Dynamic data rectification using particle filters

Computing arrival cost parameters in moving horizon estimation using sampling based filters

“…A great deal of interest is generated by the utility of these simple, accurate and fast algorithms for the generally infinite dimensional nonlinear filter [2,10,[19][20][21][22]. The central idea is to represent the non-Gauss ian densities by a large number of samples or particles distributed accordingly and update the samples and weights conditioned on measurement information according to Bayes rule.…”

On reducing false alarms in multivariate statistical process control

“…This approach was originally proposed in the literature of time-series outlier detection to differentiate the presence of randomly induced outliers from the onset of systematic process change (Abraham and Chuang, 1993;Chen et al, 2008;Singhal and Seborg, 2000).…”

“…(2). Therefore a % 100β control limit can be established for i m , the limit denoting the maximum number of out-of-control data points allowed, % 100 β m , in a time window of size n (Abraham and Chuang, 1993;Chen et al, 2008;Singhal and Seborg, 2000): for m, and then selecting the maximum value of m that satisfies Eq. (3).…”

An integrated approach to active model adaptation and on-line dynamic optimisation of batch processes

“…When these assumptions do not hold, approximate methods need to be utilised, such as the extended Kalman filters, unscented Kalman filters [27,28] and more recently particle filters [6,29,10]. These filters are sequential in the sense that they infer the system states at the current time step by utilising the current measurement only.…”