Fuzzy control of portable ECG monitor design

　　

At present, the acquisition of ECG, analysis and diagnosis of ECG monitoring system has been widely used for heart diseases prevention, diagnosis and played a significant role.

But this type of ECG monitor can only be in the patient or the patient to static special circumstances to use, the heart that is too high, especially heart disease suspected patients and early heart disease, affecting their normal working life; and the other portable ECG monitor, its 24-hour can monitor, but its storage needs lots of space for ECG playback requires a significant amount of time, in the light of the above two questions, this article design based on fuzzy control of portable ECG monitor. He overcome these two problems at the same time also a breakthrough in the past, online diagnosis disease of single-value processing, can more accurately determine the abnormal ECG and achieve a timely issued an alert. 1 system President design 1.1 system design goals according to the ECG signal characteristics, biological signal processing system and the development of modern ECG monitoring technology, this system using high-speed SOC series MCU C8051F020 as cardiac monitor's main chip implements the following features: (1) measurement with non-invasive, security, accuracy, repeatability strong, etc.; (2), measurement of simple and convenient operation, without prejudice to the question monitoring patients a normal life; (3) real time analysis ECG and judge the signal is, exception, preliminary monitoring patients diagnosed disorders of the heart; (4) 16 MB of FLASH memory to store the users 24 hours of abnormal signals; and (5) system error alarm function; (6) can be stored in abnormal ECG via the USB interface to the PC for further diagnostic ECG. 1.2 system hardware composing the system hardware diagram shown in Figure 1. First of all from the electrode circuit acquisition of the measured object Dim ECG, then to zoom in on the signal, after a filter directly to the main control chip, the other into the lead off detection circuit, then the connection with main control chip C8051F020. Signal after entering master chip C8051F020, appropriate treatment, alarm, storage, transmission, etc., judgment. 1.3 system flowchart system flowchart shown in Figure 2. 2 fuzzy analysis ECG automatically control section's main goal is to make the diagnosis of cardiac arrhythmia. The clinic for cardiac arrhythmia automatic diagnosis is combined with rhythm analysis and waveform shape analysis of measured ECG waveform recognition classification and diagnosis according to predetermined standards or criteria to make the appropriate clinical diagnosis. But the design is based on the extraction of characteristic value combination of medical knowledge and medical expert system completes the judgment. First of all, according to the following criteria for the treatment of patients with initial detection diagnosis: tachycardia r-r interval < 0.5 s (120 times/min); bradycardia r-r interval > 1.5 s (40th/min); arrest and ventricular fibrillation in a long time no QRS, generally this time > 1.6 s; leakage Bo a r-r interval is about the average r-r interval before 2 times after and no one beats is checked out as leakage Bo, r-r interval greater than the average of 2 times but less than 1.5 s, as room sinus stop checkout; ventricular premature ventricular contractions in pairs (two consecutive ventricular extrasystole): room for 2 with LV (normal and ventricular premature alternating twice over); Chamber of triple law (normal, normal and ventricular premature alternating two consecutive times above), are classified as premature ventricular contractions. Testing standard complex, the need for further analysis using fuzzy logic; judgment in T R (R on T) this is in the Ventricular Repolarization era (T-wave) appears in PVC, because T wave cannot detect, so only by rhythm analysis; insert premature is no compensation of the beats, the beats of roughly equal beats R-R interval before average R-R interval; atrial premature beats (APB) an adaptation of a premature ventricular contraction. More on heart disease is the criterion for judging the second value, this method of detecting simple easy-to-implement but on diseases of the judgment is not accurate. In fact there are a large number of medical diagnosis using fuzzy judging language and behavior, according to the prevalence in patients with multiple parameters multiple value judgment. These criteria are thus aspects of medical experts of prior knowledge, use these criteria to form multiple fuzzy rule, the medical expert clinical diagnosis disease with machines. Here's to fuzzy logic testing room of early Bo, for example, to monitor the patient's condition to conduct preliminary diagnosis, and the difference is, abnormal ECG, abnormal ECG recording only. In single-chip implementation of fuzzy control generally use 3 ways: direct transfer method, the strength of the look-up table and the formula calculation. Considering the direct consulting by offline calculation, get a fuzzy control table control table stored in computer memory, the application in the control, the speed while soon but if variables more (this system with up to five) will result in the dimensionality of fuzzy control, with high storage table lookup is not convenient. With the pretended formula also small for this system parameter calculation. While the transfer mode is based on the strength of fuzzy control of a minimal approach to inference. Each input parameter is mapped to more than one membership grade, each input activates the rules that may correspond to different results. Used by various small principle calculation corresponds to the output of the combined strength of the rule, and then press the largest membership principles that corresponds to the credibility of the conclusions. Thus, for each input can be obtained with the conclusions of the corresponding output intensity. Known as the output of membership. Get maximum output intensity of the membership as output. Experimental results show that this method most appropriate to this system. Here's a case study with beats that extract r-wave width (RW), RR interval (RR), the R wave area (RA) T wave area (TA), the T wave value (TH) these five eigenvalue diagnosis using fuzzy to be custodians of ventricular contractions occur. Some fuzzy rules table as shown in table 1. 2.1 membership function stored as single-chip memory capacity is limited, ifThe system input and output on the domain of all membership functions of continuous curve for storage, it is not possible, so this system for triangle membership function takes three-point method and storage 3 vertices of a triangle; and a half for the two sides of the waist and trapezoidal storage top 3 points. Membership function stored in ROM, Figure 3 to R wave's width description for an example. 2.2 input characteristics of fuzz extraction is the exact values, the parameter will they be compared with the membership function, find the appropriate amount of fuzzy input of membership in the range 0 to 1, the microcontroller can be said to 00 ~ FFH. For the purposes of this system, each accurate input value corresponds to a maximum of only two fuzzy input quantity is greater than zero, the remaining amount is fuzzy input to zero. For example: suppose RW = 1.1, from table 3 that he fell in (m) and large (L) two interval, in (m) and large (L) of membership: μ M (1.1) = (1.3-1.1)/(1.3-1.0) × FFH = 5 μ H A L (1.1) = (1.1-1.0)/(1.3-1.0) × FFH = 55H in RAM to open up an area, enter the depositing the fuzzy. 2.3 fuzzy rules deposited fuzzy rule expressed as: IF A and B and C and D and ETHEN Y (or Z) which "IF" immediately followed by the word says, "until THEN" after the word that follows. First of all the values the input of fuzzy S, M, L, XL, respectively, and the numbers 0, 1, 2, 3, namely: RW, RR, RA, TA, TH: S = 0, M = 1, 2, L = XL = 3. Each rule uses 3 bytes. 1 byte 4 bits high 1 front pieces of fuzzy values, low 4 bits 2 fuzzy values before parts; the first 2 bytes of the high 4 bits before 3, low 4 bits before 4; 3 byte 4 bits represent the high 5 front pieces, low 4 bits after the pieces. Where F represents the first piece, A representation is PVC, "B" may be PVC. As a rule (stored in ROM), for example, as shown in illustration. 2.4 fuzzy reasoning for each set of input data, first to blur and then iterates through every piece of fuzzy rules, take the first rule first pieces (3H) as an address offset, coupled with the fuzzy input in RAM RW deposited first address (40H), you can store from RAM RR in the zone to find out the membership grade A1 XL; take the first rule of the second front piece (3H) as an address offset, coupled with the fuzzy input in RAM RW deposited first address (44H), you can be kept from within the RW RAM area to find out the membership Grade B1 XL; and so on have membership C1, D1, E1. According to the intensity of transfer method, take the A1, B1, C1, D1, E1 in minimum value as the rules for the language variable "PVC" Y1 of membership. When all rules are: after traversal "PVC" membership Y1, Y2 ~ Ym, "may be PVC" membership Z1, Z2-Zn Y1, Y2 ~ by Ym's maximum value as the "PVC" membership Y, Z1, Z2-Zn maximum PVC as "might be" membership Z. If Y <7FH且Z<1FH,则输出"正常";如果Y>Z,输出"是PVC";如果Z> Y, output may be PVC.

3 conclusion this design of ECG monitor, set ECG collection, analysis, and the system volume is small, secure, reliable, and can not be custodians of the normal life of use, you can diagnose the illness, thereby initially save storage space, saving the time to further diagnose the illness, the experiment yielded very good results.

If you are able to combine more the aspect of medical experts of knowledge, you can achieve better diagnostic capabilities.