Energy-saving task offloading method and system for electric intelligent connected vehicles

Abstract

The present invention relates to the field of vehicle networking technology, and discloses an energy-saving task unloading method and system for electric intelligent networked vehicles. By placing a server with powerful computing power in a roadside unit not far from the vehicle, computing services are provided for vehicles to and from, reducing vehicle load, saving vehicle energy consumption, and shortening task delays; tasks are transmitted between roadside units through optical fibers and results are returned, thereby improving server load balancing, accelerating task processing speed, and improving computing resource utilization efficiency. The average task processing delay of the task unloading method of the present invention is reduced by up to 20% and 45% respectively compared with partial unloading to the nearest MEC server and full unloading to the nearest MEC server, the average energy consumption of the task is reduced by up to 14% compared with PONM, and the task completion rate is increased by up to 24% and 75% compared with PONM and AONM, effectively utilizing idle computing resources, alleviating the problem of server load imbalance, reducing task delay and vehicle energy consumption, and the unloading method has superior performance.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle networking, and in particular to a method and system for unloading energy-saving tasks of an electric intelligent networked vehicle. Background Art

[0002] The rapid development of 5G and battery technology has made electric intelligent connected vehicles the focus of the current automotive industry. The soaring price of fossil fuels has led to a sharp increase in demand for electric vehicles. However, the rapid development of technologies and services such as autonomous driving and vehicle networking has significantly increased the computing workload of various on-board tasks. Cloud servers are far away from vehicles and cannot complete user tasks in a timely manner. Electric vehicles have limited energy and computing resources, so the cost of processing tasks is high and timeliness cannot be guaranteed.

[0003] Multi-access edge computing technology (MEC) deploys servers at the edge of the network close to users and offloads tasks to nearby MEC servers, which can significantly reduce transmission latency. At the same time, it uses the computing resources of the server to process tasks and reduce local computing load. In addition, for vehicles, the energy consumed in the task offloading process is much lower than the energy consumed by local processing of the same amount of data, so using MEC servers for processing can reduce the energy consumption of vehicles. However, existing research focuses on offloading tasks to the MEC server closest to the vehicle. Due to the uneven temporal and spatial distribution of vehicles, this offloading scheme will cause some MEC servers to be overloaded, while others will be idle. The unbalanced server load reduces the efficiency of computing resource utilization, and server overload causes tasks to be unable to be processed in a timely manner. In addition, existing research does not target electric connected vehicles, so the proposed offloading scheme does not consider the problem of limited battery capacity of electric vehicles. Summary of the invention

[0004] In order to overcome the defects of the above-mentioned prior art, the purpose of the present invention is to provide an energy-saving task offloading method and system for electric intelligent connected vehicles, so as to solve the technical problems in the prior art such as unbalanced server load reducing the efficiency of computing resource utilization and server overload causing the task to be unable to be processed in time.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for unloading energy-saving tasks of an electric intelligent network-connected vehicle comprises the following steps:

[0007] Step 1: Receive vehicle task requests from vehicles within its coverage area during the simulation time and send them to the macro base station. The macro base station sorts out the vehicle set and task set based on all vehicle request information during the time.

[0008] In step 2, the macro base station starts to solve the task offloading strategy based on the information of the vehicle set and the task set to obtain the final offloading plan, and sends the final offloading plan to the roadside unit. The roadside unit sends the unloading plan of the corresponding vehicle to the vehicles within its coverage area to realize the energy-saving task offloading of the electric intelligent connected vehicle.

[0009] Preferably, in step 2, the macro base station simulates the actual situation based on the information of the vehicle set and the task set, and processes the vehicle tasks locally and on the server respectively in the simulation of the actual situation. The specific process is as follows:

[0010] The vehicle tasks are evenly divided into several blocks. According to the final unloading plan given by the macro base station, part of the vehicle tasks are unloaded to the roadside unit closest to the vehicle, or unloaded to the nearest roadside unit and then transmitted to the next roadside unit in the driving direction through the optical fiber network unit for processing. The remaining vehicle tasks are processed locally. After the nearest roadside unit receives the vehicle tasks, it waits for the server to execute them in the task queue of the MEC server. After the server completes the processing, it sends the results to the vehicle. If the roadside unit executing the task cannot process it, packet loss processing is performed and an additional penalty factor is added.

[0011] Preferably, in step 2, the macro base station starts to solve the task offloading strategy according to the information of the vehicle set and the task set to obtain the final offloading solution, wherein the process is as follows:

[0012] S1, set the particle dimension to 2D, and initialize the particle swarm and parameter settings, N = 1; where D is the number of tasks in the task set;

[0013] S2, assign the first D items of the particle dimension to each task in the task set The last D item is assigned to each task in the task set.

[0014] S3, providing a vehicle set, a task set and a simulation time for processing, and calculating the task delay and vehicle energy consumption based on the processing results to obtain an adaptation value;

[0015] S4, update the individual optimal and global optimal values ​​of all particles;

[0016] S5, determine whether iteration is still needed; if the task still needs iteration, update the speed and position of each particle; add 1 to the current iteration number; otherwise, repeat S1 to S4 until all task iterations are completed, and output the fitness value and unloading plan corresponding to the global optimal value.

[0017] Furthermore, the latency in S3 includes task local processing time, wireless communication time, server queue waiting time, and server processing time; vehicle energy consumption includes task local energy consumption and wireless communication energy consumption for task uploading;

[0018] The calculation formula for task local processing time is as follows:

[0019]

[0020] in, is the local execution ratio of vehicle tasks; The amount of calculation required to complete the task; k is the vehicle; t is the time; I t,k is the vehicle task generated by vehicle k at time t; is the computing power of vehicle k;

[0021] The calculation formula of task local energy consumption is as follows:

[0022]

[0023] Where κ is the energy density;

[0024] The calculation formula for wireless communication time is as follows:

[0025]

[0026] in, Calculate the data as needed; is the local execution ratio of vehicle tasks; is the wireless communication transmission rate;

[0027] The calculation formula for wireless communication energy consumption of task upload is as follows:

[0028]

[0029] in, represents the transmission power of vehicle k; Calculate the data as needed; is the local execution ratio of vehicle tasks; is the wireless communication transmission rate; Wireless communication time;

[0030] The server processing time is calculated as follows:

[0031]

[0032] in, represents the computing power of the server closest to vehicle k; is the local execution ratio of vehicle tasks; The amount of computation required to complete the task;

[0033] The calculation formula for server queue waiting time is:

[0034]

[0035] in Indicates that the server is The amount of remaining tasks in the server queue at this moment, t is I t,k is the generation time of , and δ is the offloading solution given by the macro base station.

[0036] Furthermore, the remaining tasks in the server queue are calculated as follows:

[0037]

[0038] Where ω is the set of all tasks.

[0039] Preferably, after the vehicle task is generated, it is first processed locally. If the local execution part of the vehicle task cannot be completed within the delay threshold, the task is directly discarded and the evaluation index penalty factor is assigned. Instead, local processing begins while the vehicle offloads the remainder of the task to the nearest roadside unit via wireless communication.

[0040] Furthermore, the roadside unit distributes the task to the vehicles within its range according to the offloading scheme determined by the macro base station. The vehicle offloads the task to the nearest roadside unit according to the offloading scheme or offloads it to the nearest roadside unit and transmits it to the next roadside unit in the driving direction through optical fiber for processing. After receiving the vehicle task, the roadside unit places the task in the task queue of the MEC server and waits for the server to execute it. The MEC server calculates the time to complete the task processing. If the part can be completed within the delay threshold, it is processed. Otherwise, the task is directly discarded and a penalty factor is added to the evaluation index. At the same time, the penalty factor needs to consider the energy consumption generated by local processing and task upload time. In the case of task discard, the energy consumption generated by the vehicle for the task can be expressed as:

[0041]

[0042] in, is the total energy consumption of task processing.

[0043] Preferably, the vehicle tasks include video tasks, navigation tasks and intelligent traffic dispatching tasks, wherein each vehicle task has a different delay threshold and task size.

[0044] An energy-saving task unloading system for an electric intelligent network-connected vehicle, used to implement the energy-saving task unloading method for an electric intelligent network-connected vehicle described above, comprising:

[0045] The first processing module is used to receive vehicle task requests from vehicles within its coverage area during the simulation time and send them to the macro base station. The macro base station sorts out the vehicle set and task set according to all vehicle request information within the time;

[0046] The second processing module is used for the macro base station to start solving the task unloading strategy according to the information of the vehicle set and the task set to obtain the final unloading plan, and send the final unloading plan to the roadside unit. The roadside unit sends the unloading plan of the corresponding vehicle to the vehicles within its coverage area, thereby realizing the energy-saving task unloading of the electric intelligent connected vehicle.

[0047] Preferably, a macro base station is set up on the side of the road, and several roadside units are evenly arranged along the extension direction of the roadside; several roadside units are communicatively connected to vehicles, and the macro base station is communicatively connected to several roadside units; the roadside unit is provided with a communication base station, an optical fiber network unit and an MEC server; the communication base station is communicatively connected to the optical fiber network unit and the MEC server respectively, the roadside unit is communicatively connected to other roadside units through the optical fiber network unit, and the roadside unit is communicatively connected to several vehicles through the communication base station.

[0048] Compared with the prior art, the present invention has the following beneficial technical effects:

[0049] The present invention provides an energy-saving task unloading method for an electric intelligent networked vehicle, which unloads tasks to a roadside unit, shortens task delays and reduces vehicle energy consumption. The roadside unit transmits tasks and processing results through optical fibers, alleviates the problem of unbalanced server loads, improves computing resource utilization, and further shortens task delays. A packet loss penalty mechanism is provided to discard unfinished tasks in advance, thereby avoiding waste of computing resources and improving task completion rates. Finally, a discrete particle swarm algorithm is used to give a near-optimal unloading solution in a very short time; the average task processing delay of the task unloading method of the present invention is reduced by up to 20% and 45% respectively compared with partial unloading to the nearest MEC server (PONM) and full unloading to the nearest MEC server (AONM), the average energy consumption of the task is reduced by up to 14% compared with PONM, and the task completion rate is increased by up to 24% and 75% compared with PONM and AONM. ​​In the high-efficiency and energy-saving task unloading method, idle computing resources are effectively utilized, the problem of unbalanced server loads is alleviated, task delays and vehicle energy consumption are reduced, and the unloading method has superior performance.

[0050] The present invention also provides an energy-saving task unloading system for electric intelligent connected vehicles, which places an MEC server in a roadside unit not far from the vehicle, communicates with the optical fiber network unit and the MEC server respectively through a communication base station, the roadside unit communicates with other roadside units through the optical fiber network unit, and the roadside unit communicates with several vehicles through the communication base station to provide computing services for the vehicles to and from, thereby reducing vehicle load, saving vehicle energy consumption, and shortening task delays; tasks are transmitted and results are returned between roadside units through optical fiber, thereby improving server load balancing, accelerating task processing, and improving computing resource utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flow chart of the energy-saving task unloading method of the electric intelligent network-connected vehicle in the present invention;

[0052] Figure 2 A distribution diagram of an energy-saving task unloading system of an electric intelligent network-connected vehicle in the present invention;

[0053] Figure 3 The energy-saving task unloading system model and simulation process diagram of the electric intelligent network-connected vehicle in the present invention;

[0054] Figure 4 This is a process diagram of the task packet loss penalty mechanism in the present invention;

[0055] Figure 5 A discrete particle swarm algorithm process diagram of the task offloading method described in the present invention;

[0056] Figure 6 is a process diagram of a solution strategy of the task offloading method of the present invention;

[0057] Figure 7 A comparison chart of the average task energy consumption of the present invention, AONM and PONM;

[0058] Figure 8 A comparison chart of the average processing delay between the present invention and AONM and PONM;

[0059] Fig. 9 A comparison chart of the average task completion rate of the present invention, AONM and PONM;

[0060] Fig.10 A comparison chart of the average task offloading ratio between the present invention and PONM;

[0061] Fig.11 The figure is a comparison chart of the average calculation rate of the present invention, AONM and PONM.

[0062] In the figure: 1-roadside unit; 2-vehicle; 3-macro base station; 11-communication base station; 12-optical network unit; 13-MEC server. DETAILED DESCRIPTION

[0063] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0064] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0065] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0066] The purpose of the present invention is to provide an energy-saving task unloading method and system for electric intelligent connected vehicles to solve the technical problems in the prior art that the server load is unbalanced, which reduces the efficiency of computing resource utilization, and the server is overloaded, resulting in the inability to process tasks in a timely manner.

[0067] according to Figure 1 As shown, the present invention provides a method for unloading energy-saving tasks of an electric intelligent network-connected vehicle, comprising the following steps:

[0068] Step 1, receiving a vehicle task request from a vehicle 2 within its coverage area during the simulation time, and sending it to the macro base station 3, the macro base station 3 collates all vehicle request information within the time to obtain a vehicle set and a task set;

[0069] In step 2, the macro base station 3 starts to solve the task unloading strategy according to the information of the vehicle set and the task set to obtain the final unloading plan, and sends the final unloading plan to the roadside unit 1. The roadside unit 1 sends the unloading plan of the corresponding vehicle to the vehicle 2 within its coverage area, thereby realizing the energy-saving task unloading of the electric intelligent connected vehicle.

[0070] Specifically, in step 2, the macro base station 3 simulates the actual situation based on the information of the vehicle set and the task set, and processes the vehicle tasks locally and on the server respectively in the simulation of the actual situation. The specific process is as follows:

[0071] Specifically, according to Figure 3 As shown, vehicle tasks include video tasks, navigation tasks, and intelligent traffic dispatching tasks, each of which has different delay thresholds and task sizes. The vehicle tasks are evenly divided into several blocks. According to the final unloading scheme given by the macro base station 3, part of the vehicle tasks are unloaded to the roadside unit 1 closest to the vehicle, or unloaded to the nearest roadside unit 1 and then transmitted to the next roadside unit 1 in the driving direction through the optical network unit 12 for processing. The remaining vehicle tasks are processed locally. If the nearest roadside unit cannot process it in time, the task is sent to other roadside units in the direction of the vehicle through the optical network unit of the roadside unit. The task is then sent by the current roadside unit to the task queue of the MEC server 13 to wait for the server to execute. After the server completes the processing, the result is sent to the vehicle. If the roadside unit 1 executing the task cannot process it, packet loss processing is performed and an additional penalty factor is added. In the process of sending the task processing result to the original vehicle, since the size of the task result is much smaller than the size when the task is uploaded, the delay of this part is negligible. Since the transmission speed of the optical fiber link is much faster than the wireless communication speed, this part of time can be ignored. Therefore, in the whole simulation process, the delay includes the time generated by the local task processing time, task upload time, server queue waiting time and server processing time. In addition, the server processing task does not consume the energy of the vehicle, so it is not calculated in the vehicle energy consumption. Therefore, the vehicle energy consumption includes the local task energy consumption and the wireless communication energy consumption of task upload.

[0072] Specifically, according to Figure 5 As shown, in step 2, the macro base station 3 starts to solve the task offloading strategy according to the information of the vehicle set and the task set to obtain the final offloading solution, wherein the process is as follows:

[0073] S1, set the particle dimension to 2D, and initialize the particle swarm and parameter settings, N = 1; where D is the number of tasks in the task set;

[0074] S2, assign the first D items of the particle dimension to each task in the task set The last D item is assigned to each task in the task set.

[0075] S3, provides vehicle collection and task collection and simulation time, and calculate the task delay and vehicle energy consumption based on the processing results to obtain the adaptation value;

[0076] S4, update the individual optimal and global optimal values ​​of all particles;

[0077] S5, determine whether iteration is still needed; if the task still needs iteration, update the speed and position of each particle; add 1 to the current iteration number; otherwise, repeat S1 to S4 until all task iterations are completed, and output the fitness value and unloading plan corresponding to the global optimal value.

[0078] Among them, in S3, the vehicle set, task set and simulation time are provided for processing, according to Figure 6 As shown, the processing steps are as follows:

[0079] S301: Input vehicle set Task Collection and end time T1; set T=1;

[0080] S302, initializing the current time;

[0081] S303, sort all tasks by task arrival time and store them in queue Q;

[0082] S304, determine whether the arrival time of the head task in queue Q is equal to the current time;

[0083] S305: When the task arrival time is not equal to the current time, execute step S315;

[0084] S306: When the task arrival time is equal to the current time, task I t,k Take it from the head of the queue and start processing; S307, determine the local execution ratio of the task Is it not equal to 0?

[0085] S308: When the local execution ratio of the task When it is equal to 0, go to step S313;

[0086] S309: When the local execution ratio of the task When it is not equal to 0, judge I t,k Whether the local part can be completed within the delay threshold;

[0087] S310, when I t,k When the local part can be completed within the delay threshold, task I t,k The local part is left to local processing;

[0088] S311, determine whether the task has only a local part, that is Is it equal to 1?

[0089] S312, when When it is equal to 1, the process goes to step S304;

[0090] S313, when When not equal to 1, task I t,k middle The part is sent to the nearest roadside unit to start task unloading; the wireless communication end time is calculated and the task is placed in the communication queue W; and step S304 is executed;

[0091] S314: When the task cannot be completed within the delay threshold, the task is discarded. t,k ; Go to step S304;

[0092] S315, judging whether the communication of the head task in the communication queue has ended, that is, whether the communication end time of the head of the W queue is equal to T;

[0093] S316: When the communication end time of the W squadron head is not equal to T, go to step S324;

[0094] S317, when the communication end time of the squadron head is equal to T, take task n out from the squadron head W;

[0095] S318, determine task n Is it equal to 1?

[0096] S319, when task n If it is not equal to 1, the process goes to step S321;

[0097] S320, when task n If it is equal to 1, task n is transmitted to the next roadside unit in the direction of the vehicle through optical fiber;

[0098] S321, determining whether task n can be processed and completed by the server within the delay threshold;

[0099] S322, when the task can be processed by the server within the delay threshold, the task is placed in the server queue of the current roadside unit; and step S315 is executed;

[0100] S323, when the task cannot be processed by the server within the delay threshold, the task n is discarded; and step S315 is executed instead;

[0101] S324, setting the serial number of the server to 1;

[0102] S325, determining whether the i-th server is idle;

[0103] S326: When the i-th server is idle, execute step S330;

[0104] S327, when the i-th server is not idle, determine whether the task on the server has reached the end time;

[0105] S328: When the task on the server has not reached the end time, execute step S333;

[0106] S329, when the task on the server reaches the end time, the task is completed and exits from the server;

[0107] S330, determining whether the queue of the i-th server is empty;

[0108] S331: When the queue of the i-th server is empty, execute step S333;

[0109] S332: When the queue of the i-th server is not empty, take the head of the server queue for processing and calculate the end time;

[0110] S333, i=i+1;

[0111] S334, determine whether all servers have been traversed, that is, i<=r;

[0112] S335, when i is less than or equal to r, go to step S325;

[0113] S336, when i is greater than r, update the vehicle position according to the vehicle speed, T=T+1;

[0114] S337, judging whether the simulation end time has been reached, that is, T is less than or equal to T1;

[0115] S338, when T is less than or equal to T1, go to step S;

[0116] S339: When it is greater than T1, output the processed task set.

[0117] Among them, the delay in S3 includes the task local processing time, wireless communication time, server queue waiting time and server processing time; the vehicle energy consumption includes the task local energy consumption and the wireless communication energy consumption of task uploading;

[0118] Assume that the road in the present invention is a two-way four-lane straight road, the number of roadside units 1 is r; the number of vehicles 2 is k; and the task generated by vehicle k at time t is I t,k , I t,k The mathematical expression of is: in is the size of the data to be calculated, The amount of computation required to complete the task, is the delay threshold of the task, and ω represents the set of all tasks within a time interval. The evaluation index of this method is composed of the energy consumed by the vehicle when processing the task, the total delay when the task is successfully completed, and and the penalty factor when the task fails Composition, where the penalty factor is defined as a large number, which is proportional to the data size of the task Directly proportional.

[0119] The computing capacity of the MEC server 13 is recorded as:

[0120] Mission I t,k The local execution ratio is recorded as The value range of is:

[0121]

[0122] The calculation formula for task local processing time is as follows:

[0123]

[0124] in, is the local execution ratio of vehicle tasks; The amount of calculation required to complete the task; k is the vehicle; t is the time; I t,k is the vehicle task generated by vehicle k at time t; is the computing power of vehicle k;

[0125] The calculation formula of task local energy consumption is as follows:

[0126]

[0127] Where κ is the energy density;

[0128] Wireless communication transmission rate in the present invention The mathematical expression is:

[0129]

[0130] in, represents the channel bandwidth between vehicle k and the sth roadside unit closest to k; represents the transmission power of vehicle k; represents the channel gain between vehicle k and roadside unit s; σ represents the urban noise power.

[0131] The calculation formula for wireless communication time is as follows:

[0132]

[0133] in, Calculate the data as needed; is the local execution ratio of vehicle tasks; is the wireless communication transmission rate;

[0134] The calculation formula for wireless communication energy consumption of task upload is as follows:

[0135]

[0136] in, represents the transmission power of vehicle k; Calculate the data as needed; is the local execution ratio of vehicle tasks; is the wireless communication transmission rate; Wireless communication time;

[0137] The server processing time is calculated as follows:

[0138]

[0139] in, represents the computing power of the server closest to vehicle k; is the local execution ratio of vehicle tasks; The amount of computation required to complete the task;

[0140] Mission I t,k The unloading decision variable is recorded as The mathematical expression is:

[0141]

[0142] Among them When , it means the task is unloaded to the nearest roadside unit. When , it indicates that the task is unloaded to the next roadside unit in the vehicle's travel direction.

[0143] The calculation formula for server queue waiting time is:

[0144]

[0145] in Indicates that the server is The amount of remaining tasks in the server queue at this moment, t is I t,k is the generation time of , and δ is the offloading solution given by the macro base station.

[0146] Calculation of remaining tasks in the server queue:

[0147] Where ω is the set of all tasks

[0148] Therefore, the total time of the task offloading process is:

[0149]

[0150] in Indicates the waiting time in the server queue of roadside unit s.

[0151] The total time for task processing is:

[0152]

[0153] The total energy consumption of task processing is:

[0154]

[0155] in, Wireless communication energy consumption for task upload; is the local energy consumption of the task.

[0156] Specifically, according to Figure 4 As shown in the figure, the present invention provides a packet loss penalty mechanism for the situation where the task cannot be processed within the delay threshold. The task is discarded in advance and an evaluation index penalty factor is assigned to improve the task completion rate while saving computing resources. After the vehicle task is generated, local processing judgment is first performed. If the local execution part of the vehicle task cannot be completed within the delay threshold, the task is directly discarded and an evaluation index penalty factor is assigned. On the contrary, local processing starts and the vehicle offloads the remaining part of the task to the nearest roadside unit 1 via wireless communication transmission.

[0157] Among them, the roadside unit 1 unloads to the nearest roadside unit 1 according to the unloading scheme determined by the macro base station 3, or unloads to the nearest roadside unit 1 and distributes the vehicle to the next roadside unit 1 in the driving direction through optical fiber transmission. After receiving the vehicle task, the nearest roadside unit 1 waits for the server to execute in the task queue of the MEC server 13, wherein the MEC server 13 calculates the time for the task to be processed. If this part can be completed within the delay threshold, it is processed, otherwise the task is directly discarded, and a penalty factor is added to the evaluation index. At the same time, the penalty factor needs to consider the energy consumption generated by local processing and task upload time. In the case of task discard, the energy consumption generated by the vehicle for the task can be expressed as:

[0158]

[0159] When a task is discarded locally, the entire task has not yet been processed, so discarding does not consume energy. When a task is discarded on the server, the task has already been processed locally and wirelessly, so discarding still needs to consider the energy consumption generated by local processing and wireless communication, which is equal to in, is the total energy consumption of task processing.

[0160] refer to Figure 7 As shown, Figure 7 A comparison chart of the energy consumption generated by the average tasks of this method and AONM and PONM is given. For the AONM strategy, since all tasks are assigned to the MEC server, and the energy consumption of local processing is much higher than the energy consumed by the vehicle during server processing, the energy consumption is the lowest. We can find that the vehicle energy consumption generated by this method is reduced by up to 14% compared with PONM, so this method can provide methods and help to reduce energy consumption of electric intelligent connected vehicles.

[0161] refer to Figure 8 As shown, Figure 8 The average processing delay comparison chart of this method and AONM and PONM is given. The average processing delay of this method is reduced by up to 45% compared with the fully unloaded AONM and by up to 20% compared with PONM. This method has superior performance in terms of delay, which is of great significance for shortening the delay of vehicle tasks and improving service quality.

[0162] refer to Fig. 9 As shown, Fig. 9 The average task completion rates of this method, AONM and PONM are given. Fig. 9 It can be seen that the average task completion rate decreases as the number of vehicles increases. We can find that the task completion rate of this method is up to 75% higher than that of AONM and up to 24% higher than that of PONM. This means that more computing resources are properly utilized at the same time compared to other methods, which improves the task completion rate. This method provides a feasible means for the timely and stable completion of tasks.

[0163] refer to Fig.10 As shown, Fig.10 The average task offloading ratio of this method and PONM is given, that is, the average ratio of a task to be handled by the server. Fig.10 It can be seen that the average task offloading ratio decreases with the increase of the number of vehicles. This is because the computing resources of the server are limited. When the amount of tasks is large, the server cannot bear it and can only process it locally. In addition, the average task offloading ratio of this method is higher than that of PONM, which proves that this method makes better use of server computing resources than PONM, allowing more tasks to be executed on the MEC server, improving the efficiency of task execution, and saving vehicle battery energy.

[0164] refer to Fig.11 As shown, Fig.11The average calculation rate of this method, AONM and PONM is given, and the average calculation rate increases with the increase of vehicles. When the number of vehicles is small, the average calculation rate of this method is similar to that of PONM, indicating that the tasks generated by both schemes can be processed in time. On the other hand, when the number of vehicles increases, this method also considers the roadside units in the direction of vehicle travel, which is different from PONM. In addition, AONM does not consider local computing resources, so its average calculation rate is the lowest among the three methods.

[0165] The present invention also provides an energy-saving task unloading system for an electric intelligent network-connected vehicle, which is used to implement the energy-saving task unloading method for an electric intelligent network-connected vehicle described above, comprising:

[0166] The first processing module is used to receive the vehicle task request proposed by the vehicle 2 within its coverage area during the simulation time, and send it to the macro base station 3. The macro base station 3 sorts out the vehicle set and the task set according to all the vehicle request information within the time;

[0167] The second processing module is used for the macro base station 3 to start solving the task unloading strategy according to the information of the vehicle set and the task set to obtain the final unloading plan, and send the final unloading plan to the roadside unit 1. The roadside unit 1 sends the unloading plan of the corresponding vehicle to the vehicle 2 within its coverage area, so as to realize the energy-saving task unloading of the electric intelligent connected vehicle.

[0168] Specifically, according to Figure 2 As shown, a macro base station 3 is set up on the side of the road, and several roadside units 1 are evenly arranged along the extension direction of the roadside; several roadside units 1 are communicatively connected with vehicles 2, and the macro base station 3 is communicatively connected with several roadside units 1; the roadside unit 1 is provided with a communication base station 11, an optical fiber network unit 12 and a MEC server 13; the communication base station 11 is communicatively connected with the optical fiber network unit 12 and the MEC server 13 respectively, the roadside unit 1 is communicatively connected with other roadside units 1 through the optical fiber network unit 12, and the roadside unit 1 is communicatively connected with several vehicles 2 through the communication base station 11.

[0169] Example

[0170] Without loss of generality, in this embodiment, the energy-saving task unloading system of the electric intelligent connected vehicle is assigned, the length of the two-way road is 800m, the number of roadside units 2 is 4, the roadside units 2 are evenly distributed on the road, and the vehicle entering the considered road scene obeys Poisson distribution.

[0171] When a vehicle enters the road, a mission is generated and the information about the vehicle and the mission is notified to the nearest roadside unit. A mission is divided into four parts, namely The sizes of the video task, navigation task, and intelligent traffic dispatching task are 10MB, 3MB, and 0.5MB respectively, and the delay thresholds are 1500ms, 400ms, and 150ms.

[0172] Wireless bandwidth of roadside units Set to 20 MHz, transmit power is 100mW, and the ambient noise power σ is 10 -13 W.

[0173] The processing capacity of the MEC server is 10G Cycles / s, and the processing capacity of the vehicle is 4G Cycles / s. The computing density is 150Cycles / bit, and the energy density is 10 -26 J / Cycles.

[0174] By comparing the task offloading method proposed in the present invention with traditional AONM and PONM, it is found that the method proposed in the present invention has certain advantages in shortening task processing delay and reducing vehicle energy consumption; by comparing the task completion rate, average task offloading ratio and average calculation rate of the three methods, it is shown that the task offloading method proposed in the present invention can utilize idle server resources and alleviate server load imbalance, which is of great significance for reducing task processing delay and reducing vehicle energy consumption.

[0175] In summary, the present invention provides an energy-saving task unloading method and system for electric intelligent networked vehicles, which unloads tasks to roadside units, shortens task delays and reduces vehicle energy consumption. The roadside unit transmits tasks and processing results through optical fibers, alleviates the problem of unbalanced server load, improves computing resource utilization, and further shortens task delays. A packet loss penalty mechanism is also provided to discard unfinished tasks in advance, avoiding waste of computing resources while improving task completion rate. Finally, a discrete particle swarm algorithm is used to give a near-optimal unloading solution in a very short time; the average task processing delay of the task unloading method of the present invention is reduced by up to 20% and 45% respectively compared with partial unloading to the nearest MEC server (PONM) and full unloading to the nearest MEC server (AONM), the average energy consumption of the task is reduced by up to 14% compared with PONM, and the task completion rate is increased by up to 24% and 75% compared with PONM and AONM. ​​In the efficient and energy-saving task unloading method, idle computing resources are effectively utilized, the problem of unbalanced server load is alleviated, task delay and vehicle energy consumption are reduced, and the unloading method has superior performance.

[0176] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for unloading energy-saving tasks of an electric intelligent network-connected vehicle, characterized in that: The steps include: Step 1, receiving a vehicle task request from a vehicle (2) within its coverage area during the simulation time, and sending it to a macro base station (3), and the macro base station (3) collates all vehicle request information within the time to obtain a vehicle set and a task set; Step 2: The macro base station starts to solve the task offloading strategy based on the information of the vehicle set and the task set to obtain the final offloading plan, and sends the final offloading plan to the roadside unit. The roadside unit sends the offloading plan of the corresponding vehicle to the vehicles within its coverage area, thereby realizing the energy-saving task offloading of the electric intelligent connected vehicle; The macro base station (3) simulates the actual situation based on the information of the vehicle set and the task set, and processes the vehicle tasks locally and on the server respectively in the simulation of the actual situation. The specific process is as follows: The vehicle task is evenly divided into several blocks. According to the final unloading plan given by the macro base station (3), part of the vehicle task is unloaded to the roadside unit (1) closest to the vehicle, or unloaded to the nearest roadside unit (1) and then transmitted to the next roadside unit (1) in the driving direction through the optical fiber network unit (12) for processing. The remaining vehicle tasks are processed locally. After the nearest roadside unit (1) receives the vehicle task, it waits for the server to execute it in the task queue of the MEC server (13). After the server completes the processing, it sends the result to the vehicle (2); if the roadside unit (1) executing the task cannot process it, it performs packet loss processing and adds a penalty factor; Among them, the final unloading scheme is processed by providing a vehicle set, a task set and a simulation time, and the task delay and vehicle energy consumption are calculated according to the processing results to obtain the fitness value, and the fitness value is obtained; The latency includes the local task processing time, wireless communication time, server queue waiting time and server processing time; the vehicle energy consumption includes the local task energy consumption and the wireless communication energy consumption for task uploading; The calculation formula for task local processing time is as follows: in, is the local execution ratio of vehicle tasks; The amount of computation required to complete the task; k is the vehicle; t is the time; I t,k is the vehicle task generated by vehicle k at time t; is the computing power of vehicle k; The calculation formula of task local energy consumption is as follows: Where κ is the energy density; The calculation formula for wireless communication time is as follows: in, Calculate the data as needed; is the local execution ratio of vehicle tasks; is the wireless communication transmission rate; The calculation formula for wireless communication energy consumption of task upload is as follows: in, represents the transmission power of vehicle k; Calculate the data as needed; is the local execution ratio of vehicle tasks; is the wireless communication transmission rate; Wireless communication time; The server processing time is calculated as follows: in, represents the computing power of the server closest to vehicle k; is the local execution ratio of vehicle tasks; The amount of computation required to complete the task; The calculation formula for server queue waiting time is: in Indicates that the server is The amount of remaining tasks in the server queue at the moment, t is I t,k The generation time of ,δ is the offloading scheme given by the macro base station; Calculation of remaining tasks in the server queue: Where ω is the set of all tasks.

2. The energy-saving task unloading method of an electric intelligent network-connected vehicle according to claim 1, characterized in that: In step 2, the macro base station (3) starts to solve the task offloading strategy based on the information of the vehicle set and the task set to obtain the final offloading solution, wherein the process is as follows: S1. Set the particle dimension to 2D, and initialize the particle swarm and parameter settings, N = 1; where D is the number of tasks in the task set; S2. Assign the first D items of the particle dimension to each task in the task set. The last D item is assigned to each task in the task set. S3, providing a vehicle set, a task set and a simulation time for processing, and calculating the task delay and vehicle energy consumption according to the processing results to obtain an adaptation value; S4, update the individual optimal and global optimal values ​​of all particles; S5. Determine whether iteration is still needed; When the task still needs to be iterated, update the speed and position of each particle; increase the current iteration number by 1; otherwise, repeat S1 to S4 until all task iterations are completed, and output the fitness value and unloading plan corresponding to the global optimal value.

3. The energy-saving task unloading method of an electric intelligent network-connected vehicle according to claim 1, characterized in that: In step 2, after the vehicle task is generated, it is first processed locally. If the local execution part of the vehicle task cannot be completed within the delay threshold, the task is directly discarded and the evaluation index penalty factor is assigned. Instead, local processing begins while the vehicle offloads the remainder of the task to the nearest roadside unit (1) via wireless communication.

4. The energy-saving task unloading method of an electric intelligent network-connected vehicle according to claim 3, characterized in that: The roadside unit (1) distributes the task to the vehicles within its range according to the unloading scheme determined by the macro base station (3). The vehicle unloads the task to the nearest roadside unit (1) according to the unloading scheme, or unloads the task to the nearest roadside unit (1) and transmits it to the next roadside unit (1) in the driving direction through optical fiber for processing. After receiving the task, the roadside unit (1) places the task in the task queue of the MEC server (13) and waits for the server to execute it. The MEC server (13) calculates the time for the task to be processed. If the part can be completed within the delay threshold, it is processed. Otherwise, the task is directly discarded and a penalty factor is added to the evaluation index. At the same time, the penalty factor needs to consider the energy consumption generated by local processing and task upload time. In the case of task discard, the energy consumption generated by the vehicle for the task can be expressed as: in, is the total energy consumption of task processing.

5. The energy-saving task unloading method of an electric intelligent network-connected vehicle according to claim 3, characterized in that: Vehicle tasks include video tasks, navigation tasks, and intelligent traffic scheduling tasks, each of which has different delay thresholds and task sizes.

6. An energy-saving task unloading system for an electric intelligent network-connected vehicle, used to implement an energy-saving task unloading method for an electric intelligent network-connected vehicle according to any one of claims 1 to 5, characterized in that: include: A first processing module is used to receive a vehicle task request from a vehicle (2) within its own coverage area during simulation time, and send it to a macro base station (3). The macro base station (3) obtains a vehicle set and a task set based on all vehicle request information during the time; The second processing module is used for the macro base station (3) to start solving the task unloading strategy according to the information of the vehicle set and the task set to obtain a final unloading plan, and send the final unloading plan to the roadside unit (1). The roadside unit (1) sends the unloading plan of the corresponding vehicle to the vehicle (2) within its coverage area, so as to realize the energy-saving task unloading of the electric intelligent networked vehicle.

7. The energy-saving task unloading system for electric intelligent networked vehicles according to claim 6 is characterized in that: The macro base station (3) is arranged on the side of the road, and a plurality of roadside units (1) are evenly arranged along the extension direction of the roadside; the plurality of roadside units (1) are connected in communication with the vehicles (2), and the macro base station (3) is connected in communication with the plurality of roadside units (1); the roadside unit (1) is provided with a communication base station (11), an optical fiber network unit (12) and an MEC server (13); the communication base station (11) is connected in communication with the optical fiber network unit (12) and the MEC server (13) respectively, the roadside unit (1) is connected in communication with other roadside units (1) via the optical fiber network unit (12), and the roadside unit (1) is connected in communication with the plurality of vehicles (2) via the communication base station (11).

Images