CN-122022678-A - Dynamic data flow prediction method for data processing service of industrial big data

CN122022678ACN 122022678 ACN122022678 ACN 122022678ACN-122022678-A

Abstract

The invention discloses a dynamic data flow prediction method of a data processing service of industrial big data, which relates to the technical field of data processing and comprises the steps of calculating tail section duty ratio, tail section concentration degree and front-back section balance degree based on distance measurement distance of intelligent material handling equipment, performing scene judgment on a current time window based on the tail section duty ratio, the tail section concentration degree and the front-back section balance degree to obtain a scene marking result, calculating a reinterpretation occupation evidence value corresponding to the current time window based on the scene marking result, and determining a reinterpretation occupation evidence value corresponding to a next time window through the reinterpretation occupation evidence value corresponding to the current time window. The invention improves the continuous operation capability of the intelligent material handling equipment in a cold chain low-temperature scene.

Inventors

LUO XIANLI
SHAN PINGPING

Assignees

上海安茁信息技术有限公司

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (8)

1. A dynamic data flow prediction method for industrial big data processing service is characterized by comprising the following steps: s1, calculating a tail section duty ratio, a tail section concentration degree and a front-rear section balance degree based on a distance measurement distance of intelligent material handling equipment; S2, performing scene judgment on the current time window based on the tail section duty ratio, the tail section concentration degree and the front and rear section balance degree to obtain a scene marking result; s3, calculating a reinterpretation occupation evidence value corresponding to the current time window based on the scene marking result; s4, determining the reinterpretation occupation evidence value corresponding to the next time window through the reinterpretation occupation evidence value corresponding to the current time window.
2. The method for predicting dynamic data flow of data processing service of industrial big data according to claim 1, wherein the specific steps of calculating the tail section duty ratio are as follows: Collecting a plurality of ranging distances of intelligent material handling equipment within a preset time window; sequencing a plurality of distance measurement distances to obtain an ascending distance sequence; Determining the median distance of the ascending distance sequence; screening distance measurement distances with values larger than the median distance from the ascending distance sequence to determine the distance measurement number of the tail section; dividing the distance measurement number of the tail sections by the total distance measurement number to obtain the tail section duty ratio.
3. The method for predicting dynamic data flow of data processing service of industrial big data according to claim 2, wherein the specific steps of calculating the tail segment concentration degree are as follows: determining a 90 percentile distance of the ascending sequence of distances; Calculating the distance standard deviation of the ascending sequence of distances; Performing difference calculation on each ranging distance in the ascending sequence of distances and 90 percentile distances to obtain a distance deviation; comparing and counting each ranging distance in the ascending distance sequence based on the absolute value of the distance deviation and the distance standard deviation to obtain the concentrated ranging number of the tail section; dividing the number of the concentrated ranging distances of the tail sections by the total number of the ranging distances to obtain the concentration degree of the tail sections.
4. A method for predicting dynamic data flow of data processing service of industrial big data according to claim 3, wherein the specific steps of calculating the degree of balance of the front and rear segments are as follows: Determining the number of forward ranging based on the comparison result of each ranging distance and the median distance; Determining the number of tail section ranging based on the comparison result of each ranging distance and the median distance; Dividing the forward ranging number by the tail section ranging number to obtain the front and rear section balance degree.
5. The method for predicting dynamic data flow of data processing service of industrial big data according to claim 1, wherein the specific steps of obtaining the scene marking result are as follows: combining the tail section duty ratio, tail section concentration degree and front and rear section balance degree into a tail track structure coding vector; Calculating the structural distance between the current time window and each historical time window according to the wake structure coding vector corresponding to the current time window and the wake structure coding vector corresponding to the historical time window; Ascending order is carried out on all the structure distances to obtain a structure distance sequence; calculating a structural distance median of the structural distance sequence; counting the number of targets with the structural distance smaller than or equal to the structural distance median; determining a total number of windows of the historical time window; Comparing the target number with half of the total number of windows to obtain a numerical comparison result; and performing scene judgment on the current time window according to the numerical comparison result to obtain a scene marking result of the current time window.
6. The method for dynamic data flow prediction for industrial big data processing service according to claim 2, wherein calculating the reinterpretation occupation evidence value corresponding to the current time window based on the scene mark result comprises: if the scene marking result is a non-trail dominant scene: allocating an occupation mark for each ranging distance in the current time window; summing all occupation marks in the current time window to obtain a reinterpretation occupation evidence value; if the scene marking result is a trail leading scene: dividing an ascending distance sequence corresponding to the current time window through the median distance corresponding to the current time window to obtain a forward ranging set and a tail section ranging set; and calculating a reinterpretation occupation evidence value corresponding to the current time window according to the forward ranging set and the tail section ranging set.
7. The method for dynamic data stream prediction for industrial big data processing service according to claim 6, wherein calculating the reinterpretation occupancy evidence value corresponding to the current time window according to the forward ranging set and the tail ranging set comprises: subtracting the tail section duty ratio corresponding to the current time window from 1 to obtain a tail weight coefficient corresponding to the current time window; assigning an occupancy contribution of a ranging distance in the forward ranging set to 1; Assigning a value to the occupation contribution of the ranging distance in the tail section ranging set according to the trail weight coefficient; Summing all occupation contributions in the forward ranging set to obtain a forward occupation summation value; summing all occupied contributions in the tail section ranging set to obtain a tail section occupied summation value; And adding the forward occupation summation value and the tail section occupation summation value to obtain a reinterpretation occupation evidence value.
8. The method for predicting dynamic data flow of data processing service of industrial big data according to claim 1, wherein determining the reinterpretation occupation evidence value corresponding to the next time window through the reinterpretation occupation evidence value corresponding to the current time window comprises: Determining a historical time window corresponding to the current time window; Based on the reinterpretation occupation evidence value corresponding to the historical time window and the current time window, the reinterpretation occupation evidence value corresponding to the next time window is predicted.

Description

Dynamic data flow prediction method for data processing service of industrial big data Technical Field The invention relates to the technical field of data processing, in particular to a dynamic data flow prediction method for data processing service of industrial big data. Background In the warehouse entry, door opening, buffer room and channel area of cold chain warehouse and low temperature warehouse, intelligent material handling equipment is required to finish operations such as picking, placing, traversing, transferring and the like under the environment of frequent exchange of low temperature and external hot and humid air. The distance measuring device carried by the carrying equipment is divided into a plurality of time windows according to time in the continuous operation process, a plurality of groups of distance measuring distances in the time windows are acquired and used as a part of industrial big data resources for aggregation, storage and calling, based on the data continuously arriving along time, data processing service often needs to output channel occupation evidence, and further gives out an occupation evidence prediction result of the next time window, so that the upper task allocation and path planning are adjusted in advance before channel congestion changes, and the continuity and safety of carrying operation are ensured. When freezing fog medium appears in the areas such as low-temperature door opening, the ranging echo can be reflected from the surface of a solid object, multiple scattering echo caused by fog drops or frost crystals can appear, so that obvious tailing and aggregation distribution phenomenon appears on one side of a larger distance of the ranging distance, if the ranging result in a time window is uniformly regarded as a real obstacle and directly accumulated as occupied evidence according to a conventional method, the channel occupancy degree is systematically overestimated, the deviation is transferred to occupied evidence prediction of the next time window, and channel blocking risk and traffic capacity judgment are distorted, so that task selection of conveying equipment and stability of a travelling path are affected. Disclosure of Invention The invention aims to solve the defects of channel blocking risk and traffic capacity judgment distortion in the prior art, and provides a dynamic data flow prediction method for data processing service of industrial big data. In order to solve the problems existing in the prior art, the invention adopts the following technical scheme: A data processing service dynamic data flow prediction method for industrial big data, comprising: s1, calculating a tail section duty ratio, a tail section concentration degree and a front-rear section balance degree based on a distance measurement distance of intelligent material handling equipment; S2, performing scene judgment on the current time window based on the tail section duty ratio, the tail section concentration degree and the front and rear section balance degree to obtain a scene marking result; s3, calculating a reinterpretation occupation evidence value corresponding to the current time window based on the scene marking result; s4, determining the reinterpretation occupation evidence value corresponding to the next time window through the reinterpretation occupation evidence value corresponding to the current time window. Preferably, the specific steps of calculating the tail section duty cycle are as follows: Collecting a plurality of ranging distances of intelligent material handling equipment within a preset time window; sequencing a plurality of distance measurement distances to obtain an ascending distance sequence; Determining the median distance of the ascending distance sequence; screening distance measurement distances with values larger than the median distance from the ascending distance sequence to determine the distance measurement number of the tail section; dividing the distance measurement number of the tail sections by the total distance measurement number to obtain the tail section duty ratio. Preferably, the specific steps of calculating the tail section concentration are as follows: determining a 90 percentile distance of the ascending sequence of distances; Calculating the distance standard deviation of the ascending sequence of distances; Performing difference calculation on each ranging distance in the ascending sequence of distances and 90 percentile distances to obtain a distance deviation; comparing and counting each ranging distance in the ascending distance sequence based on the absolute value of the distance deviation and the distance standard deviation to obtain the concentrated ranging number of the tail section; dividing the number of the concentrated ranging distances of the tail sections by the total number of the ranging distances to obtain the concentration degree of the tail sections. Preferably, the specific steps for calculating the fr