployed in UE, then positioning accuracy verification capability is improved, but device complexity increases

Engineering Contradiction:
Improvepositioning accuracy verificationVSAvoidtesting equipment complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces testing equipment as an intermediary component that coordinates between the UE's AI/ML positioning model and the verification process. The testing equipment sends measurement requests, receives model inferences, compares results with ground truth, and determines accuracy - effectively mediating the complex verification process without requiring the UE itself to handle all verification complexity

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The verification process is segmented into distinct functional components: (1) sending measurement requests, (2) receiving model inferences with location coordinates, (3) comparing inferences against ground truth, and (4) determining accuracy based on threshold criteria. This segmentation allows each component to be implemented independently and simplifies the overall system architecture

Inventive Principle:
Principle #1Segmentation

2Reliability

If AI/ML model performance monitoring is implemented in real-time, then positioning reliability is improved, but processing time increases

Engineering Contradiction:
Improvepositioning accuracyVSAvoidverification time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent implements a feedback mechanism where the testing equipment continuously monitors AI/ML model performance by comparing model inferences against ground truth location coordinates. The system provides real-time feedback on positioning accuracy by determining whether the difference between model inference and ground truth falls within threshold values, enabling continuous reliability verification without significant time penalty

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The verification process applies partial action by checking positioning accuracy against predefined threshold values for a defined number of samples rather than verifying every single measurement. This approach provides sufficient reliability assurance while reducing the total processing time required for comprehensive verification

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS20250048319A1Test mechanism for artificial intelligence/machine learning based positioning accuracy verification
Publication Date: 2025.02.06 NOKIA TECHNOLOGIES OY
  • US20250048319A1 patent drawing
  • US20250048319A1 patent drawing
  • US20250048319A1 patent drawing

AI summary

Systems, methods, apparatuses, and computer program products for verifying positioning accuracy. One method may include transmitting, by testing equipment, a request to measure location coordinates to a device, receiving, by the testing equipment, from the device, a model inference comprising at least one location coordinate, and determining, by the testing equipment, whether a difference between the model inference and at least one ground truth is within a threshold value for a defined number of samples, wherein the at least one ground truth comprises at least one known location coordinate of a test point.