A method for maximizing computing rate by combining edge computing with collaborative intelligent transportation system

Abstract

The present invention discloses a method for maximizing computing rate of edge computing combined with a collaborative intelligent transportation system, wherein the edge computing combined with a collaborative intelligent transportation system includes a base station, a roadbed unit and a wireless device. The method for maximizing computing rate of edge computing combined with a collaborative intelligent transportation system constructs a computing rate expression for each wireless device regarding the offloading decision and the bandwidth allocation of each wireless device to the roadbed unit, and solves the offloading decision and the bandwidth allocation of each wireless device to the roadbed unit to obtain the computing rate of each wireless device, and the computing rate of each wireless device is the maximum computing rate of each wireless device, and the sum of the maximum computing rates of all wireless devices is the maximum computing rate of the edge computing combined with the collaborative intelligent transportation system, thereby significantly reducing system delay and improving computing efficiency.

Description

Technical Field

[0001] The present invention belongs to the field of edge computing technology, and specifically relates to a method for maximizing computing rate by combining edge computing with a collaborative intelligent transportation system. Background Art

[0002] With the widespread application of intelligent transportation systems (ITS) around the world, the intelligence and automation of traffic management are gradually improving. Intelligent transportation systems rely on advanced technologies such as big data, cloud computing, and the Internet of Things (IoT) to achieve functions such as traffic flow prediction, accident detection, and vehicle management. However, with the rapid increase in the number of connected vehicle devices, the demand for data processing is growing exponentially, and traditional cloud computing-based transportation systems face problems such as latency, bandwidth pressure, and computing power bottlenecks. Especially in application scenarios with high real-time requirements, large amounts of data to be processed, and wide distribution, existing technologies can no longer meet the needs of modern intelligent transportation systems.

[0003] Existing related patent technologies mainly focus on data processing and decision optimization through cloud computing. These technologies usually rely on a centralized computing architecture, in which all data needs to be transmitted to a remote server for centralized processing. However, this architecture has significant shortcomings: on the one hand, long-distance data transmission often brings high latency, affecting the efficiency of real-time decision-making; on the other hand, the centralized processing of massive data increases the computing burden of cloud servers, resulting in a decrease in computing speed, affecting the system's response time and processing capabilities. Summary of the invention

[0004] The purpose of the present invention is to solve the problems raised in the background technology and to propose a method for maximizing the computing rate by combining edge computing with a collaborative intelligent transportation system.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] The present invention proposes a method for maximizing computing rate by combining edge computing with a collaborative intelligent transportation system, including:

[0007] The edge computing combined with the collaborative intelligent transportation system includes a base station, Subgrade units and wireless devices;

[0008] Random Generation The channel gain from the base station to each wireless device in each time slot, and the channel gain from each wireless device unit to each road base unit;

[0009] First time slot In the example, each wireless device is randomly initialized as a reinforcement learning policy network;

[0010] Constructing a computational rate expression for each wireless device regarding offloading decisions and bandwidth allocation from each wireless device to the roadbed unit;

[0011] Input the channel gains associated with each wireless device into their respective policy networks to obtain corresponding offloading decisions;

[0012] The channel gain from the base station to each wireless device and the channel gain from each wireless device to each roadbed unit are input into CVX to obtain the optimal bandwidth allocation for each roadbed unit.

[0013] According to the obtained unloading decision of each wireless device and the optimal bandwidth allocation of each roadbed unit, the calculation rate expression of each wireless device is calculated, and the calculation rate of each wireless device is the maximum calculation rate of each wireless device, and the sum of the maximum calculation rates of all wireless devices is the maximum calculation rate of edge computing combined with collaborative intelligent transportation system. ;

[0014] use Update the policy network of each wireless device, and use the updated policy network as the policy network of each wireless device in the next time slot, repeatedly calculate the maximum calculation rate, and then obtain The edge computing of each time slot is combined with the maximized computing rate of the collaborative intelligent transportation system.

[0015] Preferably, the base station transmits radio frequency energy to each wireless device to power each wireless device, and each wireless device communicates with the road-based unit based on the captured radio frequency energy.

[0016] Preferably, the constructing of a calculation rate expression for each wireless device regarding offloading decision and bandwidth allocation from each wireless device to the roadbed unit comprises:

[0017] Taking each wireless device as a leader, construct the calculation rate expression of each leader regarding the offloading decision and the bandwidth allocation of each wireless device to the roadbed unit as follows:

[0018] ;

[0019] in, For the The computation rate of each wireless device, and , For the The wireless devices can calculate locally or offload calculations to the roadbed unit, and , indicating the Each wireless device computes locally, , indicating the The number of cycles required for each wireless device to process one bit of data is calculated by unloading each wireless device to the roadbed unit. For the The energy capture efficiency of each wireless device, is the power of the radio frequency energy sent by the base station, For the The channel gain from a wireless device to the base station, For the The computational efficiency of each wireless device, For the Wireless device to The bandwidth allocation of each roadbed unit, and , For the The transmission power offloaded by each wireless device, For the Wireless device to The channel gain of each roadbed unit, is the power spectral density of the noise of each wireless device, The duration that the base station provides energy to each wireless device.

[0020] Preferably, the step of inputting the channel gain associated with each wireless device into the respective policy network to obtain the corresponding offloading decision includes:

[0021] For each wireless device, the channel gain from the wireless device to the base station and the channel gain from the wireless device to each roadbed unit are input into the corresponding strategy network to obtain an offloading decision for the corresponding wireless device;

[0022] and The set of policy networks of wireless devices is represented as , For the policy network of wireless devices, The set of offloading decisions of wireless devices is expressed as , For the offload decision for each wireless device.

[0023] Preferably, the uninstall decision satisfies: or ;

[0024] when When The offloading decision of each wireless device is calculated locally. , and at this time the calculation rate of each wireless device as a leader is: ;

[0025] when When The offloading decision for the wireless device is to offload to the The calculation of roadbed units is ;

[0026] At this time, the calculation rate of each wireless device as a leader is expressed as: Convert the calculation rate formula of each wireless device as a leader to the calculation rate expression of each roadbed unit as a follower:

[0027] ;

[0028] in, For the The calculation rate of each roadbed unit, For the The offloading decision for the wireless device is to offload to the The indicator of a roadbed unit, according to the above formula, About The perspective function of is a convex function, solved by CVX .

[0029] Preferably, the solution is performed by CVX The process is as follows:

[0030] For The base station to each wireless device channel gain, and each wireless device to the first The channel gain of each road base unit is input into CVX to obtain the channel gain of each wireless device to the first The optimal bandwidth allocation for each roadbed unit is expressed as ,in Indicates Wireless device to Optimal bandwidth allocation for each roadbed unit;

[0031] Then the optimal bandwidth allocation for all roadbed units is .

[0032] Preferably, the calculation rate expression of each wireless device is calculated according to the obtained unloading decision of each wireless device and the optimal bandwidth allocation of each roadbed unit, and the calculation rate of each wireless device is the maximum calculation rate of each wireless device, and the sum of the maximum calculation rates of all wireless devices is the maximum calculation rate of edge computing combined with collaborative intelligent transportation system. , the formula is as follows:

[0033] .

[0034] Preferably, the use Update the policy network of each wireless device, the formula is as follows:

[0035] ;

[0036] in, For the updated policy network of wireless devices, Indicates The learning rate of the policy network of each wireless device, For the The policy network function of each wireless device, For the The output of the policy network for each wireless device, i.e., the offloading decision, From base station to The channel gain of the wireless device, and The channel gain from each wireless device to each roadbase unit.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] This method for maximizing the computing rate of edge computing combined with a collaborative intelligent transportation system constructs a computing rate expression for each wireless device regarding the offloading decision and the bandwidth allocation of each wireless device to the roadbed unit, and solves the offloading decision and the bandwidth allocation of each wireless device to the roadbed unit to obtain the computing rate of each wireless device. The computing rate of each wireless device is the maximum computing rate of each wireless device, and the sum of the maximum computing rates of all wireless devices is the maximum computing rate of the edge computing combined with the collaborative intelligent transportation system, thereby significantly reducing system latency and improving computing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a flow chart of the method for maximizing computing rate by combining edge computing with a collaborative intelligent transportation system according to the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] like Figure 1 As shown, a method for maximizing computing rate by combining edge computing with a collaborative intelligent transportation system is provided, including:

[0042] Step 1: Edge computing combined with collaborative intelligent transportation system includes a base station, Subgrade units and wireless devices; the base station sends radio frequency energy to each wireless device to power each wireless device, and each wireless device communicates with the roadbed unit based on the captured radio frequency energy (unloading data to the roadbed unit based on the captured energy). In this embodiment, if the number of roadbed units is three, the number of wireless devices is fifteen.

[0043] Step 2: Random Generation The channel gain from the base station to each wireless device in each time slot, and the channel gain from each wireless device unit to each road base unit;

[0044] Step 3: First time slot Inside (including time slots, and the length of each time slot is ), randomly initialize each wireless device as a reinforcement learning policy network;

[0045] Step 4: Construct a calculation rate expression for each wireless device regarding the offloading decision and the bandwidth allocation from each wireless device to the roadbed unit, including:

[0046] Taking each wireless device as a leader, construct the calculation rate expression of each leader regarding the offloading decision and the bandwidth allocation of each wireless device to the roadbed unit as follows:

[0047] ;

[0048] in, For the The computation rate of each wireless device, and , For the The wireless devices can calculate locally or offload calculations to the roadbed unit, and , indicating the Each wireless device computes locally, , indicating the The number of cycles required for each wireless device to process one bit of data is calculated by unloading each wireless device to the roadbed unit. For the The energy capture efficiency of each wireless device, is the power of the radio frequency energy sent by the base station, For the The channel gain from a wireless device to the base station, For the The computational efficiency of each wireless device, For the Wireless device to The bandwidth allocation of each roadbed unit, and , For the The transmission power offloaded by each wireless device, For the Wireless device to The channel gain of each roadbed unit, is the power spectral density of the noise of each wireless device, The duration that the base station provides energy to each wireless device.

[0049] Step 5: Input the channel gain related to each wireless device into the respective policy network to obtain the corresponding offloading decision, which is specifically:

[0050] For each wireless device, the channel gain from the wireless device to the base station and the channel gain from the wireless device to each roadbed unit are input into the corresponding strategy network to obtain an offloading decision for the corresponding wireless device;

[0051] and The set of policy networks of wireless devices is represented as , For the policy network of wireless devices, The set of offloading decisions of wireless devices is expressed as , For the offload decision for each wireless device.

[0052] The uninstall decision satisfies: or ;

[0053] when When The offloading decision of each wireless device is calculated locally. , and at this time the calculation rate of each wireless device as a leader is: To determine the value;

[0054] when When (in this embodiment, , or ), indicating the The offloading decision for the wireless device is to offload to the The calculation of roadbed units is ;

[0055] At this time, the calculation rate of each wireless device as a leader is expressed as: Convert the calculation rate formula of each wireless device as a leader to the calculation rate expression of each roadbed unit as a follower:

[0056] (1);

[0057] in, For the The calculation rate of each roadbed unit, For the The offloading decision for the wireless device is to offload to the The indicator of a roadbed unit, according to formula (1), About The perspective function of is a convex function, solved by CVX It should be noted that by taking the wireless device as the leader and the roadbed unit as the follower, the Stackelberg game algorithm is introduced, so that the offloading decision of each wireless device and the optimal bandwidth allocation of each roadbed unit will reach the Stackelberg equilibrium.

[0058] Step 6: Solve using CVX The process is as follows: the channel gain from the base station to each wireless device and the channel gain from each wireless device to each roadbed unit are input into CVX (CVX, Convex Optimization, is a toolbox for modeling and solving convex optimization problems) to obtain the optimal bandwidth allocation for each roadbed unit;

[0059] Among them, for the The base station to each wireless device channel gain, and each wireless device to the first The channel gain of each road base unit is input into CVX to obtain the channel gain of each wireless device to the first The optimal bandwidth allocation for each roadbed unit is expressed as ,in Indicates Wireless device to Optimal bandwidth allocation for each roadbed unit;

[0060] Then the optimal bandwidth allocation for all roadbed units is , Indicates that each wireless device The optimal bandwidth allocation for each roadbed unit.

[0061] Step 7: Calculate the computing rate expression of each wireless device according to the obtained unloading decision of each wireless device and the optimal bandwidth allocation of each roadbed unit, and the computing rate of each wireless device is the maximum computing rate of each wireless device, and the sum of the maximum computing rates of all wireless devices is the maximum computing rate of edge computing combined with collaborative intelligent transportation system. , the specific formula is (the unloading decision of each wireless device and the optimal bandwidth allocation of each roadbed unit obtained by proving will reach the Stackelberg equilibrium, that is, the computing rate of each wireless device calculated is the maximum computing rate of each wireless device, and then the sum of the maximum computing rates of all wireless devices is the maximum computing rate of edge computing combined with the collaborative intelligent transportation system, and the specific proof process belongs to the prior art and is not described in detail, such as the existing literature Xia S, Yao Z, Wu G, et al. Distributed offloading for cooperative intelligent transportation under heterogeneous networks[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(9): 16701-16714. and the existing literature Gao Z, Yang L, Dai Y. Large-scale computation offloading using a multi-agent reinforcement learning in heterogeneous multi-access edge computing[J]. IEEE Transactions on Mobile Computing, 2022, 22(6): 3425-3443. There are specific proof processes):

[0062] .

[0063] Step 8. Utilization Update the policy network of each wireless device, and use the updated policy network as the policy network of each wireless device in the next time slot, repeatedly calculate the maximum calculation rate, and then obtain Edge computing in each time slot is combined with the maximum computing rate of the collaborative intelligent transportation system;

[0064] Among them, the use Update the policy network of each wireless device, the formula is as follows:

[0065] ;

[0066] in, For the updated policy network of wireless devices, Indicates The learning rate of the policy network of each wireless device, For the The policy network function of each wireless device, For the The output of the policy network for each wireless device, i.e., the offloading decision, From base station to The channel gain of the wireless device, and The channel gain from each wireless device to each road base unit;

[0067] The updated strategy network is used as the strategy network of each wireless device in the next time slot, and then steps 5 to 8 are repeated to calculate the maximum calculation rate in each time slot, thereby obtaining The edge computing of each time slot is combined with the maximized computing rate of the collaborative intelligent transportation system.

[0068] This method optimizes the data exchange between edge nodes and the central cloud platform while ensuring the maximum computing rate through intelligent decision scheduling and dynamic bandwidth resource allocation, significantly reducing system latency and improving computing efficiency. Specifically, by reasonably allocating computing tasks to roadbed units, the data transmission distance is reduced, and faster response and efficient real-time decision-making are achieved. The Stackelberg game algorithm is used to dynamically adjust the allocation strategy of computing resources according to the real-time load and network conditions of the traffic system to ensure that the computing rate can reach the optimal value under different load conditions. Different wireless devices can work together, and through data sharing and intelligent collaboration, the efficiency of wireless device computing is improved, and computing latency and data redundancy are reduced. By comprehensively considering the urgency of computing tasks and the capabilities of wireless devices, the use of computing resources is optimized, excessive consumption or idleness of computing resources is avoided, and the overall performance of the system is improved.

[0069] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0070] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for maximizing computing rate by combining edge computing with a collaborative intelligent transportation system, characterized in that: The edge computing combined with the collaborative intelligent transportation system to maximize the computing rate method includes: The edge computing combined with the collaborative intelligent transportation system includes a base station, Subgrade units and wireless devices; Random Generation The channel gain from the base station to each wireless device in each time slot, and the channel gain from each wireless device unit to each road base unit; First time slot In the example, each wireless device is randomly initialized as a reinforcement learning policy network; Constructing a computational rate expression for each wireless device regarding offloading decisions and bandwidth allocation from each wireless device to the roadbed unit; Input the channel gains associated with each wireless device into their respective policy networks to obtain corresponding offloading decisions; The channel gain from the base station to each wireless device and the channel gain from each wireless device to each roadbed unit are input into CVX to obtain the optimal bandwidth allocation for each roadbed unit. According to the obtained unloading decision of each wireless device and the optimal bandwidth allocation of each roadbed unit, the calculation rate expression of each wireless device is calculated, and the calculation rate of each wireless device is the maximum calculation rate of each wireless device, and the sum of the maximum calculation rates of all wireless devices is the maximum calculation rate of edge computing combined with collaborative intelligent transportation system. ; use Update the policy network of each wireless device, and use the updated policy network as the policy network of each wireless device in the next time slot, repeatedly calculate the maximum calculation rate, and then obtain Edge computing in each time slot is combined with the maximum computing rate of the collaborative intelligent transportation system; The step of constructing a calculation rate expression for each wireless device regarding offloading decisions and bandwidth allocation from each wireless device to a roadbed unit includes: Taking each wireless device as a leader, construct the calculation rate expression of each leader regarding the offloading decision and the bandwidth allocation of each wireless device to the roadbed unit as follows: ； in, For the The computation rate of each wireless device, and , For the The wireless devices can calculate locally or offload calculations to the roadbed unit, and , indicating the Each wireless device computes locally, , indicating the The number of cycles required for each wireless device to process one bit of data is calculated by unloading each wireless device to the roadbed unit. For the The energy capture efficiency of each wireless device, is the power of the radio frequency energy sent by the base station, For the The channel gain from the wireless device to the base station, For the The computational efficiency of each wireless device, For the Wireless device to The bandwidth allocation of each roadbed unit, and , For the The transmission power offloaded by each wireless device, For the Wireless device to The channel gain of each roadbed unit, is the power spectral density of the noise of each wireless device, The duration for which the base station provides energy to each wireless device; The uninstall decision satisfies: or ; when When The offloading decision of each wireless device is calculated locally. , and at this time the calculation rate of each wireless device as a leader is: ; when When The offloading decision for the wireless device is to offload to the The calculation of roadbed units is ; At this time, the calculation rate of each wireless device as a leader is expressed as: Convert the calculation rate formula of each wireless device as a leader to the calculation rate expression of each roadbed unit as a follower: ； in, For the The calculation rate of each roadbed unit, For the The offloading decision for the wireless device is to offload to the The indicator of a roadbed unit, according to the above formula, About The perspective function of is a convex function, solved by CVX ; The utilization Update the policy network of each wireless device, the formula is as follows: ； in, For the updated policy network of wireless devices, Indicates The learning rate of the policy network of each wireless device, For the The policy network function of each wireless device, For the The output of the policy network for each wireless device, i.e., the offloading decision, From base station to The channel gain of the wireless device, and The channel gain from each wireless device to each roadbase unit.

2. The edge computing combined with the collaborative intelligent transportation system maximization computing rate method as claimed in claim 1 is characterized by: The base station transmits radio frequency energy to each wireless device to power each wireless device, and each wireless device communicates with the road-based unit based on the captured radio frequency energy.

3. The edge computing combined with the collaborative intelligent transportation system maximization computing rate method as claimed in claim 1 is characterized by: The inputting the channel gain related to each wireless device into the respective policy network to obtain the corresponding offloading decision includes: For each wireless device, the channel gain from the wireless device to the base station and the channel gain from the wireless device to each roadbed unit are input into the corresponding strategy network to obtain an offloading decision for the corresponding wireless device; and The set of policy networks of wireless devices is represented as , For the policy network of wireless devices, The set of offloading decisions of wireless devices is expressed as , For the offload decision for each wireless device.

4. The edge computing combined with the collaborative intelligent transportation system maximization computing rate method as claimed in claim 1 is characterized by: The solution is obtained by CVX The process is as follows: For The base station to each wireless device channel gain, and each wireless device to the first The channel gain of each road base unit is input into CVX to obtain the channel gain of each wireless device to the first The optimal bandwidth allocation for each roadbed unit is expressed as ,in Indicates Wireless device to Optimal bandwidth allocation for each roadbed unit; Then the optimal bandwidth allocation for all roadbed units is .

5. The edge computing combined with the collaborative intelligent transportation system maximization computing rate method as claimed in claim 4 is characterized by: According to the obtained unloading decision of each wireless device and the optimal bandwidth allocation of each roadbed unit, the calculation rate expression of each wireless device is calculated, and the calculation rate of each wireless device is the maximum calculation rate of each wireless device, and the sum of the maximum calculation rates of all wireless devices is the maximum calculation rate of the edge computing combined with the collaborative intelligent transportation system. , the formula is as follows: 。

Images