UWB Position Prediction With ML Correction for Indoor Robotics

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

RTK GPS systems are expensive, power-hungry, and less effective in indoor or obstructed environments, while classical UWB trilateration is prone to inaccuracies due to multipath propagation and interference, limiting their use in cost-effective and accurate robotic platform positioning.

Innovation Solution

A computer-implemented method using machine-learning models to correct UWB-measured distance and pose data, assigning quality indicators, and optimizing a cost function to determine the most-likely position and pose of an autonomous robotic platform, utilizing a local coordinate system based on UWB beacons, which can include dynamic transceivers and anchors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If RTK GPS systems are used for positioning, then positioning accuracy is improved, but cost and power consumption increase

Engineering Contradiction:

Improvepositioning accuracyVSAvoidsystem cost

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces expensive RTK GPS systems with cheaper UWB transceivers and anchors. The UWB-based positioning system uses cost-effective components to achieve comparable positioning accuracy, making the system more economically viable for robotic platforms without sacrificing measurement precision

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

Solution Approach 2:

The patent substitutes the GPS satellite-based electromagnetic system with a ground-based UWB electromagnetic positioning system. This replacement eliminates dependence on satellite signals and provides better performance in indoor and obstructed environments while reducing cost and power consumption

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If RTK GPS systems are used for positioning, then positioning accuracy is improved, but power consumption increases

Engineering Contradiction:

Improvepositioning accuracyVSAvoidpower consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The UWB transceivers and anchors consume significantly less power than RTK GPS systems while maintaining positioning accuracy. This energy-efficient approach enables longer operational duration for mobile robotic platforms without requiring frequent recharging

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

3Device complexity

If classical UWB trilateration is used for positioning, then cost is reduced, but positioning accuracy deteriorates due to multipath propagation

Engineering Contradiction:

Improvesystem costVSAvoidpositioning accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent introduces a machine learning model as an intermediary between raw UWB measurements and final position estimates. This model processes distance measurements from multiple anchors, corrects for multipath propagation errors, and produces accurate position estimates, thereby resolving the accuracy issue while maintaining cost-effectiveness

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system uses quality indicators and uncertainty estimates as feedback to continuously improve positioning accuracy. The machine learning model learns from measurement quality patterns and adjusts its corrections accordingly, enabling the system to maintain high accuracy even in challenging environments with multipath propagation

Inventive Principle:

Principle #23Feedback

4Device complexity

If classical UWB trilateration is used for positioning, then cost is reduced, but positioning reliability deteriorates due to interference

Engineering Contradiction:

Improvesystem costVSAvoidpositioning reliability

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The machine learning model acts as a mediator that filters out interference and produces reliable position estimates. By learning from training data that includes various interference scenarios, the model becomes robust against different types of interference, thereby improving positioning reliability while keeping the system cost-effective

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentEP4657107A1Position prediction method and device

Publication Date: 2025.12.03 EINHELL GERMANY AG

Figure 1
Figure 2
Figure 3

AI summary

Described is a computer-implemented position prediction method (200) comprising a machine-learning model (300) configured to correct UWB-measured distance data (22) and/or pose data (23) of an autonomous robotic platform (10) indicating the distance and/or pose of the autonomous robotic platform (10) relative to at least a first UWB beacon (4a) and a second UWB beacon (4b) and for assigning at least one quality indicator to the UWB-measured distance data (22) and/or pose data (23) indicating the quality of a certain predicted distance (24), predicted position (25) and/or predicted pose (26), the method (200) comprising the steps of: - transforming (206) the UWB-measured distance data (22) into distance distributions (350) by assigning the at least one quality indicator of the UWB-measured distance data (22) to the UWB-measured distance data (22); and/or transforming (206) the UWB-measured pose data (23) into pose distributions (351) by assigning the at least one quality indicator of the UWB-measured pose data (23) to the UWB-measured pose data (23); and - determining (208) the most-likely position (25) in the local coordinate system by optimizing a cost function; and/or determining (208) the most-likely pose of the autonomous robotic platform (10) in the local coordinate system by optimizing a cost function.