in the formula,. DELTA.f_i(j) Representing a sequence of differences Δ F_i(j) Of (d) is the j-th difference, Δ d_i(t) is a coefficient of difference measure, and Δ d_i(t)＝Δf_i(max)-Δf_i(min), wherein,. DELTA.f_i(max) denotes a sequence of differences Δ F_iMaximum value of (t), Δ f_i(min) represents the sequence of differences Δ F_i(t) minimum value, k is a given number of intervals, and

is a value function; when in use

When it is, then

When in use

When it is, then

e is the number of valid intervals, let e's initial value be 1, and increase by step 1, when theta_i(t) first satisfaction

Then, taking e at this time as the final effective interval number, and recording as e', selecting window sequence F_i(t) is satisfied

Data f of_i(j) Make up set F'_i(t) wherein f_i(j) And f_i(j-1) are window sequences F_i(t) th and (j-1) th data of the physiological parameter i;

according to set F'_i(t) determining a first detection threshold H₁(t) and a second detection threshold H₂(t), then H₁(t) and H₂The expression of (t) is:

in the formula,

denotes a set F'_iMean of the data in (t), f_i(k) Denotes a set F'_i(t) kth data of physiological parameter i, N (F'_i(t)) represents a set F'_i(t) number of data;

when f is_i(t)＜H₁(t) or f_i(t)＞H₂(t) determining the data f_i(t) is noise data, order

When f is_i(t) satisfies H₁(t)≤f_i(t)≤H₂(t) if f is judged_i(t) is valid data, let f'_i(t)＝f_i(t)。

The preferred embodiment is used for processing the acquired physiological parameter data, detecting the physiological parameter data by adopting a moving window mode, simultaneously including the physiological parameter data before the current moment and the physiological parameter data acquired after the current moment in the window sequence, screening the physiological parameter data in the window sequence to construct a difference sequence so as to avoid the influence of noise data acquired after the current moment in the window sequence on the processing of the physiological parameter data at the current moment, constructing k intervals according to the obtained difference measurement coefficients, reflecting the difference condition between adjacent data in the window sequence by the constructed k intervals, counting the data in the intervals, selecting the data in the interval with smaller difference to participate in the detection of the physiological parameter data at the current moment, and effectively screening the noise data in the window sequence, the influence of noise data acquired after the current moment in the window sequence on the physiological parameter data at the current moment is avoided; the first detection threshold and the second detection threshold constructed according to the selected physiological parameter data accord with the change rule of the physiological parameters, noise data can be effectively detected, and meanwhile the phenomenon that the fluctuation of the physiological parameter data caused by the movement of a user is mistaken for noise is avoided.

Preferably, the state evaluation unit is configured to evaluate the body state of the user at the current time according to the processed physiological parameter data, and includes an offline classification unit and an online evaluation unit, where the offline classification unit is configured to classify the collected historical physiological parameter data, and the online evaluation unit is configured to evaluate the body state of the user according to the collected physiological parameter data.

Preferably, the historical physiological parameter data includes labeled physiological parameter data and unlabeled physiological parameter data, the label includes a health label and a risk label, the offline classification unit is configured to classify the historical physiological parameter data, H represents the historical physiological parameter data set, and H ═ H {, where H represents the historical physiological parameter data set¹,H²,H³In which H¹Historical physiological parameter data set, H, representing labeled and labeled health²Historical physiological parameter data set, H, representing labeled and dangerous³Representing a non-labeled historical physiological parameter data set, let H (i) represent a set H³And h (i) ═ h (h) for the ith data point in (1)_x(i) X is 1,2, … n, where h is_x(i) The data points h (i) represent the values of the physiological parameters x corresponding to the data points h, n represents the types of the acquired physiological parameters; is provided with L_iA reference data set representing data points h (i), and L_i(j) h (i) -h (j) r (i), j 1,2, … n (i), where r (i) is a given reference threshold, and

h (L) is a data point directly adjacent to the data point h (i), L (i) represents a data point directly adjacent to the data point h (i), and h (j) represents the reference data set L_iWherein (j) th data point, n (i) represents the reference data set L_iIn (1)Counting the number of data points; for reference data set L_iDetecting when the reference data set L is detected_iWhen at least one data point with a label exists, predicting the label of the data point h (i), defining the label prediction function corresponding to the data point h (i) as P (i), wherein the expression of P (i) is as follows:

in the formula, eta (H (j), H¹) Is a value function, and

ρ(h(j),H²) Is a value function, and

eta (i) is a value function eta (H (j), H¹) Corresponding correction coefficient rho (i) is a value function rho (H (j), H²) Corresponding correction coefficients, and the expressions of η (i) and ρ (i) are:

wherein L is_j(j ═ 1,2, …, n (i)) represents a reference data set of data points h (j ═ 1,2, …, n (i)), and h (j ═ 1,2, …, n (i)) represents a reference data set L_iThe j-th data point in (a), n (j) represents the reference data set L_jH (L) represents a reference data set L_jThe ith data point in (1);

when the label prediction function P (i) > 1, the label of the data point H (i) is judged to be healthy, when the label prediction function P (i) < 1, the label of the data point H (i) is judged to be dangerous, when the label prediction function P (i) < 1, the data point H (i) is marked as quadratic prediction data, and when the set H is the set H³After the label prediction of all the data points is finished, carrying out label prediction on the marked secondary prediction data by adopting the method again;

after all data points in the set H have labels, the data points in the set H having health labels are classified into class C according to the labels of the data points¹Classifying data points in set H having danger labels into class C²。

The preferred embodiment is used for classifying the collected historical physiological parameter data and taking the final classification result as a reference value of the state evaluation unit; the historical physiological parameter data adopted by the preferred embodiment comprises a small amount of labeled physiological parameter data and a large amount of unlabeled physiological parameter data, when the historical physiological parameter data are classified, the labels of the unlabeled physiological parameter data are predicted according to the labeled physiological parameter data, and finally the classification of the historical physiological parameter data is completed according to the labels of the physiological parameter data; when the label of the non-label physiological parameter data is predicted, a data point which is closer to the data to be predicted is selected to construct a reference data set of the data to be predicted, and the label of the data to be predicted is predicted according to the labeled data in the reference data set, so that the label of the data to be predicted can be effectively predicted; defining a tag prediction function, wherein a value function in the tag prediction function can effectively count tagged data in a reference data set, and data points which are closer to the data to be predicted reflect the attributes of the data to be predicted to a great extent, so that the value function in the tag prediction function can effectively predict tags of the data to be predicted through counting tags of adjacent data points; correction coefficients are introduced into the label prediction function aiming at the value functions, the label prediction function can effectively predict labels of edge data between a healthy class and a dangerous class by the aid of the correction coefficients, when data to be predicted are located at the edge between the healthy class and the dangerous class, misjudgment is easily caused by label prediction only through counting labeled data in a reference data set of the data to be predicted, and the correction coefficients can effectively avoid detection errors caused by the fact that the data to be predicted are located at the edge by counting labeled data in a slightly far range, so that accuracy of classification results is improved, and accuracy of body state evaluation results of users is further improved.

Preferably, let f '(t) denote the physiological parameter data point at time t after processing, and f' (t) { f } { (t) }_i' (t), i ═ 1,2, … n }, where f_i' (t) represents the value of the physiological parameter i at the time t after processing, n represents the type of the acquired physiological parameter, and the physical state of the user at the current time is evaluated, specifically:

v. the¹Class C representing classification of offline classification units¹Is like the center of (1), and

wherein,

represents class C¹Class center of middle physiological parameter i, let B¹Represents class C¹A set of edge data points in (c), and

wherein,

represents a set B¹X-th data of middle physiological parameter i, m¹Represents a set B¹Number of data points in, let B²Represents class C²A set of edge data points in (c), and

wherein,

represents a set B²X-th data of middle physiological parameter i, m²Represents a set B²The number of data points in (1), defining a first evaluation coefficient G₁(t) and a second evaluation coefficient G₂(t), and G₁(t) and G₂The expressions of (t) are respectively:

in the formula,

represents a set B¹Middle distance data f_i' (t) the nearest edge data point,

represents a set B²Middle distance data f_i' (t) the nearest edge data point,

as a comparison function when

When it is, then

Otherwise

As a function of value when

When it is, then

Otherwise

Using a first evaluation coefficient G₁(t) evaluating the physical state of the user at the current moment, when the first evaluation coefficient G is larger than the first evaluation coefficient₁(t)≤d_min(B¹,v¹) If so, judging that the body state of the user at the current moment is healthy; when the first evaluation coefficient G₁(t)＞d_max(B¹,v¹) If so, judging that the physical state of the user at the current moment is dangerous; when d is_min(B¹,v¹)＜G₁(t)≤d_max(B¹,v¹) Then continue to use the second evaluation coefficient G₂(t) evaluating the physical state of the user at the current moment, when G₂When (t) is 1, the body state of the user at the current moment is judged to be healthy, and when G is used₂When (t) is 0, the physical state of the user at the current moment is judged to be dangerous, wherein d_min(B¹,v¹) And d_max(B¹,v¹) Respectively represent a set B¹From the edge data point to the class center v¹A minimum distance value and a maximum distance value.

The preferred embodiment compares the processed physiological parameter data with class centers of classes classified according to the offline classification result, thereby determining the physical health status of the user, and defining a first evaluation coefficient, which can effectively detect that the physical status of the user at the current time is in a relatively healthy or relatively dangerous status by comparing the processed physiological parameter data with the class center of the health class, and further, the preferred embodiment defines a second evaluation coefficient, the second evaluation coefficient is obtained by comparing the processed physiological parameter data with the marginal data points of the health class and the marginal data points of the risk class, therefore, when the physical state of the user is at a critical value of health and danger, the physical state of the user at the current moment can still be effectively detected, and the accuracy of the detection result is greatly improved.

Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the protection scope of the present invention, although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. The utility model provides an intelligence digit system of getting involved, characterized by, includes a plurality of intelligent runway display module, power module, mobile client and intelligent runway control end, intelligent runway display module arranges along runway surface, road surface or curb in succession for show image or characters, link to each other through the power cord between the adjacent intelligent runway display module, power module is used for giving intelligent runway display module supplies power, mobile client wears on one's body the user, including parameter setting unit, data acquisition unit and danger early warning unit, parameter setting unit is used for the user to set up the parameter of getting involved, data acquisition unit is used for gathering user's physiological parameter data, the parameter of getting involved of setting and the physiological parameter data transmission who gathers to intelligent runway control end, intelligent runway control end includes intelligent runway controller, intelligent runway control end, The intelligent runway controller controls images or characters to be sequentially displayed on the intelligent runway display module in a moving mode according to the running parameters set by a user, the data processing unit is used for processing the received physiological parameter data, the state evaluation unit is used for evaluating the body state of the user at the current moment according to the processed physiological parameter data, when the current body state of the user is evaluated to be dangerous, an early warning signal is sent to a mobile user end, and the danger early warning unit of the mobile user end carries out early warning;

the data processing unit is used for filtering the received physiological parameter data and setting f_i(t) data of physiological parameter i acquired at time t, and a window sequence F with a length of (2m +1) is set_i(t) and F_i(t)＝{f_i(t-m)，f_i(t-m+1)，...，f_i(t-1)，f_i(t)，f_i(t+1)，...，f_i(t+m-1)，f_i(t + m) }, wherein f_i(t-m)、f_i(t-m +1) data representing the physiological parameter i acquired at the time (t-m) and (t-m +1), respectively, f_i(t-1) and f_i(t +1) data representing the physiological parameter i acquired at the time of (t-1) and (t +1), respectively, f_i(t + m-1) and f_i(t + m) represents the data of the physiological parameter i collected at the time of (t + m-1) and (t + m), respectively, and a difference sequence delta F is set_i(t)＝{|f_i(t-m+1)-f_i(t-m)|，...，|f_i(t)-f_i(t-1)|，|f_i(t+1)-f_i(t)|，...，|f_i(t+m)-f_i(t+m-1)|}＝{Δf_i(j) J-t-m +1, t-m +2,.., t + m }, for window sequence F_i(t) counting the data, defining a statistical coefficient theta_i(t), and θ_iThe expression of (t) is:

is a value function; when in use

When it is, then

When in use

When it is, then

e is effective interval number, let e's initial value be 1, and carry out iteration increase with step length 1, when theta_i(t) first satisfaction

in the formula,

When f is_i(t) satisfies H₁(t)≤f_i(t)≤H₂(t) if f is judged_i(t) is valid data, let f_i′(t)＝f_i(t) wherein f_i' (t) denotes the data f_i(t) the value after the treatment.

2. The intelligent digital running system according to claim 1, wherein the intelligent runway display module is a full color LED matrix.

3. The intelligent digital running system according to claim 2, wherein the state evaluation unit is used for evaluating the physical state of the user at the current time according to the processed physiological parameter data, and comprises an offline classification unit and an online evaluation unit, the offline classification unit is used for classifying the collected historical physiological parameter data, and the online evaluation unit is used for evaluating the physical state of the user at the current time according to the processed physiological parameter data.

4. The intelligent digital running system according to claim 5, wherein the historical physiological parameter data comprises labeled physiological parameter data and unlabeled physiological parameter data, the labels comprise health labels and danger labels, and the offline classification unit is configured to classify the historical physiological parameter data.

5. The intelligent digital getting-off system according to claim 4, wherein H is said historical physiological parameter data set, and H ═ H¹，H²，H³In which H¹Historical physiological parameter data set, H, representing labeled and labeled health²Historical physiological parameter data set, H, representing labeled and dangerous³Representing a non-labeled historical physiological parameter data set, let H (i) represent a set H³And h (i) ═ h (h) for the ith data point in (1)_x(i) N, where h is 1,2_x(i) The data points h (i) represent the values of the physiological parameters x corresponding to the data points h, n represents the types of the acquired physiological parameters; is provided with L_iA reference data set representing data points h (i), and L_iN (i) }, where r (i) is a given reference threshold, and (j) | h (i) -h (j) | < r (i)

h (L) is a data point directly adjacent to the data point h (i), L (i) represents a data point directly adjacent to the data point h (i), and h (j) represents the reference data set L_iWherein (j) th data point, n (i) represents the reference data set L_iThe number of data points in; for reference data set L_iDetecting when the reference data set L is detected_iWhen at least one data point with a label exists, predicting the label of the data point h (i), defining the label prediction function corresponding to the data point h (i) as P (i), wherein the expression of P (i) is as follows:

in the formula, eta (H (j), H¹) Is a value function, and

ρ(h(j)，H²) Is a value function, and

wherein L is_j(j 1, 2.. and n (i)) represents a reference data set of data points h (j) (1, 2.. and n (i)), and h (j) (1, 2.. and n (i)) is a reference data set L (j)_iThe j-th data point in (a), n (j) represents the reference data set L_jH (L) represents a reference data set L_jThe ith data point in (1);

when all data points in the set H have labels, the set H has health according to the labels of the data pointsData point classification of tags into class C¹Classifying data points in set H having danger labels into class C²。