CN114928653B - Data processing methods and devices for crowd intelligence sensing - Google Patents

Abstract

The embodiment of the invention discloses a data processing method and a device for crowd sensing, wherein the data processing method for crowd sensing comprises the following steps: acquiring perception task data to be processed of the terminal equipment i; determining a task segmentation strategy according to reference conditions, wherein the reference conditions comprise at least one of the following: the terminal equipment i, the edge equipment and the cloud center are busy, task processing speed and data transmission rate respectively; the control terminal equipment i segments the perception task data according to a segmentation strategy; and controlling at least one of the terminal equipment i, the edge equipment and the cloud center to process the segmented data according to the segmentation condition of the perception task data. The invention solves the technical problem of slower processing speed when processing the perception data in the related technology.

Description

Data processing methods and devices for crowd intelligence sensing

Technical field

The present invention relates to the field of crowd intelligence sensing, and in particular to a data processing method and device for crowd intelligence sensing.

Background technique

With the rapid popularization of crowd intelligence mobile devices and the exponential growth of IOT sensing devices, the types and data volumes of crowd sensing (MCS: Mobile Crowd Sensing) tasks have also increased, especially computationally intensive and delay-sensitive sensing. The emergence and rapid increase of tasks, such as autonomous driving road condition analysis, AR/VR, disaster scene information modeling such as earthquakes, floods and fires, etc. This leads to the problem of excessive computational delay. Unreasonable allocation of computing resources.

In related technologies, during the processing of sensing tasks, when faced with a large amount of calculation, the data of the end device will be completely offloaded to the edge device or the cloud side. The method of increasing the computing speed is the same as increasing the end device. Consider it from the perspective of computing performance, or increasing the number and performance of edge nodes. However, there is an unreasonable task allocation situation, and the computing resources of each terminal and edge cloud cannot be fully utilized, which not only causes a waste of resources, but also causes the crowd intelligence sensing task processing speed to be slow.

In response to the above problems, no effective solution has yet been proposed.

The above information disclosed in the Background section is only provided to enhance understanding of the background of the technology described herein. Therefore, the Background Art may contain information that does not form the prior art known to those skilled in the art.

Contents of the invention

Embodiments of the present invention provide a data processing method and device for crowd intelligence sensing, so as to at least solve the technical problem of slow processing speed when processing sensing data existing in related technologies.

According to the first aspect of the embodiment of the present invention, a data processing method for crowd sensing is provided, including: obtaining the sensing task data to be processed by the terminal device i; determining a task division strategy according to reference conditions, and the reference conditions include at least the following: One: the respective busyness, task processing speed and data transmission rate of the terminal device i, edge device and cloud center; control the terminal device i to segment the sensing task data according to the segmentation strategy; according to the segmentation of the sensing task data, control At least one of the terminal device i, the edge device and the cloud center processes the divided data.

Further, the reference conditions also include the power of the terminal device i And the total energy consumption E _i of the terminal device i for processing the sensing task data acquired by itself.

Further, determining the task division strategy based on reference conditions includes: based on goals and constraints Determine the segmentation ratio of the perception task data; where, /> T _i ^L is the total delay for terminal device i to process the sensing task data it acquires,/> is the total delay for edge device j to process the sensing task data obtained by terminal device i, T _i ^C is the total delay for the cloud center to process the sensing task data obtained by terminal device i, Π is the offloading decision set of all terminal devices ∏= {∏ _i ,i∈N}, ∏ _i is the offloading decision vector of the sensing task data obtained by the terminal device i, Π _i = {xi _, y _i,j ,z _i }, _xi is the offloading decision vector for the terminal device i The proportion of sensing task data of device i, yi _,j is the proportion of sensing task data divided to edge device j, z _i is the proportion of sensing task data divided to cloud center, N is the total number of terminal devices, f is all The computing resource allocation set of edge devices f = {f _i,j ,i∈N,j∈M}, f _i,j is the computing resource allocation vector of edge device j that processes sensing task data from terminal device i, M is the total number of edge devices.

further, 1-μ _i1 is the proportion of CPU cycles required by terminal device i to process the sensing task data acquired by itself, K _i is the total number of CPU cycles required to process the sensing task data acquired by terminal device i, and f _i is the number of CPU cycles processed by terminal device i. The amount of computing power allocated when it obtains the sensing task data itself;/> is the delay for edge device m _i to transmit sensing task data obtained by terminal device i to edge device j,/> Calculate the latency of sensing task data obtained by terminal device i for edge device j;/> T _c is the preset delay for the edge device to transmit the sensing task data obtained by the terminal device i to the cloud center,/> Calculate the latency of sensing task data obtained by terminal device i for the cloud center; Calculate the energy consumption of the sensing task data acquired by the terminal device i, It is the energy consumption of the terminal device i to transmit the sensing task data acquired by itself to the edge device _mi .

Further, under energy consumption constraints, according to the goals and constraints Determine the split ratio of perception task data, including: in/> In the case of , it is determined that x _i =1, y _i,j =0, z _i =0; where, C _i is the calculation rate when the terminal device i calculates the sensing task data it acquires,/> is the transmission rate when the terminal device i transmits the sensing task data it acquires to the edge device m _i .

Further, under energy consumption constraints, according to the goals and constraints Determining the segmentation ratio of perception task data also includes:/> In the case of , if the amount of task data to be processed by edge device j is less than or equal to B _i , then it is determined that x _i <1, y _i,j <0, z _i =0; where, //> C _i,j is the calculation rate at which edge device j calculates the sensing task data obtained by terminal device i.

Further, under energy consumption constraints, according to the goals and constraints Determining the split ratio of perception task data also includes: in/> And when the amount of pending task data of edge device j is greater than B _i , according to the formula/> Calculate _Xi , Yi _, Zi _, and then determine the split ratio based on Xi _, Yi _{, Zi} ; where, _Xi is the amount of sensing task data split to terminal device i, and _Yi is split to edge device j The data volume of sensing task data, _Zi is the data volume of sensing task data divided to the cloud center, Xi ₊ Y _i +Z _i =L _i , x _i :y _i :z _i =X _i :Y _i : Z _i and _Li are the total amount of sensing task data.

Further, under energy consumption constraints, according to the goals and constraints Determining the division ratio of sensing task data also includes: when the terminal device i is busy, according to the formula /> Calculate the data volume threshold X _Ti divided to terminal device i; and/or, when edge device j is busy, according to the formula Calculate the threshold Y _Ti for the amount of data split to edge device j; determine the split ratio based on at least one of X _Ti and Y _Ti and the total amount of sensing task data; where,/> is the waiting time on terminal device i;/> Indicates the waiting time on edge device j,/> is the time for calculation on terminal device i,/> is the transmission rate when terminal device i transmits the sensing task data acquired by itself to edge device m _i, and C _i,j is the calculation rate at which edge device j calculates the sensing task data acquired by terminal device i.

According to a second aspect of the embodiment of the present invention, a data processing device for crowd sensing is also provided, including: an acquisition unit for acquiring sensing task data to be processed by the terminal device; and a determining unit for determining based on reference conditions Determine the task division strategy, and the reference conditions include at least one of the following: the respective busyness, task processing speed, and data transmission rate of the terminal device, edge device, and cloud center; the division unit is used to control the terminal device to perform sensing tasks according to the division strategy The data is divided; the control unit is used to control at least one of the terminal device, the edge device and the cloud center to process the divided data according to the division of the sensing task data.

The data processing method for crowd sensing in the embodiment of the present invention includes: obtaining the sensing task data to be processed by the terminal device i; determining the task division strategy according to the reference conditions, and the reference conditions include at least one of the following: terminal device i, edge device and The respective busy status, task processing speed and data transmission rate of the cloud center; control the terminal device i to segment the sensing task data according to the segmentation strategy; control the terminal device i, edge device and cloud center according to the segmentation of the sensing task data At least one of them processes the split data. By adopting this processing method, the task division strategy is determined based on at least one of the busyness, task processing speed and data transmission rate of the terminal device i, the edge device and the cloud center, and the task data is processed according to the division strategy. segmentation, and then controls at least one of the terminal device i, the edge device, and the cloud center to process the segmented data according to the segmentation result of the task. In this way, the perception can be combined with the actual situation of the terminal device i, the edge device, and the cloud center. The task data is divided and then processed separately, thereby making full use of the computing power of end devices, edge devices and cloud centers, optimizing resource allocation, allowing the crowd intelligence sensing platform to specify resource scheduling plans in a more efficient way, and deepen end-end Edge-cloud collaboration enables the entire platform to process crowd sensing task data and feedback results in less time, effectively increasing the processing speed of sensing task data and shortening processing waiting time. It is especially suitable for computing-intensive and delay-sensitive applications. The sensory data processing task solves the problem of slow processing speed in processing sensory data in related technologies.

Description of drawings

The drawings described here are used to provide a further understanding of the present invention and constitute a part of this application. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the attached picture:

Figure 1 is a schematic flow chart of a data processing method for crowd intelligence sensing provided by an embodiment of the present invention;

FIG. 2 is a schematic diagram of a data processing device for crowd intelligence sensing provided by an embodiment of the present invention.

Detailed ways

In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only These are some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts should fall within the scope of protection of the present invention.

It should be noted that the terms "first", "second", etc. in the description, claims and drawings of the present invention are used to distinguish different objects, rather than to limit a specific order.

Figure 1 is a data processing method for crowd intelligence sensing according to an embodiment of the present invention. As shown in Figure 1, the method includes the following steps:

Step S102, obtain the sensing task data to be processed by the terminal device i;

Step S104, determine the task division strategy based on reference conditions, which include at least one of the following: the respective busyness, task processing speed, and data transmission rate of the terminal device i, the edge device, and the cloud center;

Step S106, control the terminal device i to segment the sensing task data according to the segmentation strategy;

Step S108: Control at least one of the terminal device i, the edge device, and the cloud center to process the segmented data according to the segmentation situation of the sensing task data.

The data processing method for crowd sensing using the above solution includes: obtaining the sensing task data to be processed by the terminal device i; determining the task division strategy according to the reference conditions, which include at least one of the following: terminal device i, edge device and cloud The respective busy status, task processing speed and data transmission rate of the center; control the terminal device i to segment the sensing task data according to the segmentation strategy; control the terminal device i, edge device and cloud center according to the segmentation of the sensing task data At least one of them processes the split data. By adopting this processing method, the task division strategy is determined based on at least one of the busyness, task processing speed and data transmission rate of the terminal device i, the edge device and the cloud center, and the task data is processed according to the division strategy. segmentation, and then controls at least one of the terminal device i, the edge device, and the cloud center to process the segmented data according to the segmentation result of the task. In this way, the perception can be combined with the actual situation of the terminal device i, the edge device, and the cloud center. The task data is divided and then processed separately, thereby making full use of the computing power of end devices, edge devices and cloud centers, optimizing resource allocation, allowing the crowd intelligence sensing platform to specify resource scheduling plans in a more efficient way, and deepen end-end Edge-cloud collaboration enables the entire platform to process crowd sensing task data and feedback results in less time, effectively increasing the processing speed of sensing task data and shortening processing waiting time. It is especially suitable for computing-intensive and delay-sensitive applications. The sensory data processing task solves the problem of slow processing speed in processing sensory data in related technologies.

Specifically, the reference conditions also include the power of the terminal device i And the total energy consumption E _i of the terminal device i for processing the sensing task data acquired by itself.

By incorporating the power of the terminal device into the reference range, the energy consumption of the sensing task data processing process and the power of the terminal device can be combined to better optimize the division of tasks, which is conducive to controlling the energy consumption of the terminal device and reducing the risk of Unreasonable task division leads to the risk of increased energy consumption of terminal equipment or even insufficient power of terminal equipment.

Specifically, determining the task segmentation strategy based on reference conditions includes: based on goals and constraints Determine the segmentation ratio of the perception task data; where, /> T _i ^L is the total delay for terminal device i to process the sensing task data it acquires,/> is the total delay for edge device j to process the sensing task data obtained by terminal device i, T _i ^C is the total delay for the cloud center to process the sensing task data obtained by terminal device i, Π is the offloading decision set of all terminal devices Π = {Π _i ,i∈N}, Π _i is the offloading decision vector of the sensing task data obtained by the terminal device i, Π _i ={xi _, y _i,j ,z _i }, xi _i is the offloading decision vector for the terminal device i The proportion of sensing task data of device i, yi _,j is the proportion of sensing task data divided to edge device j, z _i is the proportion of sensing task data divided to cloud center, N is the total number of terminal devices, f is all The computing resource allocation set of edge devices f = {f _i,j ,i∈N,j∈M}, f _i,j is the computing resource allocation vector of edge device j that processes sensing task data from terminal device i, M is the total number of edge devices.

By setting the above goals and constraints, reasonable task segmentation conditions are designed. All segmentation situations that meet the above goals and constraints can well optimize the perception task data processing process, thereby improving the processing speed of perception task data.

specifically, 1-μ _i1 is the proportion of CPU cycles required by terminal device i to process the sensing task data acquired by itself, K _i is the total number of CPU cycles required to process the sensing task data acquired by terminal device i, and f _i is the number of CPU cycles processed by terminal device i. The amount of computing power allocated when it obtains the sensing task data itself;/> is the delay for edge device m _i to transmit sensing task data obtained by terminal device i to edge device j,/> Calculate the latency of sensing task data obtained by terminal device i for edge device j;/> T _c is the preset delay for the edge device to transmit the sensing task data obtained by the terminal device i to the cloud center,/> Calculate the latency of sensing task data obtained by terminal device i for the cloud center; Calculate the energy consumption of the sensing task data acquired by the terminal device i, It is the energy consumption of the terminal device i to transmit the sensing task data acquired by itself to the edge device _mi .

In actual implementation, after the terminal device i transmits the sensing task data to the edge device _mi , the edge device _mi may have different busy states. If the tasks queued by the edge device _mi exceed the threshold, the edge device _mi will receive the data. The received data is transmitted to edge device j. Of course, edge device m _i and edge device j can be the same device, that is, the data to be processed by edge device m _i does not exceed the threshold. At this time, it can process the sensing task data from terminal device i by itself. At this time, both No data transmission is required between That is 0.

In a specific embodiment, under energy consumption constraints, according to the goals and constraints Determine the split ratio of perception task data, including: in/> In the case of , it is determined that x _i =1, y _i,j =0, z _i =0; where, C _i is the calculation rate when the terminal device i calculates the sensing task data it acquires,/> is the transmission rate when the terminal device i transmits the sensing task data it acquires to the edge device m _i .

Under energy consumption constraints, according to goals and constraints Determining the split ratio of perception task data also includes: in/> In the case of , if the amount of task data to be processed by edge device j is less than or equal to B _i , then it is determined that x _i <1, y _i,j <0, z _i =0; where, //> C _i,j is the calculation rate at which edge device j calculates the sensing task data obtained by terminal device i.

Specifically, under energy consumption constraints, according to the goals and constraints Determining the split ratio of perception task data also includes: in/> And when the amount of pending task data of edge device j is greater than B _i , according to the formula/> Calculate _Xi , Yi _, Zi _, and then determine the split ratio based on Xi _, Yi _{, Zi} ; where, _Xi is the amount of sensing task data split to terminal device i, and _Yi is split to edge device j The data volume of sensing task data, _Zi is the data volume of sensing task data divided to the cloud center, Xi ₊ Y _i +Z _i =L _i , x _i :y _i :z _i =X _i :Y _i : Z _i and _Li are the total amount of sensing task data.

In the presence of busy end devices and/or edge devices, subject to energy consumption constraints, according to goals and constraints Determining the division ratio of sensing task data also includes: when the terminal device i is busy, according to the formula /> Calculate the data volume threshold X _Ti divided to terminal device i; and/or, when edge device j is busy, according to the formula /> Calculate the threshold Y _Ti for the amount of data split to edge device j; determine the split ratio based on at least one of X _Ti and Y _Ti and the total amount of sensing task data; where,/> is the waiting time on terminal device i;/> Indicates the waiting time on edge device j,/> is the time for calculation on terminal device i,/> is the transmission rate when terminal device i transmits the sensing task data acquired by itself to edge device m _i, and C _i,j is the calculation rate at which edge device j calculates the sensing task data acquired by terminal device i.

Secondly, as shown in Figure 2, embodiments of the present invention also provide a data processing device for crowd sensing, which includes: an acquisition unit for acquiring sensing task data to be processed by the terminal device; a determining unit for Determine the task division strategy according to the reference conditions, which include at least one of the following: the respective busyness, task processing speed and data transmission rate of the terminal device, edge device and cloud center; the division unit is used to control the terminal device according to the division strategy Segment the sensing task data; the control unit is used to control at least one of the terminal device, the edge device and the cloud center to process the segmented data according to the segmentation of the sensing task data.

Specifically, the reference conditions also include the power of the terminal device i And the total energy consumption E _i of the terminal device i for processing the sensing task data acquired by itself.

By incorporating the power of the terminal device into the reference range, the energy consumption of the sensing task data processing process and the power of the terminal device can be combined to better optimize the division of tasks, which is conducive to controlling the energy consumption of the terminal device and reducing the risk of Unreasonable task division leads to the risk of increased energy consumption of terminal equipment or even insufficient power of terminal equipment.

Specifically, the determination unit is used to: According to the goals and constraints Determine the segmentation ratio of the perception task data; where, /> T _i ^L is the total delay for terminal device i to process the sensing task data it acquires,/> is the total delay for edge device j to process the sensing task data obtained by terminal device i, T _i ^C is the total delay for the cloud center to process the sensing task data obtained by terminal device i, П is the offloading decision set П= for all terminal devices {Π _i ,i∈N}, П _i is the offloading decision vector of the sensing task data obtained by the terminal device i, П _i = {xi _, y _i,j ,z _i }, _xi is the offloading decision vector for the terminal device i The proportion of sensing task data of device i, yi _,j is the proportion of sensing task data divided to edge device j, z _i is the proportion of sensing task data divided to cloud center, N is the total number of terminal devices, f is all The computing resource allocation set of edge devices f = {f _i,j ,i∈N,j∈M}, f _i,j is the computing resource allocation vector of edge device j that processes sensing task data from terminal device i, M is the total number of edge devices.

By setting the above goals and constraints, reasonable task segmentation conditions are designed. All segmentation situations that meet the above goals and constraints can well optimize the perception task data processing process, thereby improving the processing speed of perception task data.

specifically, 1-μ _i1 is the proportion of CPU cycles required by terminal device i to process the sensing task data acquired by itself, K _i is the total number of CPU cycles required to process the sensing task data acquired by terminal device i, and f _i is the number of CPU cycles processed by terminal device i. The amount of computing power allocated when it obtains the sensing task data itself;/> is the delay for edge device m _i to transmit sensing task data obtained by terminal device i to edge device j,/> Calculate the latency of sensing task data obtained by terminal device i for edge device j;/> T _c is the preset delay for the edge device to transmit the sensing task data obtained by the terminal device i to the cloud center,/> Calculate the latency of sensing task data obtained by terminal device i for the cloud center; Calculate the energy consumption of the sensing task data acquired by the terminal device i,/> It is the energy consumption of the terminal device i to transmit the sensing task data acquired by itself to the edge device _mi .

In actual implementation, after the terminal device i transmits the sensing task data to the edge device _mi , the edge device _mi may have different busy states. If the tasks queued by the edge device _mi exceed the threshold, the edge device _mi will receive the data. The received data is transmitted to edge device j. Of course, edge device m _i and edge device j can be the same device, that is, the data to be processed by edge device m _i does not exceed the threshold. At this time, it can process the sensing task data from terminal device i by itself. At this time, both No data transmission is required between That is 0.

In a specific embodiment, under energy consumption constraints, the determination unit includes a first determination module: the first determination module is used to In the case of , it is determined that x _i =1, y _i,j =0, z _i =0; where, C _i is the calculation rate when the terminal device i calculates the sensing task data it acquires,/> is the transmission rate when the terminal device i transmits the sensing task data it acquires to the edge device m _i .

Under the energy consumption constraint, the determination unit also includes a second determination module, and the second determination module is used to In the case of , if the amount of task data to be processed by edge device j is less than or equal to B _i , then it is determined that x _i <1, y _i,j <0, z _i =0; where, //> C _i,j is the calculation rate at which edge device j calculates the sensing task data obtained by terminal device i.

Specifically, under energy consumption constraints, the determination unit also includes a third determination module: the third determination module is used to And when the amount of pending task data of edge device j is greater than B _i , according to the formula Calculate _Xi , Yi _, Zi _, and then determine the split ratio based on Xi _, Yi _{, Zi} ; where, _Xi is the amount of sensing task data split to terminal device i, and _Yi is split to edge device j The data volume of sensing task data, _Zi is the data volume of sensing task data divided to the cloud center, Xi ₊ Y _i +Z _i =L _i , x _i :y _i :z _i =X _i :Y _i : Z _i and _Li are the total amount of sensing task data.

In the case that the terminal device and/or the edge device is busy, under the energy consumption constraint, the determination unit is also used: in the case that the terminal device i is busy, according to the formula Calculate the data volume threshold X _Ti divided to terminal device i; and/or, when edge device j is busy, according to the formula/> Calculate the threshold Y _Ti for the amount of data split to edge device j; determine the split ratio based on at least one of X _Ti and Y _Ti and the total amount of sensing task data; where,/> is the waiting time on terminal device i;/> Indicates the waiting time on edge device j,/> is the time for calculation on terminal device i,/> is the transmission rate when terminal device i transmits the sensing task data acquired by itself to edge device m _i, and C _i,j is the calculation rate at which edge device j calculates the sensing task data acquired by terminal device i.

Below, the data processing method for crowd sensing of the present invention will be described with reference to a specific embodiment, which includes the following steps:

Step 1: Record all information about sensing tasks, network and device attributes in the cloud;

Step 2: Calculate the energy consumption and delay of processing the sensing task data _Qi on the end node:

The total local processing delay is

Among them, 1-μ _i1 is the proportion of CPU cycles required by terminal device i to process the sensing task data acquired by itself, and K _i is the total number of CPU cycles required to process the sensing task data acquired by terminal device i.

The energy consumption of local processing _Qi is expressed as

Step 3: Calculate the processing delay and transmission energy consumption of sensing task data _Qi on edge node j. The total computing and transmission delay on the execution device is The energy consumption of data transmission is:

Step 4: Calculate the total delay in processing the sensing task data _Qi on the cloud center;

Step 5: Combining the above steps, the total delay and total energy consumption of Q _i processing can be obtained as:

Through the above description, the partial offloading decision vector of device i is recorded as Π _i = {x _i , y _i,j , z _i }, the offloading decision set of all devices is Π = {Π _i , i∈N}, and the The computing resource allocation vector is recorded as f={f _i,j ,i∈N,j∈M}, and the battery capacity of the device is Therefore, the goals and constraints of the resource allocation plan under partial offloading are expressed as follows:

Step 6: Determine partial segmentation thresholds

The first is the segmentation threshold between the end device and the edge node. Ci represents the processing rate of the task on the end device i. A _i represents the maximum amount of locally computable data when it is only calculated locally. It is expressed by the formula: (/> is the transmission rate, f _i is the local computing power allocation) is simplified to:

By default, the delay in establishing a transmission connection and receiving is not included. This is satisfied under the energy consumption constraint, that is, on the premise that the local device i can normally handle all its tasks. This formula indicates the processing rate of task data of device i/> Then it chooses to perform calculations locally under energy consumption constraints, otherwise it is partially offloaded or transmitted to edge nodes for processing.

Secondly, in the case of collaboration between the edge node and the cloud center, B _i represents the maximum amount of sensing task data that the edge node can process when the cloud is not required for collaborative computing. On the basis of minimizing the delay, it satisfies the formula Ignoring the impact of network instability and the delay in establishing a network connection, it can be simplified to:

B _i ·C _i,j = _TC '·f _i,j ; where, _TC '= _TC , is a known quantity, so B _i in the critical state can be determined, and then the task division strategy can be determined;

Again, in a locally busy state, that is, the task needs to wait for local processing. At this time, the waiting time of the queue needs to be taken into account when dividing the task. This situation is similar when the edge node is busy. The division of task processing in the local busy state is expressed by the formula as follows:

Among them, X _Ti represents the data volume threshold for offloading in the local busy state; Indicates the waiting time on local device i;/> Indicates the time that task data needs to wait on the edge node.

Again, the situation analysis when the edge node is busy satisfies the formula This formula needs to ensure that the data is just completely transmitted from the local device to the edge node, and the local device begins to calculate the remaining sensing data except A _i .

in Indicates the time to perform calculations locally on the data, and Y _Ti indicates the data volume threshold for partial unloading of data when the edge node is busy.

Finally, in the case of device-edge-cloud collaborative work, the partial offloading of device-edge-cloud collaborative computing is as follows:

And it also satisfies:

X _i +Y _i +Z _i =L _i , x _i :y _i :z _i =X _i :Y _i :Z _i ;

According to the above formula, the values of the data volume thresholds X _i , Y _i and Z _i in various situations can be calculated, and the offloading plan and execution plan of each task in the entire partial offloading system can be obtained.

Step 7: Then segment the sensing task processing process according to the segmentation threshold, and then use the optimized ADMM algorithm to partially offload the sensing data to minimize the total delay.

At this point, the minimized delay of partial offloading under crowd sensing terminal-edge-cloud collaboration can be achieved. The above-mentioned terminal and local refer to the terminal.

The above serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. Furthermore, the steps shown in the flowchart illustrations of the figures may be performed in a computer system, such as a set of computer-executable instructions, and, although a logical sequence is shown in the flowchart illustrations, in some cases, may be performed in a different order. The steps shown or described are performed in the order herein.

In the above-mentioned embodiments of the present invention, each embodiment is described with its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units may be a logical functional division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or may be Integrated into another system, or some features can be ignored, or not implemented. On the other hand, the coupling or direct coupling or communication connection between each other shown or discussed may be through some interfaces, and the indirect coupling or communication connection of the units or modules may be in electrical or other forms.

The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or software functional units.

If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or contributes to the existing technology or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions to cause a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in various embodiments of the present invention. The aforementioned storage media include: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program code. .

The above are only preferred embodiments of the present invention. It should be noted that those skilled in the art can make several improvements and modifications without departing from the principles of the present invention. These improvements and modifications can also be made. should be regarded as the protection scope of the present invention.

Claims (2)

1. A data processing method for crowd intelligence sensing, which is characterized by including: Obtain the sensing task data to be processed by terminal device i; Determine the task division strategy according to reference conditions, which include at least one of the following: the respective busyness, task processing speed, and data transmission rate of the terminal device i, edge device, and cloud center; Control the terminal device i to segment the sensing task data according to the segmentation strategy; According to the segmentation situation of the sensing task data, control at least one of the terminal device i, the edge device and the cloud center to process the segmented data; The reference condition also includes the power of the terminal device i And the total energy consumption E _i of the terminal device i for processing the sensing task data acquired by itself; Determining task segmentation strategies based on reference conditions includes: According to goals and constraints Determine the segmentation ratio of the perception task data; in, T _i ^L is the total delay for the terminal device i to process the sensing task data acquired by itself, /> is the total delay for the edge device j to process the sensing task data obtained by the terminal device i, T _i ^C is the total delay for the cloud center to process the sensing task data obtained by the terminal device i, Π is A set of uninstall decisions for all terminal devices/> is the offloading decision vector of the sensing task data obtained by the terminal device i,/> _xi is the proportion of the sensing task data divided to the terminal device i, y _i,j is the proportion of the sensing task data divided to the edge device j, z _i is the proportion of the sensing task data divided to the cloud center The proportion of the sensing task data, N is the total number of the terminal devices, f is the computing resource allocation set of all the edge devices f = {f _{i, j} , i∈N, j∈M}, fi _{, j} is the computing resource allocation vector of the edge device j that processes the sensing task data from the terminal device i, and M is the total number of edge devices; 1-μ _i1 is the proportion of CPU cycles used by the terminal device i to process the sensing task data acquired by itself, and K _i is the total number of CPU cycles required to process the sensing task data acquired by the terminal device i , f _i is the computing power allocation amount when the terminal device i processes the sensing task data obtained by itself;/> is the delay for edge device m _i to transmit the sensing task data obtained by terminal device i to edge device j,/> Calculate the delay of the sensing task data obtained by the terminal device i for the edge device j;/> _Tc is the preset delay for the edge device to transmit the sensing task data obtained by the terminal device i to the cloud center,/> Calculate the delay of the sensing task data obtained by the terminal device i for the cloud center;/> Calculate the energy consumption of the sensing task data acquired by the terminal device i,/> The energy consumption for the terminal device i to transmit the sensing task data obtained by itself to the edge device _mi ; Under energy consumption constraints, according to goals and constraints Determining the segmentation ratio of the perception task data includes: exist In the case of , it is determined that x _i =1, y _i,j =0, z _i =0; Wherein, C _i is the calculation rate when the terminal device i calculates the sensing task data obtained by itself, is the transmission rate when the terminal device i _transmits the sensing task data obtained by itself to the edge device mi; Under energy consumption constraints, according to goals and constraints Determining the segmentation ratio of the perception task data also includes: exist In the case of , if the amount of task data to be processed by the edge device j is less than or equal to B _i , then it is determined that x _i <1, y _i,j <0, and z _i =0; in, C _i,j is the calculation rate at which edge device j calculates the sensing task data obtained by terminal device i; Under energy consumption constraints, according to goals and constraints Determining the segmentation ratio of the perception task data also includes: exist And when the amount of task data to be processed by the edge device j is greater than B _i , according to the formula Calculate _Xi , _Yi , _Zi , and then determine the division ratio based on Xi _{, Yi} , _Zi ; _{Wherein , _} The _data amount of the sensing task data in the center _is , _Xi + Y _i + Z _i = _Li , xi _: y _{i :} z _{i =} total amount; Under energy consumption constraints, according to goals and constraints Determining the segmentation ratio of the perception task data also includes: When the terminal device i is busy, according to the formula Calculate the data volume threshold X _Ti allocated to the terminal device i; and/or, when the edge device j is busy, calculate according to the formula/> Calculate the data volume threshold Y _Ti to be divided to the edge device j; Determine the division ratio according to at least one of X _Ti and Y _Ti and the total amount of the sensing task data; in, is the waiting time required on the terminal device i;/> Indicates the waiting time on the edge device j,/> is the time for calculation on the terminal device i,/> is the transmission rate when the terminal device i _transmits the sensing task data acquired by itself to the edge device mi, C _i,j is the edge device j to calculate the sensing task acquired by the terminal device i The rate at which data is calculated. 2. A device for performing the method of claim 1, characterized in that it includes: The acquisition unit is used to acquire the sensing task data to be processed by the terminal device; A determination unit configured to determine a task division strategy based on reference conditions, the reference conditions including at least one of the following: respective busy conditions, task processing speed conditions, and data transmission rate conditions of the terminal device, edge device, and cloud center; A segmentation unit configured to control the terminal device to segment the sensing task data according to the segmentation strategy; A control unit configured to control at least one of the terminal device, the edge device, and the cloud center to process the segmented data according to the segmentation situation of the sensing task data.