I have a signal which will be either 1,0 or -1 for a discrete time interval (say 10s) then change depending on other parameters. I want a counter to keep a running count of an Initial Value modified by the input signal.
example:
t0->t1, input=1, output=initialvalue+1
t1->t2, input=0, output=initialvalue+1
t2->t3, input=1, output=(initialvalue+1)+1
t3->t4, input=-1, output=((initialvalue+1)+1)-1
etc. Ideally I would want this to work for input being any integer but I'll take what I can get...
There's a counter block but it only counts up or down, not both. I tried using edge detection but I'm not sure how to implement the discrete counter on which it operates. Any help highly appreciated.
Related
I would like to predict the switching behavior of time-dependent signals. Currently the signal has 3 states (1, 2, 3), but it could be that this will change in the future. For the moment, however, it is absolutely okay to assume three states.
I can make the following assumptions about these states (see picture):
the signals repeat periodically, possibly with variations concerning the time of day.
the duration of state 2 is always constant and relatively short for all signals.
the duration of states 1 and 3 are also constant, but vary for the different signals.
the switching sequence is always the same: 1 --> 2 --> 3 --> 2 --> 1 --> [...]
there is a constant but unknown time reference between the different signals.
There is no constant time reference between my observations for the different signals. They are simply measured one after the other, but always at different times.
I am able to rebuild my model periodically after i obtained more samples.
I have the following problems:
I can only observe one signal at a time.
I can only observe the signals at different times.
I cannot trigger my measurement with the state transition. That means, when I measure, I am always "in the middle" of a state. Therefore I don't know when this state has started and also not exactly when this state will end.
I cannot observe a certain signal for a long duration. So, i am not able to observe a complete period.
My samples (observations) are widespread in time.
I would like to get a prediction either for the state change or the current state for the current time. It is likely to happen that i will never have measured my signals for that requested time.
So far I have tested the TimeSeriesPredictor from the ML.NET Toolbox, as it seemed suitable to me. However, in my opinion, this algorithm requires that you always pass only the data of one signal. This means that assumption 5 is not included in the prediction, which is probably suboptimal. Also, in this case I had problems with the prediction not changing, which should actually happen time-dependently when I query multiple predictions. This behavior led me to believe that only the order of the values entered the model, but not the associated timestamp. If I have understood everything correctly, then exactly this timestamp is my most important "feature"...
So far, i did not do any tests on Regression-based approaches, e.g. FastTree, since my data is not linear, but keeps changing states. Maybe this assumption is not valid and regression-based methods could also be suitable?
I also don't know if a multiclassifier is required, because I had understood that the TimeSeriesPredictor would also be suitable for this, since it works with the single data type. Whether the prediction is 1.3 or exactly 1.0 would be fine for me.
To sum it up:
I am looking for a algorithm which is able to recognize the switching patterns based on lose and widespread samples. It would be okay to define boundaries, e.g. state duration 3 of signal 1 will never last longer than 30s or state duration 1 of signal 3 will never last longer 60s.
Then, after the algorithm has obtained an approximate model of the switching behaviour, i would like to request a prediction of a certain signal state for a certain time.
Which methods can I use to get the best prediction, preferably using the ML.NET toolbox or based on matlab?
Not sure if this is quite what you're looking for, but if detecting spikes and changes using signals is what you're looking for, check out the anomaly detection algorithms in ML.NET. Here are two tutorials that show how to use them.
Detect anomalies in product sales
Spike detection
Change point detection
Detect anomalies in time series
Detect anomaly period
Detect anomaly
One way to approach this would be to first determine the periodicity of each of the signals independently. This could be done by looking at the frequency distribution of time differences between measurements of state 2 only and separately for each signal.
This will give a multinomial distribution. The shortest time difference will be the duration of the switching event (after discarding time differences less than the max duration of state 2). The second shortest peak will be the duration between the end of one switching event and the start of the next.
When you have the 3 calculations of periodicity you can simply calculate the difference between each of them. Given you have the timestamps of the measurements of state 2 for each signal you should be able to calculate the time of switching for all other signals.
I need a subsystem that needs to output 1 at interval or 30 seconds or slightly over 30 seconds.
Written in matlab code it should work like that
function y = fcn(time,uplinkTimeInterval)
%#codegen
persistent lastTriggerTime
if isempty(lastTriggerTime)
lastTriggerTime = 0;
end
if time>=lastTriggerTime || time == 0
y = 1;
lastTriggerTime = time + uplinkTimeInterval;
else
y = 0;
end
end
where ulplinkTimeInterval is 30 seconds. Of course I tried to use the matlab function block with this code but for some reason it does not work (in debug mode I can see that y takes value 1 as it should but it simply does not ouput the value outside the block), therefore I wanted to use blocks but I do not know how to do it.
Thanks very much for your help
You can make this logic relatively easily with code or blocks. As you requested a solution using blocks, here it is!
Use the clock block to keep track of time, and some constant block to determine the interval (in seconds) at which to give 1 instead of 0.
Use the memory block to delay the clock signal by 1 timestep, so we can compare consecutive steps' values.
Divide the times by the interval, and round down, to give how many intervals have passed.
Finally, compare consecutive "number of intervals passed" using a relational operator. If more intervals have passed on the upper line, then you have just stepped over the interval threshold.
Note: this will return a 0 for every timestep where you have not crossed a new interval, and a 1 at each individual timestep where you have. The accuracy of the output will depend on the step size of your model.
Edit: It may be clearer / easier to just add the memory block after the floor block, so you are only doing the division / rounding once. It would still allow you to do a comparison to the previous time step. That would look like:
Easiest way to do this is with a just a single Pulse Generator block, set to have a "high" of 1 every 30 seconds. That is shown as part of the image below. The signal will be high for whatever the percentage of the period is specified in the block dialog.
If for some reason you really need to use a subsystem then use a Triggered and Enabled Subsystem (See top right of image). Feed the same pulse signal into both the trigger and the enable port, and set the outport inside the subsystem to have Output when disabled to reset, and to have an Initial Output of 0 (See the lower right of the image).
The model below shows how to do this. In this instance the pulse has been set to have a period of 30s with the rising edge happening every 1% of that period (See the top left of the image).
The output signal will be high for one time step every time the input rises (assuming the trigger is set to rising edge.)
Assume that you have a Simulink simulation where a certain signal is first positive and after some time t in a given interval, it becomes negative. Your goal is to find the zero-crossing.
A first approach would be to plot the signal over the given interval, save it and calculate the zero-crossing.
When repeating this simulation numerous times for varying parameters, it would be beneficial to be able to stop the simulation after the signal has become negative. Then there is already enough information for the calculation of the zero-crossing in Matlab. How could you do that?
Yes, use the Stop Simulation block with the appropriate logical input to the block:
You can use an if / else block to control the flow in the Simulink model. In the if / else block, you can choose the condition if u > 0, continue as normal if it's true, and use the else-option to bypass the rest of the flow you would otherwise run. For instance jump directly to the scope.
Another ways:
You can use Hit Crossing Block in Simulink to find time at the moment of hitting zero.
Another way - use any Trigger or Resettable system. It detects the zero crossing too. For example: this question at SO.
Of course you can also use User Defined function to detect zero crossing by your hand and do whatever you want with the signal at the same time.
About making a lot of simulations and then stops:
you can use Check Upper Static Bound for automatically stops simulation at the moment when zero will be crossed in nth time. For example:
I set upper bound = 10 for this block and this stops at 10th crossing.
There are a lot of ways to save function values in this points or just array of times but this is an another question :)
In Simulink how can I delay the transport block to the exact time of when a certain threshold has been reached?
Basically I have an S-function which switches to an alternate input source, a sine wave, after a certain value has been reached. As soon as the switch occurs I want to start the sine wave. For that I need to delay the wave generation until the switch occurs.
How can I send that time to the Variable transport delay block?
Maybe you can try this way:
you need Variable Integer Delay for example. It changes delay to what you need. Calculate needed value of delay you can in User Defined Function, route to it your threshold value and current time and all other values you need.
Hope i understood your question correctly!
I'm using Matlab 7 and have a problem in creating a monoflop which shall raise for a specific time to "1" and after that time fall to "0". How can I do this with Matlab/Simulink 7?
I don't have any other version, so I can't use the "Monostable" Block from newer versions.
Any ideas?
greets, poeschlorn
There are a couple of ways to do this, depending on whether or not you want the pulse (i.e. "monoflop") to occur at a predetermined time or in response to another signal (like a rising edge)...
Creating a pulse at a predetermined time:
If you want to create a single pulse that steps from 0 to 1 at time tOnset, then steps back to 0 after a time tDur has elapsed, you can do this using a Step block, a Transport Delay block, and a Sum block. Here is what the layout would look like:
You would set the Step time of the Step block to tOnset, the Time delay of the Transport Delay block to tDur, and then subtract the delayed signal from the original signal.
Creating a pulse in response to a rising edge:
This is will be a bit more complicated. It will require two Detect Increase blocks, a Relay block, a Transport Delay block, a Gain block, and a Sum block. Here's what the layout would look like:
Assuming your input signal is either a 1 or a 0, the first Detect Increase block will output a 1 when the input steps from 0 to 1. By setting the Switch on point to 0.5 and the Switch off point to -0.5 for the Relay block, this will create hysteresis in the Relay such that the output will persist in the "on" state (i.e. an output of 1) after the brief pulse that occurs when the rising edge is detected.
To get the Relay block to switch back into the "off" state (i.e. an output of 0) after a specified time tDur, you would set the Time delay of the Transport Delay block to tDur. The Detect Increase block in the feedback loop will output a 1 when the delayed signal steps from 0 to 1, and this output will then be scaled by setting the Gain of the Gain block to 2.
When subtracted from the input signal, this gain will ensure that the output from the Sum block can be pulled below -0.5 no matter what the positive input is (0 or 1), thus ensuring that the Switch off point of the Relay block is reached and its output is turned off when the delayed signal has a rising edge (i.e. after tDur has elapsed). One result of this is that any additional rising edges occurring in the model input after the first rising edge and during the time tDur will be ignored. Once the output from the model returns to 0, the next rising edge on the model input will trigger another pulse.