Update the Kalman Filter in Simulink model

I want to creat a Filter in Simulink. For example:

kf = trackingKF('MotionModel', '2D Constant Acceleration', 'State', zeros(6, 1));

Then in simulink model, I want to update the filter and predict the states after receive feedback during the simulation. The function is defined as:

function predictedState = fcn(x,y)
    
    % Receive feedback about the present position
    presentPosition = [x;y]; 
    
    % Correct step: Update the filter with observed position
    correct(filter, presentPosition);
    
    % Predict future positions after receiving feedback
    predictedState = predict(filter);
 
end

My problem is: How to define this Filter in Simulink that can be updated by using 'correct' command, and how can I build a structure that can update this filter using my above function during my simulation. Thank you for your helps !

I try to define the Kalman Filter in my Predict function but it seems not good because the filter need to be updated to become more precise. Can the callback function in Model Explorer give a help ?

Answer 1

[MULI]

Hi Pham,

I understand that you need to update the filter states based on feedback received during simulation. To integrate and continuously update a Kalman Filter in Simulink, you can use a "MATLAB Function" block. Within this block, utilize a "persistent" variable to maintain the filter's state across function calls, as given below:

function predictedState = fcn(x, y) 

       % Persistent variable to hold the filter 

       persistent filter  

       % Initialize the filter if it's the first execution 

       if isempty(filter) 

           filter = trackingKF('MotionModel', '2D Constant Acceleration', 'State', zeros(6, 1)); 

       end 

       % Correct the filter's state based on new measurements 

       presentPosition = [x; y]; 

       correct(filter, presentPosition); 

       % Predict the next state 

       predictedState = predict(filter); 

   end 

You may refer to this documentation link for more information on “persistent” variable.

https://www.mathworks.com/help/releases/R2022a/matlab/ref/persistent.html