Assignment 2, Digital Signal processing: FIR filtersThe task of this assignment is to filter an ECG with FIR filters. Before you record from yourselfplease read the info-sheet and sign the consent form.1 ECG filteringForm. groups of 2 students. Record at least 2 different ECGs so that every member has his or herpersonal ECG. There are different amplifiers with different gains and converter boards available. Whenyou record the ECG make sure that you take a note of the gain, bits, etc. Make sure the “raw” buttonis pressed while recording. If you record with a one channel amplifier record Einthoven II (google itwhat it means). Upload the files onto the ECG database on moodle or e-mail it to yourself. Give thefile a name which anonymises the ECG. Use only numbers, characters and the underscore for yourfilename.Every student has to work on a different ECG. If your initial recording has been too noisy just recordanother one. You can minimise artefacts by lying down and/or placing the electrodes on the shoulders/ hips instead of wrists/ankles. In Python load the file with the command numpy.loadtxt(’myecg.dat’.You’ll get a numpy matrix with time-stamps in ms in the first column and 3 columns of ecg with rawunsigned integer numbers. Choose the channel with the highest amplitude. Also note that the ECGdevices will be around all Oct / Nov so there is plenty of time to record your ECG. You can use oneof the example ECGs from moodle to kick start your work.No high level Python functions from scipy except of fft and the window functions are allowed inthis section.1. Display the original ECG in mV (!!!) against msec. Show the whole recording and one heartbeatto identify the so called PQRST waves (google it). Remember to take into account the gain ofthe amplifier. The A/D converter has either 12bit or 24bit resolution and has an input rangefrom −4.096V ...+4.096V or −1.325V ...1.325 respectively. Do not use the pre-defined pythonfunctions to convert it. Rather develop your own formula to convert the data. Plot also thefrequency spectrum of the ECG (abs(F(ω))) (omitting the mirror). Discuss what is signal andwhat is noise. In particular, identify the unwanted 50Hz contamination and the DC drift ofthe ECG. Label the ECG properly with the help of a vector based drawing program such asInkscape. [10%]2. Create a Python FIR filter class which implements an FIR filter which has a method of the formvalue FIR(value) where both the value argument and return value are scalars and not vectorsso that it could be used in a realtime system. The constructor of the class takes the coefficientsas its input:class FIR_filter:def __init__(self,_coefficients):# your code here1def filter(self,v):# your code herereturn result[20%]3. Calculate the coefficients of an FIR filter analytially (by using the sinc formula from the lecture)for a 50Hz notch filter and filter the ECG with it. [10%].4. Calculate the coefficients of an FIR filter to remove the baseline shift and 50Hz of the ECGnumerically with the help of the inverse Fast Fourier Transform. Filter the ECG and make surethat the ECG is not distorted. [10%].2 ECG heartrate 调试C/C++程序、C/C++FIR filters 程序代做留学生、代做C/C++编程作业、C/C++课程设计代做留学detectionThe task is to detect the momentary heart rate r(t) over a longer period of time. For example, afterexercise you should see a slow decay of the heart rate to the baseline of perhaps 60 beats per minute.It is not the average heart rate but the frequency derived from the times between adjacent heartbeats.Record a separate long ECG over a couple of minutes from at least two different people. Ask theperson to breathe in quite deeply which will change the heart rate or anything which changes the heartrate for example sending the person down to level 1 and back.High level filter scipy filter commands are allowed to speed up the filtering operation.1. Design a matched filter which detects the individual heartbeats in the ECG. Think of the opti-mum length of the filter and give a reason. Optimise the detection by pre-filtering both the ECGand the template. Remember the detection works best when both the signal and the templatesare completely DC free. [20%]2. Implement a small program which uses the output of the matched filter to calculate the momen-tary heart rate r(t) over time (not the average!) by measuring the intervals between the detectedhearbeats over the whole period of the ECG. Detect the heartbeats by employing a threshold(adaptive, if required). Square the output of the detector to improve the signal to noise ratio ortry other methods. Generally use any heuristics to weed our wrong detection, for example thata heartbeat is usually below 200bpm and above 30bpm. [30%]Every report must be based on different ECG recordings. Please keep it short but it should makeclear what you have done and why you have done it. Include the Python code as well as plots of theECGs (timedomain) and their frequency representation (with proper lables). If necessary enhance theplots with InkScape and remember to export plots in vector based image formats, for example EPS,SVG, PDF or EMF and not pixel based formats. Also, show zoomed in ECG traces of one heartbeat sothat it is possible to identify the different parts of a single heartbeat (see Fig. 1) and that it’s possibleto check if it’s still intact. Hand in your report at the teaching office (Room 620 James Watt BuildingSouth). Deadline is 27th November 3pm.3 Bonus work (instead of assignment 3 = double the marks)3.1 Option 1: FIR module in Python/C++/SWIGWrite a class which does both the filtering and the FIR filter design for lowpass, highpass, bandpassand bandstop. Optimise the actual FIR filter routine by implemeting it in C or C++ and create anFIR package/module. Publish this class on github including a demo and howto. If successful this willgive you guaranteed 100%. It can be done under any operating system but its easiest under Linux andnearly impossible under Windows (but a nice chanllenge!).3.2 Option 2: Realtime demoWrite a realtime demo which shows that your FIR filter is able filter the ECG in realime. The USB-DUX boards use a library called comedi which has a Python wrapper. Then on the other handmatplotlib can animate realtime data. This needs to be done under Linux as the USB-DUX boardsare made for Linux. Alternatively that can be also done with the Attys (www.attys.tech) which hasalso python libraries but needs to be adjusted.转自：http://ass.3daixie.com/2018060769869946.html
