[embodiment]
The application's feature and advantage will be elaborated in conjunction with the accompanying drawings by embodiment.
Medical Devices (for example patient monitor) generally include physiologic parameter monitoring device and alarm triggering device, physiologic parameter monitoring device is used for the physiological parameter of detection of biological body, generally, most of physiological parameters all have certain physiological range, in the time of outside physiological parameter is in physiological range, may be that unfavorable condition or dangerous situation have appearred in monitored biosome, this moment, alarm triggering device need judge whether through default rule to trigger to report to the police to notify medical personnel.
Mostly be that the monitoring result of physiological parameter is done deterministic judgement, i.e. two-valued function in the prior art: report to the police/do not report to the police, and in fact, warning itself means that just dangerous probability is bigger, itself not clear and definite boundary of report to the police/not reporting to the police.
For can in time reporting to the police and reduce false alarm to dangerous situation, the present invention adopts at least two alarm factors, and utilize fuzzy logic theory to carry out fuzzy logic inference based on some semantic rules (reflecting its mutual restriction relation), and further reflect the information of warning intensity (priority) in the output of reporting to the police, thereby final As soon as possible Promising Policy is in time reported to the police and is reduced the requirement of frequent false alarm.
In an embodiment of the present invention, alarm triggering device comprises computing module, fuzzy reasoning module and alarm decision module as shown in Figure 2.Computing module is used for calculating according to the measured value of monitored physiological parameter the value of at least two alarm factors.The fuzzy reasoning module is used for drawing warning intensity by the value of at least two alarm factors through fuzzy logic inference.In one embodiment, the fuzzy reasoning module comprises and is used for respectively being determined the first module of the warning tendency of described at least two alarm factors and degree of membership and being used for warning tendency and degree of membership by at least two alarm factors draw warning intensity through fuzzy logic inference Unit second by the membership function of the value of described at least two alarm factors and default this alarm factors.Described Unit second is used for the warning of at least two alarm factors tendency is determined to be preset in alarm condition on the warning membership function figure through default semantic rule, draw the area that the degree of membership determined according to each semantic rule and alarm condition curve surround then, calculate the center of gravity that all surround the union of area at last, the horizontal ordinate of described center of gravity is warning intensity.In other embodiments, the fuzzy reasoning module also can adopt other fuzzy logic inference method to carry out reasoning.The alarm decision module is used for warning intensity and default final alarm limit are compared, if warning intensity surpasses final alarm limit, the triggering of then reporting to the police, otherwise the triggering of not reporting to the police.
In one embodiment, be that example describes to adopt three alarm factors to carry out fuzzy logic inference as input variable.Three alarm factors can be the degree of depth of crossing the border (Depth), the width that crosses the border (Width) and the ratio that crosses the border (CrossRate), also can be cross the border any two and other alarm factors in the degree of depth, the cross the border width and the ratio that crosses the border.In the present embodiment, be that the cross the border degree of depth, cross the border width and the ratio that crosses the border is that example describes with alarm factors.
Fig. 1 has provided the synoptic diagram of lower bound warning more.The degree of depth of crossing the border is the function of measured value and the difference d (t) of default alarming threshold value of the physiological parameter of current location, and what pre-set threshold value had been stipulated this physiological parameter goes up (descending) boundary, promptly
Depth=f(±d(t)),
Wherein ,+(-) correspondence goes up (descending) boundary's degree of depth more.F () can be made as the linear monotonically increasing function of single order or other increasing function.
The width that crosses the border is the function that continues the time span outside default alarming threshold value from the measured value that current location is searched for physiological parameter forward, promptly
Width=g(w(t)),
G () can be made as the linear monotonically increasing function of single order or other increasing function.
The ratio that crosses the border is meant in the past the certain time interval T O on the occasion of warning difference d (t) to the absolute value of the integration of time and warning difference d (t) ratio to time integral, promptly
Area ratio (A1+A3)/(A1+A2+A3) among Fig. 1 just.
The flow process that the alarm triggering device triggering is reported to the police may further comprise the steps as shown in Figure 3:
In step 31, calculate the value of the cross the border degree of depth, the cross the border width and three alarm factors of ratio of crossing the border according to the measured value of monitored physiological parameter and above-mentioned formula, execution in step 33 then.
In step 33, determine the warning tendency and the degree of membership of described at least two alarm factors respectively by the membership function of the value of the degree of depth of crossing the border, the cross the border width and three alarm factors of ratio of crossing the border and default this alarm factors.Cross the border the degree of depth, the cross the border width and the ratio that crosses the border the membership function curve as shown in Figure 4, membership function can be provided with as required, and membership function can be linear function (as the primary line shape function) or nonlinear function (as trigonometric function, quadratic function or trapezoidal function etc.).Shown in (a) and (b) among Fig. 4, the membership function of the degree of depth of crossing the border and the width that crosses the border has two states, represents two kinds of tendencies of reporting to the police, the probability height that " high " representative is reported to the police, and the probability that " low " representative is reported to the police is low.The membership function of ratio of crossing the border has three kinds of states shown in (c) among Fig. 4, represent three kinds of tendencies of reporting to the police.In this step, on its corresponding membership function curve map, determine the warning tendency of three alarm factors according to the value of the degree of depth of crossing the border that calculates, the cross the border width and three alarm factors of ratio of crossing the border, for example being " high " or " low ", perhaps both may be that " high " also may be " low ".And on its corresponding membership function curve map, determine the degree of membership of three alarm factors according to the value of the degree of depth of crossing the border that calculates, the cross the border width and three alarm factors of ratio of crossing the border, among Fig. 4, horizontal ordinate is represented the value of alarm factors, ordinate is represented degree of membership, and different warning tendencies may have different degrees of membership.Execution in step 35 then.
In step 35, the warning tendency of three alarm factors is determined alarm condition through default semantic rule on warning membership function figure.What semantic rule reflected is the relation of mutual restriction between the input variable, the output variable state that can draw according to the relation regulation of mutual restriction between one or more input variables, but semantic rule people is for setting.In the present embodiment, input variable is the degree of depth of crossing the border, the cross the border width and three alarm factors of ratio of crossing the border, output variable is report to the police (Alarm), preferably comprise three states: report to the police (YES), critical warning (Critical) and do not report to the police (NO), the membership function of output variable as shown in Figure 5.Alarm condition also can be to comprise two states: report to the police (YES) and do not report to the police (NO).The warning membership function also can be linear function (as the primary line shape function) or nonlinear function (as trigonometric function, quadratic function or trapezoidal function etc.).
In the present embodiment, preferably adopt the semantic rule of 5 IF-THEN as follows:
1.If(depth?is?high)and(width?is?high)then(alarm?is?Yes);
2.If(width?is?low)then(alarm?is?No);
3.If(depth?is?low)and(width?is?high)and(crossrate?is?low)then(alarm?is?No);
4.If(depth?is?low)and(width?is?high)and(crossrate?is?high)then(alarm?is?Yes);
5.If(depth?is?low)and(width?is?high)and(crossrate?is?medium)then(alarm?is?Critical)。
Semantic rule also can be set at 2,3 or 4 as required.
According to of the reasoning of the warning of three alarm factors tendency, determine 5 alarm condition through 5 semantic rules.Execution in step 37 then.
In step 37, draw the area that the degree of membership determined according to each semantic rule and alarm condition curve surround.If the logic of semantic rule be " with " logic, i.e. " AND ", then select in the degree of membership of All Alerts factor related in this semanteme rule minimum value for participating in the degree of membership of back reasoning, if the logic of semantic rule be " or " logic, i.e. " OR ", then select in the degree of membership of All Alerts factor related in this semanteme rule maximal value for participating in the degree of membership of back reasoning, write down the area that the minimum of each semantic rule correspondence or maximum membership degree and alarm condition curve surround.In the present embodiment, the logic of semantic rule be " with " logic, so should adopt minimum value in the degree of membership of All Alerts factor related in each semantic rule.Execution in step 38 then.
In step 38, calculate the center of gravity that all surround the union of area, can adopt the center of gravity of prior art reference area, in two-dimensional coordinate, center of gravity has horizontal stroke, vertical two coordinates, and in the present embodiment, the horizontal ordinate of center of gravity is a warning intensity, and execution in step 39 then.
In step 39, with warning intensity and default final alarm limit (for example 0.5) relatively, if warning intensity surpasses final alarm limit, the triggering of then reporting to the police, otherwise the triggering of not reporting to the police.
Fig. 6 is a synoptic diagram of determining warning intensity among a kind of embodiment, for the purpose of illustration, supposes input variable [depth, width, crossrate]=[3.5,3.2,0.65], and its degree of membership is respectively:
The degree of depth of crossing the border is " high ", and degree of membership is 0.28,
The degree of depth of crossing the border is " low ", and degree of membership is 0.3,
The width that crosses the border is " high ", and degree of membership is 0.2,
The width that crosses the border is " low ", and degree of membership is 0.4,
The ratio that crosses the border is " high ", and degree of membership is 0.18,
The ratio that crosses the border is " medium ", and degree of membership is 0.25.
Reasoning through above-mentioned 5 semantic rules, minimum degree of membership in adopting in the related degree of membership of every semantic rule, as a among Fig. 6, b, c, d, shown in the e, it is 0.2 and the area that surrounds for the curve of " YES " of alarm condition that shown area is respectively degree of membership, degree of membership is 0.4 and the area that surrounds for the curve of " NO " of alarm condition, degree of membership is 0 and the area that surrounds for the curve of " NO " of alarm condition, degree of membership is 0.18 and the area that surrounds for the curve of " YES " of alarm condition, degree of membership is 0.2 and the area that surrounds for the curve of " YES " of alarm condition.
Calculate the center of gravity that all surround the union of area, the horizontal ordinate of center of gravity is 0.411, draw warning intensity and equal 0.411, when warning intensity when formulating threshold values (as 0.5), start and report to the police.Among the figure, because 0.411<0.5, warning therefore do not triggered.
In another embodiment, to adopt two alarm factors to carry out fuzzy logic inference as input variable is that example describes, two alarm factors can be any two in crossing the border the degree of depth, the cross the border width and the ratio that crosses the border, and in the present embodiment, are that example describes with the degree of depth of crossing the border, the width that crosses the border.
Be depicted as input variable the cross the border degree of depth and the membership function of width of crossing the border as Fig. 7 (a) and (b), it has " high ", " low " two states, represent two kinds to report to the police and be inclined to.Fig. 8 is the membership function of output variable, and alarm condition comprises two states: report to the police (YES) and do not report to the police (NO).
In the present embodiment, adopt the semantic rule of 2 IF-THEN as follows:
1.If(depth?is?high)and(width?is?high)then(alarm?is?Yes)
2.If(width?is?low)then(alarm?is?No)
The logic of semantic rule be " with " logic, select that minimum value is the degree of membership of participation back reasoning in the degree of membership of All Alerts factor related in this semanteme rule.
Fig. 9 is the synoptic diagram of definite warning intensity of another kind of embodiment, for the purpose of illustration, supposes input variable [depth, width]=[7,4], and its degree of membership is respectively:
The degree of depth of crossing the border is " high ", and degree of membership is 0.62,
The width that crosses the border is " high ", and degree of membership is 0.31,
The width that crosses the border is " low ", and degree of membership is 0.25.
Reasoning through above-mentioned 2 semantic rules, adopt degree of membership minimum in every semantic rule, shown in a, b among Fig. 9, shown area be respectively degree of membership be 0.31 and the alarm condition area, the degree of membership that surround for the curve of " YES " be 0.25 and the area that surrounds for the curve of " NO " of alarm condition.
Calculate the center of gravity that all surround the union of area, the horizontal ordinate of center of gravity is 0.55, equals 0.55 so draw warning intensity, greater than final alarm limit 0.5, reports to the police so trigger.
The selected alarm factors of the present invention is during more than three, preferred scheme be wherein have at least two alarm factors to be selected from the degree of depth of crossing the border, the width and the ratio that crosses the border cross the border.
The membership function of above input variable and output variable should be selected according to experience, also can adjust when being necessary.
In the above-described embodiments, when calculating described area, the weight of described at least two alarm factors equates that for example weight all is 1, all is 1 as degree of membership maximal value among Fig. 4, Fig. 7.Can be unequal value with weight design as required also, the weight of the width that for example crosses the border be 2, and the weight of other alarm factors is 1.Weight is big more, and big more to the influence of the center of gravity of calculating all unions that surround area, the influence to warning intensity just is big more, thus the weight (being the maximal value of degree of membership) by alarm factors is set, the priority level that also can regulate warning.
In sum, because the warning demand has certain subjectivity, the present invention is by fuzzy set and fuzzy theory, and easier anthropomorphic dummy is to the requirement of warning.By defining suitable semantic rule, the present invention more can respond in time than existing method and report to the police and the minimizing false alarm.The proposition of warning intensity can be distinguished different priority of reporting to the police, and can allow the high patient of warning intensity obtain preferential intervention, further reduces dangerous generation.
Figure 10 uses blood oxygen saturation of the present invention (descending) boundary warning triggering synoptic diagram more, Figure 11 be use heart rate of the present invention more (on) boundary reports to the police and triggers synoptic diagram, as seen for the physiological parameter that exceeds alarming threshold value, alarm triggering device can both in time reflect the information of warning intensity (priority), and according to the triggering of reporting to the police of default final alarm limit.
The present invention has carried out off-line test on PC, the result shows that on to the correct base of recognition of reporting to the police, the present invention has shortened alarm time and release time than prior art.
Above content be in conjunction with concrete preferred implementation to further describing that the present invention did, can not assert that concrete enforcement of the present invention is confined to these explanations.For the general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, can also make some simple deduction or replace, all should be considered as belonging to protection scope of the present invention.