Filtering noise from signal

aand74

Member
Join Date
Dec 2005
Location
Deinze
Posts
131
I know there are a lot of less or more complicated filters,
but if you only choose between moving average filter, and first order filter,
which filter do you use when, what are the advantages, disadvantages?
When should you switch from first order filter to second or higher order filter?
 
I know there are a lot of less or more complicated filters,
but if you only choose between moving average filter, and first order filter,
which filter do you use when, what are the advantages, disadvantages?
When should you switch from first order filter to second or higher order filter?

I guess there's many more types of filtering algorithm that I've never even heard of, but I also guess that the better the filtering, the more complex it will be, and hence more computational power will be needed. It most likely is a trade-off situation between effectiveness and cost.

I'm sure Peter will have something to say on this when he sees it.
 
First Order:
Adv - Easy to implement, easy to model, minimal program memory req'd
Disadv - "Flier" data takes a long time to settle out (i.e., infinite impulse response), may require long lag time to achieve desired filter performance, potentially affecting feedback control performance.

Moving Average:
Adv - Can attenuated a specific frequency if "noise" (i.e., undesirable signal component) is known occur at that frequency, affect of "flier" data is gone after the moving average time base.
Disadv - Memory requirements and programming complexity, potential start-up time.

Higher Order (e.g., Butterworth) - can get better performance - noise reduction without phase shift - with modest increase in programming complexity.
 
I've got an Arduino project in the late stages of development for a Radio Control FPV receiving station that tracks the flying model.

It uses the RSSI (Received Signal Strength Indicator) off 3 receivers to determine the direction to point the antennae.

RSSI values can be affected by so many things, and a simple filtering algorithm wouldn't cut it.

I'm experimenting with Kalman filtering, and getting great results so far. Field trials (when the weather gets a bit warmer) to follow.

Kalman filtering (the cheap and cheerful explanation) looks at each "sample", and basically assigns a "confidence vote" to it, by comparing it to previous samples. It effectively only uses a proportion of the change to enter the filter, depending on how confident the algorithm thought the sample was.

It appears to be working well for filtering out the sort of things that can affect radio signals, noise, multi-path reflections, blinding, etc.
 

Similar Topics

Hi all how do i configure this AI card to filter noise coming from my input device? Thanks
Replies
2
Views
2,264
Hi All, I'm currently designing a faceplate for an analog input object, and I'm trying to create an alarm tab with an Alarm and Event Summary...
Replies
1
Views
405
Hello, I am looking to use the same header on every page (global object) that has an alarm summary button on it that will go to a display with an...
Replies
4
Views
1,774
Hi, I have just started with WinCC unified (v17) and there are alot of things I used on Comfort but now are not available. Currently I am finding...
Replies
3
Views
2,665
i have FactoryTalkView Studio ME ver 12; i have a project with more than 40 displays. 1) How to print the entire project into a pdf. 2) I need to...
Replies
3
Views
756
Back
Top Bottom