Parameters
Name
The name of the ModFrame component. This name will be visible below the ModFrame component in the ModFrame Client configuration panel.
 | The name of a ModFrame component must be unique within the container it resides in. |
Active
This parameter is not used.
Report Event Level
The event level severity sent to the event log. All events with equal or higher severity will be reported. The levels are:
- Info - All events are reported
- Warning - Warnings and Alarms are reported
- Alarm - Alarms are reported
- None - No events are reported
Model type
The Model type parameter decides what kind of model to use. Possible choices are:
- Linear. Linear state space model. The standard Kalman filter algorithm will be used in this case. The model is defined by the BMExKalmanFilter property matrices or by model matrices connected to BMExKalmanFilter ports.
- Nonlinear. Nonlinear state space model. The extended Kalman filter algorithm will be used in this case. Nonlinear models must be defined in separate Java classes. These classes must implement the MFSSModelInterface.
Model path
The path for nonlinear model classes
Model
The Model parameter defines which nonlinear model to use. All the models in the path defined by the Model path parameter will be listed here. This property has no function when the Model type parameter is set to Linear.
Reload model
This property forces the nonlinear model parameters to be reloaded to initial values from the model class defined by the Model parameter. This property has no function when the Model type parameter is set to Linear.
Parameterization
With this property the type of parameterization is set. Possible choices are:
- Internal. When this choice is activated, all model parameters are defined in the BMExKalmanFilter property pages.
- External. When this choice is activated, parameterization ports will be visible at the left side of the module. Model parameters will be read from these ports. This choice is advantageous when parameters change. When linear models are used there will be one input port for each model matrix. When nonlinear models are used, there will be one input port for the model parameter vector θ.
External lin. point
When this property is true, nonlinear models will be linearized around a linearization point defined by input ports.
Export linear model
When this boolean parameter is activated, the linear model will be exported to BMExKalmanFilter output ports. In the case of nonlinear models, the linear version linearized around the current operating point is exported.
Manipulative matrices for Vk and K
Matrices can be defined on ports which are pre- multiplied with the process covariance matrix Vk and the Kalman gain matrix K. Useful for temporarily masking out parts of the filter, i.e. due to sensor failure.
Export interval
This integer property defines how often to export linearized models. The default value is 1. This means that the model will be exported every time step. If the property is set to n, the model is exported every n'th time step.
Constant gain
This is a boolean parameter defining the type of gain to use. When true, constant Kalman gain will be used. When false, the Kalman gain will be computed at each time step.
Only states in readState
ModFrame has a feature of persisting the current state of modules. If this property is true, only the state vectors will be persisted (fastest). If false, filtering matrices will be persisted as well.
Check observability
This property is used to check the observability of the model. When true, an observability report is written to the ModFrame Client Log view. The property should only be true during configuration, since the computations are time consuming and the Log view will be overloaded with information.
Update initval on save
When this property is set to true, the initial values of the Kalman filter is updated to the current state of the filter when the application is saved. If false, the initial values will be read from the model class file the next time the application is loaded.