RUKF
The RUKF class implements the randomized unscented Kalman filter and its iterative version.
Configuration
- Configure the number of samples used for state prediction and measurement update
Set the linear factors to determine the number of samples used for state prediction and measurement update, i.e., the number of samples, with the setNumSamplesFactors() method.
- Enable the iterative measurement update
Enable the iterative measurement update with the setMaxNumIterations() method. Additionally, you can check for convergence, i.e., if no further iterations are required, with the setConvergenceCheck() method.
- Enable measurement gating
Enable measurement gating with the setMeasGatingThreshold() method.
- Use a semi-analytic measurement update
Enable a semi-analytic measurement update if a measurement model does not require all state variables with the setStateDecompDim() method.
- Enable post-processing of the predicted state estimate
Set a post-processing method for the state prediction with the setPredictionPostProcessing() method.
- Enable post-processing of the updated state estimate
Set a post-processing method for the measurement update with the setUpdatePostProcessing() method.
Supported Models
Literature
Jindřich Dunı́k, Ondřej Straka, and Miroslav Šimandl, “The Development of a Randomised Unscented Kalman Filter,” in Proceedings of the 18th IFAC World Congress (IFAC 2011), Milano, Italy, Aug. 2011, pp. 8–13.
Ángel F. Garcı́a-Fernández, Lennart Svensson, Mark Morelande, and Simo Särkkä, “Posterior Linearisation Filter: Principles and Implementation Using Sigma Points,” IEEE Transactions on Signal Processing, vol. 63, no. 20, pp. 5561–5573, Oct. 2015.