Designing Butterworth high pass filter for accelerometer

i have a data set of 5 people. this data is collected from the user by performing different activities. the collected by the people mounted the accelerometer device on different body location ,i-e; waist, arm and pocket. i need to estimate the kinetic power traces from those acceleration traces. the data set i have is a 3 axial set but i need the overall magnitude of the acceleration and that one i got it. now the next step is the filter out the gravity component from the acceleration data.now i need to use 3rd order Butter-worth high pass filter with a 0.1 Hz cutoff frequency.

could anybody tell me how can i design and implement filter?? thank u

looking forward to a positive reply Thank you

Answer 1

[Star Strider]

I would use a bandpass filter to filter out both the d-c component (here, gravity), and high-frequency noise in one step. Use the fft of your data to determine what is valid signal data and what is noise. I usually begin with the stopband frequencies for a bandpass filter being 0.5 and 1.25 times the lower and upper passband frequencies respectively. You will likely have to experiment with the design to get the results you want.

My filter design procedure is here: How to design a lowpass filter for ocean wave data in Matlab?