Distributed Kalman Filter Architecture for GNSS Carrier Range Ambiguity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Kalman filter architectures for carrier range ambiguity estimation in GNSS receivers are computationally intensive, making them unsuitable for resource-constrained environments, as they require a large number of states and increased computational costs due to the need for separate Kalman filter implementation for each observed carrier range.
Innovation Solution
A distributed Kalman filter architecture is introduced, comprising a first Kalman filter to estimate carrier range ambiguity values and a second Kalman filter to determine navigation data, where the second Kalman filter receives ambiguity data from the first, reducing the computational burden by estimating ambiguities separately and allowing implementation on less powerful processors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a separate Kalman filter state is implemented for each observed carrier range to estimate ambiguities, then carrier range ambiguity resolution accuracy is improved, but device complexity and computational cost increase non-linearly
Solution Approach 1:
The patent segments the Kalman filter implementation into two separate filters: a first Kalman filter dedicated to estimating carrier range ambiguities and a second Kalman filter for navigation state estimation. This segmentation divides the originally monolithic high-dimensional filter into two lower-dimensional filters, reducing computational complexity while maintaining the ability to resolve ambiguities for multiple satellites. The first filter handles ambiguity states for C*M satellites, and the second filter handles navigation states independently.
2Measurement precision
If the number of Kalman filter states is increased to include carrier range ambiguity estimates for multiple satellites, then positioning precision is improved, but processing time and computational power requirements increase
Solution Approach 1:
By segmenting the filter into two separate Kalman filters, the patent reduces the computational burden of matrix operations. The first filter computes ambiguity estimates with dimension 8+2*C*M, and the second filter computes navigation states with dimension 8. This segmentation allows the system to maintain high positioning accuracy through comprehensive ambiguity resolution while reducing the non-linear computational growth that would result from a single monolithic filter handling all states simultaneously.
Solution Approach 2:
The first Kalman filter performs preliminary estimation of carrier range ambiguities before the second filter uses these estimates for navigation state determination. This preliminary action allows the second filter to operate with reduced complexity, as the ambiguity parameters are pre-computed and can be treated as known or constrained values during navigation state estimation, thereby reducing overall processing time.
3Reliability
If a monolithic Kalman filter is used to estimate both navigation states and carrier range ambiguities, then measurement redundancy is maximized, but ease of operation and implementation on resource-constrained devices is worsened
Solution Approach 1:
The patent divides the monolithic filter into two specialized filters, each optimized for its specific function. The first filter focuses exclusively on ambiguity resolution using carrier phase measurements, while the second filter focuses on navigation state estimation using pseudorange and carrier range measurements. This segmentation maintains measurement redundancy by allowing both filters to process the same GNSS measurements independently, while making implementation feasible on resource-constrained receivers by reducing the computational burden of each individual filter.
Solution Approach 2:
The first Kalman filter acts as an intermediary that processes carrier phase measurements to generate ambiguity estimates, which then serve as inputs or constraints for the second Kalman filter. This intermediary structure allows the system to maintain the benefits of using both pseudorange and carrier range measurements for redundancy while simplifying the operational complexity by breaking down the processing into manageable stages that can be implemented on resource-constrained GNSS receivers.
Data Source
AI summary
Methods and apparatus for determining navigation data (124) using carrier range measurements obtained by a GNSS receiver for a plurality of GNSS satellites, the apparatus comprising: a first Kalman filter (100) configured to determine an a posteriori estimate of a first state vector based at least in part on a first set of GNSS measurements (104) obtained by the, and/or a further, GNSS receiver and an a priori estimate of the first state vector, the first state vector comprising carrier range ambiguity values relating to the plurality of GNSS satellites and a position of the, and/or the further, GNSS receiver; and a second Kalman filter (102) configured to determine an a posteriori estimate of a second state vector, which comprises the navigation data (124), based at least in part on a second set of GNSS measurements (106) obtained by the, and/or a further, GNSS receiver, an a priori estimate of the second state vector, and carrier range ambiguity data (126) based on the carrier range ambiguity values determined in the a posteriori estimate of the first state vector.