Stepped Kalman Filter for Low-Power Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation systems combining GPS and INS face computational challenges due to the high processing demands of extended Kalman filters, limiting the update rate and making them unsuitable for low-performance navigation computers, especially in cost and power-constrained applications.
Innovation Solution
The implementation of a stepped Kalman filter, which reformulates the EKF equations into sub-steps based on time separation assumptions, allowing for lower update rates of the filter while maintaining accuracy and stability, enabling implementation on small microcontrollers with high update rates of state estimates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an extended Kalman filter is used to integrate GPS and INS data, then navigation accuracy is improved, but computational complexity and processing power requirements increase
Solution Approach 1:
The patent segments the Kalman filter update process into multiple sub-steps, where the filter is applied at a lower rate (e.g., once per GPS update) while the inertial navigation system runs at a higher rate. This segmentation allows the computationally intensive EKF to be executed less frequently while maintaining overall navigation accuracy through the high-rate INS predictions filling the gaps between updates.
2Speed
If the Kalman filter update rate is increased, then navigation responsiveness is improved, but processing power consumption increases
Solution Approach 1:
The patent implements dynamic update rates where the Kalman filter operates at a lower rate than the inertial navigation system. The INS continuously updates at a high rate to maintain responsiveness, while the GPS-aided EKF updates at a lower rate to reduce computational load. This dynamic differentiation of update rates allows the system to achieve both responsiveness and energy efficiency.
3Power
If a high-performance computer is used for navigation calculations, then processing capability is improved, but cost and weight increase
Solution Approach 1:
The patent changes the operational parameters of the Kalman filter by reducing its update rate and adjusting the covariance matrices to account for the lower update frequency. This parameter modification allows the same filter algorithm to run efficiently on low-power microcontrollers rather than requiring high-performance computers, thereby reducing weight and cost while maintaining navigation capability.
Data Source
AI summary
Embodiments include systems and methods of navigation. In on embodiment, a plurality of position and motion states of a vehicle are estimated. The states may be estimated based on information received from a satellite receiver and an inertial measurement sensor. Estimating the states comprises performing one or more of a plurality of update steps at the rate that information is received from the satellite receiver. The states are estimated at a rate greater than the rate at which the update steps are performed. In one embodiment, the states are estimated using a stepped extended Kalman filter.