Kalman Filter Correction via Adsorption Data for GPS Drift

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing GPS positioning algorithms, such as the least square method and Kalman filtering, suffer from unsatisfactory performance in dynamic multipath scenarios and slow GPS drift due to error accumulation, leading to inaccurate positioning.

Innovation Solution

A positioning data processing method that corrects the Kalman filtering algorithm using an adsorption data sequence to reduce noise and error accumulation, improving positioning accuracy by performing filtering and adsorption calculations on GPS data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If Kalman filtering algorithm is used for GPS positioning, then positioning effect in multipath scenarios is improved, but error accumulation occurs leading to slow GPS drift

Engineering Contradiction:
Improvepositioning accuracyVSAvoidpositioning stability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent introduces a feedback mechanism where the calculated route information is continuously fed back to correct the Kalman filtering algorithm. The system compares the filtered positioning data with the actual route and uses this feedback to adjust the algorithm parameters, preventing error accumulation while maintaining filtering effectiveness in multipath scenarios.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent dynamically changes the parameters of the Kalman filtering algorithm based on route information. By adjusting the state transition model and measurement model parameters according to the vehicle's actual route, the system adapts to different driving conditions and prevents error accumulation that would otherwise occur with fixed parameters.

Inventive Principle:
Principle #35Parameter changes

2Ease of manufacture

If least square method is used for positioning, then calculation simplicity is maintained, but positioning effect in dynamic multipath scenarios is extremely unsatisfactory

Engineering Contradiction:
Improvealgorithm simplicityVSAvoidpositioning accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent segments the positioning process into two parts: first using the simple least square method for initial positioning, then applying Kalman filtering with route-based correction for refined positioning. This segmentation allows the system to benefit from both the simplicity of least square and the accuracy of Kalman filtering in multipath scenarios.

Inventive Principle:
Principle #1Segmentation

3Productivity

If Kalman filtering algorithm uses previous optimal estimated value as calculation basis, then iterative calculation is achieved, but error accumulation cannot be avoided

Engineering Contradiction:
Improveiterative processing efficiencyVSAvoidpositioning accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent performs preliminary action by calculating the route information before using it to correct the Kalman filtering algorithm. This preliminary route calculation provides a reference framework that prevents error accumulation in the iterative process, as the algorithm is guided by the known route rather than solely relying on previous estimated values.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3748297B1Processing method and processing apparatus for positioning data, computing device and storage medium
Publication Date: 2025.06.25 TENCENT TECHNOLOGY (SHENZHEN) CO LTD
  • EP3748297B1 patent drawingFigure 1~2
  • EP3748297B1 patent drawingFigure 3~4
  • EP3748297B1 patent drawingFigure 5

AI summary

This application discloses a positioning data processing method and processing apparatus, a computing device, and a storage medium. The positioning data processing method includes: obtaining a first positioning data sequence generated by a moving target chronologically; performing filtering processing on the first positioning data sequence according to a preset filtering algorithm to obtain a filtered data sequence, and performing adsorption calculation on the filtered data sequence to obtain an adsorption data sequence, where the preset filtering algorithm is an algorithm obtained after a Kalman filtering algorithm is corrected according to the adsorption data sequence; outputting the filtered data sequence to obtain a second positioning data sequence of the moving target; and displaying a position corresponding to second positioning data in the second positioning data sequence. The adsorption data sequence is used to correct the Kalman filtering algorithm. In a process of chronological recursive calculation, a positioning offset of the moving target is eliminated to some extent, and in particular, an impact of error accumulation caused by the slow positioning offset is eliminated, so that the position corresponding to the second positioning data in the second positioning data sequence accurately reflects an actual position of the moving target, thereby improving accuracy of positioning and navigation, and improving user experience of positioning and navigation products such as an in-vehicle navigation product.