V2X power allocation method, system, storage medium and terminal

Abstract

The present invention provides a V2X power allocation method, system, storage medium and terminal, comprising the following steps: mapping the relationship between V2X throughput, QoS constraints and power allocation from three-dimensional space to two-dimensional plane, obtaining the power allocation feasible domain that guarantees the maximum throughput under QoS constraints; for each communication time slot, calculating the optimal power allocation solution corresponding to each power allocation feasible domain; normalizing the channel state information and the corresponding optimal power allocation solution; training a DNN model based on the normalized channel state information under a preset number of communication time slots and the corresponding optimal power allocation solution, so as to obtain the V2X power allocation optimal solution under the channel state information characteristics to be predicted based on the trained DNN model. The V2X power allocation method, system, storage medium and terminal of the present invention implement V2X power allocation based on deep neural network and classification algorithm, guarantee QoS constraints, improve system throughput, and ensure the reliability of vehicle network.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless communications, and in particular to a V2X power allocation method, system, storage medium and terminal. Background Art

[0002] 5G technology has the characteristics of high speed, low latency, and high reliability. With the commercialization of 5G technology, many fields around the world will undergo earth-shaking changes, such as medical, education, energy, and Internet of Vehicles. Thanks to the development of 5G and intelligent transportation system technology, the feasibility of Internet of Vehicles technology has been further demonstrated, so Internet of Vehicles technology will be widely used in the near future. Internet of Vehicles technology has great potential to improve traffic efficiency and provide rich online information and multimedia services. This will bring a lot of convenience to vehicle drivers, such as communicating with nearby drivers and improving traffic operation efficiency, and has therefore received great attention from industry and academia.

[0003] In the early research of Internet of Vehicles, Dedicated Short Range Communication (DSRC) technology based on IEEE 802.11p was relatively popular, mainly used for communication between vehicles or between vehicles and infrastructure. However, DSRC technology has the following shortcomings:

[0004] (1) The communication coverage distance is short, and the vehicles are traveling at high speed, which leads to frequent interruptions in communication between vehicles;

[0005] (2) Vehicles accessing the Internet through DSRC gateways requires extensive infrastructure deployment, and the system deployment cost is too high;

[0006] (3) DSRC technology lacks a centralized control unit;

[0007] (4) Due to the random access characteristics of the Carrier Sense Multiple Access with Collision Avoidance (CSMA / CA) mechanism of the IEEE 802.11p protocol, when the vehicle density on the road is high, the channel competition between vehicles will be very fierce.

[0008] In recent years, with the continuous popularization and updating of cellular communication technology, Vehicle-to-Everything (V2X) technology defined by the 3rd Generation Partnership Project (3GPP) has gradually become the main communication technology in the field of Internet of Vehicles research.

[0009] In vehicle communications, V2X technology has excellent application prospects, and its main advantages are as follows:

[0010] First, V2X supports low-latency direct communication without relying on network assistance;

[0011] Second, the Intelligent Transportation System (ITS) for safety applications uses the 5.9 GHz unified spectrum. The 5.9 GHz band is open to the intelligent transportation and Internet of Vehicles industries, and there is no exclusivity in the use of the band.

[0012] Third, it supports high-speed car scenarios and supports relative speeds up to 500km / h;

[0013] Fourth, V2X follows strict minimum performance requirements defined by 3GPP, so the technology is safer and performance is predictable;

[0014] Fifth, V2X can provide better coverage and more reliable communications in any density deployment scenario.

[0015] In V2X technology, vehicle to vehicle (V2V) and vehicle to infrastructure (V2I) communication have great potential. V2V technology supports direct device to device (D2D) communication. This method effectively reduces communication latency and provides an important guarantee for the safety and reliability of vehicle communication technology. With the continuous popularization and updating of cellular networks, V2I technology based on base stations (BS) is also becoming more mature. V2I technology mainly provides vehicle users with rich online information and multimedia services through vehicle-to-infrastructure communication. In order to improve spectrum utilization, V2V links and V2I links are advocated to share the same spectrum. However, the transmission power of V2V and V2I will interfere with each other, so this problem needs to be solved through power allocation. For power allocation technology, channel state information (CSI) feedback is crucial. Due to the high-speed movement of vehicles in the Internet of Vehicles, its channel characteristics are fast-changing and CSI feedback is also delayed. Although the BS samples the CSI, the sampled information is often inaccurate. This is a critical challenge for vehicle communications. In addition, due to the high-speed movement of vehicles, the reliability of the vehicle network needs to be guaranteed, which is another issue that cannot be ignored. Therefore, a good power allocation solution design is of great significance and value for V2X communication.

[0016] There have been many studies on the application of power allocation technology in V2X systems. Power allocation technology mainly refers to reducing interference between each other and improving system performance by controlling the transmission power of V2V and V2I links. The existing power control technologies mainly adopt intelligent optimization theory and mathematical theory, including: water injection algorithm, auction algorithm, heuristic algorithm, machine learning algorithm, etc. These algorithms have improved system performance to a certain extent, but there are still many problems. For example, the existing power allocation algorithm relies on instantaneous CSI feedback and ignores strict reliability guarantees. However, due to the high-speed mobility of vehicles in V2X, CSI feedback is usually delayed, which means that traditional power allocation algorithms are no longer applicable. In addition, the reliability of each link in V2X, that is, the quality of service (QoS), is usually difficult to be reliably guaranteed. Summary of the invention

[0017] In view of the shortcomings of the prior art mentioned above, the purpose of the present invention is to provide a V2X power allocation method, system, storage medium and terminal, which implement V2X power allocation based on deep neural networks (DNN) and classification algorithms, ensure QoS constraints and improve system throughput under delayed CSI feedback in the V2X scenario, and ensure the reliability of the vehicle network.

[0018] To achieve the above-mentioned purpose and other related purposes, the present invention provides a V2X power allocation method, comprising the following steps: mapping the relationship between V2X throughput, QoS constraints and power allocation from three-dimensional space to a two-dimensional plane, and obtaining a power allocation feasible domain that ensures maximum throughput under QoS constraints; for each communication time slot, calculating the power allocation optimal solution corresponding to each power allocation feasible domain; normalizing the channel state information and the corresponding power allocation optimal solution; training a DNN model based on the normalized channel state information under a preset number of communication time slots and the corresponding power allocation optimal solution, so as to obtain the V2X power allocation optimal solution under the channel state information characteristics to be predicted based on the trained DNN model.

[0019] In one embodiment of the present invention, obtaining a feasible region of power allocation for maximizing throughput under QoS constraints includes the following steps:

[0020] Assume that the V2I link set is M = {1, ..., M}, the V2V link set is K = {1, ..., K}, and the transmitter uses the maximum transmission power P max , according to the channel gain {h m,B 、h k 、h k,B 、h m,k}, calculate the corresponding maximum receiving / interference power {P maxh m,B , P max h k , P max h k,B , P max h m,k}; where h m,B is the signal gain of the mth V2I link, h k is the signal gain of the kth V2V link, h k,B is the interference gain of the mth V2I link, h m,k is the interference gain of the kth V2V link, P max h m,B is the maximum received power of the mth V2I link, P max h k is the maximum received power of the kth V2V link, P max h k,B is the maximum interference power of the mth V2I link, P max h m,k is the maximum interference power of the kth V2V link, where m∈M and k∈K;

[0021] According to the maximum receiving / interference power, the maximum allowed interference power / minimum allowed receiving power {P * k,B , P * m,k , P * m,B , P * k}; Among them, P * k,B is the maximum allowed interference power corresponding to the mth V2I link, P * m,k is the maximum allowed interference power corresponding to the kth V2V link, P * m,B is the minimum allowed receiving power corresponding to the mth V2I link, P * k is the minimum allowed receiving power corresponding to the kth V2V link;

[0022] The power allocation feasible region is acquired according to the maximum allowed interference power / minimum allowed received power.

[0023] In one embodiment of the present invention, when 0 <P * k,B <P max And P * k <P * k,B When , the power allocation feasible region is above the baseline;

[0024] When 0 <P * m,k <P max And P * m,B <P * m,k When , the power allocation feasible region is below the baseline;

[0025] When P * k <P max And P * m,B <P max When , the feasible region of power allocation is located on both sides of the baseline;

[0026] In other cases, the power allocation feasible domain has no solution.

[0027] In one embodiment of the present invention, when the power allocation feasible region is above the baseline, the power allocation optimal solution (P k ,P m )= (P * k,B ,P max );

[0028] When the power allocation feasible region is below the baseline, the power allocation optimal solution (P k ,P m )=(P max ,P * m,k );

[0029] When the power allocation feasible region is located on both sides of the baseline, the optimal power allocation solution (P k ,P m )=arg max(Γ), where Γ={C(P * k ,P max ),C(P max ,P * m,B )}, C(·) represents the V2X system throughput function, P k represents the transmission power of the kth V2V link, P m represents the transmit power of the mth V2I link.

[0030] In one embodiment of the present invention, when the channel state information and the corresponding optimal power allocation solution are normalized, (C i -C min ) / (C max -C min ) maps the value to the interval (0,1), where Ci The i-th value, C min is the minimum value of the interval, C max The maximum value of the interval.

[0031] In an embodiment of the present invention, the preset number is no greater than ten.

[0032] In one embodiment of the present invention, the input layer of the DNN model includes 4*6 neurons, the three hidden layers are respectively provided with 40, 40, and 50 neurons, and the output layer includes 24*2 neurons.

[0033] Correspondingly, the present invention provides a V2X power allocation system, including an acquisition module, a calculation module, a normalization module and an allocation module;

[0034] The acquisition module is used to map the relationship between the throughput, QoS constraints and power allocation of V2X from a three-dimensional space to a two-dimensional plane, and obtain a feasible domain of power allocation that guarantees maximum throughput under QoS constraints;

[0035] The calculation module is used to calculate the optimal power allocation solution corresponding to each power allocation feasible domain for each communication time slot;

[0036] The normalization module is used to normalize the channel state information and the corresponding optimal power allocation solution;

[0037] The allocation module is used to train a DNN model based on the normalized channel state information under a preset number of communication time slots and the corresponding optimal power allocation solution, so as to obtain the V2X power allocation optimal solution under the channel state information characteristics to be predicted based on the trained DNN model.

[0038] The present invention provides a storage medium having a computer program stored thereon, which implements the above-mentioned V2X power allocation method when executed by a processor.

[0039] Finally, the present invention provides a terminal, including: a processor and a memory;

[0040] The memory is used to store computer programs;

[0041] The processor is used to execute the computer program stored in the memory so that the terminal performs the above-mentioned V2X power allocation method.

[0042] As described above, the V2X power allocation method, system, storage medium and terminal of the present invention have the following beneficial effects:

[0043] (1) V2X power allocation is implemented based on deep neural networks and classification algorithms, which effectively reduces the mutual interference between V2V and V2I links and optimizes the throughput of V2V and V2I links while ensuring QoS constraints;

[0044] (2) It can use machine learning algorithms to predict power allocation for V2X links based on previous CSI without obtaining the current CSI, effectively ensuring the reliability of the Internet of Vehicles;

[0045] (3) It meets the communication delay of V2V, saves the power of the transmitter, has low computational complexity and strong practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Shown is a flow chart of a V2X power allocation method in one embodiment of the present invention;

[0047] Figure 2 A schematic diagram showing a scenario of V2X communication in an embodiment;

[0048] Figure 3 It is a schematic diagram showing a feasible domain of power allocation of the present invention in one embodiment;

[0049] Figure 4 Shown is a schematic diagram of the structure of a DNN model of the present invention in one embodiment;

[0050] Figure 5 A schematic diagram showing a comparison of V2X throughputs of a power allocation algorithm of the present invention and a power allocation algorithm in the prior art in one embodiment;

[0051] Figure 6 A schematic diagram showing a comparison between the V2X system throughput obtained by the DNN model of the present invention and the optimal system throughput in one embodiment;

[0052] Figure 7 It is a schematic diagram showing the structure of a V2X power distribution system in one embodiment of the present invention;

[0053] Figure 8 Shown is a schematic structural diagram of a terminal in one embodiment of the present invention.

[0054] Component number description

[0055] 71 Get Module

[0056] 72 Computing Modules

[0057] 73 Normalization module

[0058] 74 Allocation Module

[0059] 81 Processor

[0060] 82 Memory DETAILED DESCRIPTION

[0061] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0062] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0063] The V2X power allocation method, system, storage medium and terminal of the present invention first classify the power allocation feasible domain to ensure QoS constraints and improve system throughput, and then train a DNN model to predict the power allocation scheme based on the data set obtained by the classification, thereby effectively solving the problem of delayed CSI feedback in V2X communication and achieving better performance gain.

[0064] like Figure 1 As shown, in one embodiment, the V2X power allocation method of the present invention includes the following steps:

[0065] Step S1: Map the relationship between V2X throughput, QoS constraints and power allocation from three-dimensional space to a two-dimensional plane to obtain a feasible domain of power allocation that maximizes throughput under QoS constraints.

[0066] Specifically, the channel state information includes a signal link gain and an interference link gain. Figure 2The figure shows an application scenario of uplink communication in spectrum sharing mode of a typical V2X scenario. On the one hand, the V2I link exchanges information between mobile vehicles and roadside infrastructure. On the other hand, the V2V link is used for information transmission between two mobile vehicles. Assume that M vehicles communicate through the V2I link and K pairs of vehicles transmit local data through the V2V link. Each vehicle is only allowed to use one link (V2V or V2I) to transmit data. In order to reduce the interference between different links, it is assumed that the spectrum of a V2I link can only be reused by a single V2V link. In addition, a V2V link is only allowed to share the spectrum of a single V2I link. First, the projection constraint analysis (PCA) method is used to map the relationship between the throughput, QoS constraints and power allocation of V2X from three-dimensional space to a two-dimensional plane, so that the power allocation feasible domain for the problem of maximizing throughput under QoS constraints can be obtained from the two-dimensional plane. The power allocation feasible domain is irregular.

[0067] In one embodiment of the present invention, obtaining a feasible region of power allocation for maximizing throughput under QoS constraints includes the following steps:

[0068] 11) Assume that the V2I link set is M = {1, ..., M}, the V2V link set is K = {1, ..., K}, and the transmitter uses the maximum transmission power P max , according to the channel gain {h m,B 、h k 、h k,B 、h m,k}, calculate the corresponding maximum receiving / interference power {P max h m,B , P max h k , P max h k,B , P max h m,k}; where h m,B is the signal gain of the mth V2I link, h k is the signal gain of the kth V2V link, h k,B is the interference gain of the mth V2I link, h m,k is the interference gain of the kth V2V link, P max h m,B is the maximum received power of the mth V2I link, P max h k is the maximum received power of the kth V2V link, P max h k,B is the maximum interference power of the mth V2I link, P max h m,k is the maximum interference power of the kth V2V link, where m∈M and k∈K.

[0069] 12) According to the maximum receiving / interference power, obtain the maximum allowed interference power / minimum allowed receiving power {P * k,B , P * m,k , P * m,B , P * k}; Among them, P * k,B is the maximum allowed interference power corresponding to the mth V2I link, P * m,k is the maximum allowed interference power corresponding to the kth V2V link, P * m,B is the minimum allowed receiving power corresponding to the mth V2I link, P * k is the minimum allowed receiving power corresponding to the kth V2V link.

[0070] Specifically, with the maximum received power P max h m,B Take λ as an example to calculate the corresponding maximum allowed interference power. Given the QoS constraint λ, the corresponding signal-to-noise ratio γ can be calculated according to the formula log2(1+γ)≥λ. Then, according to the formula The corresponding maximum permissible interference power P is derived * k,B Among them, σ 2 is the noise power, ρ m,k is the spectrum reuse indicator, ρ m,k =1 means that the spectrum of the mth V2I link is reused by the kth V2V link, otherwise ρ m,k =0; P m represents the transmission power of the mth V2I link, P k represents the transmission power of the kth V2V link, and P m =P max , P k =P * k,B For P max h k , P max h k,B , P max h m,k The corresponding maximum allowed interference power / minimum allowed reception power can also be calculated similarly.

[0071] 13) According to the maximum allowed interference power / minimum allowed receiving power, the power allocation feasible domain can be obtained by using a mathematical problem model.

[0072] Specifically, taking the diagonal of the two-dimensional plane, i.e., (0,0)-(23,23) as the baseline, the power allocation feasible domain is classified as follows: Figure 3 Four scenarios are shown:

[0073] a) above the baseline;

[0074] b) below the baseline;

[0075] c) on both sides of the baseline;

[0076] d) No solution.

[0077] In one embodiment of the present invention, the algorithm for determining the feasible region of power allocation is as follows:

[0078] When 0 <P * k,B <P max And P * k <P * k,B When , the power allocation feasible region is above the baseline;

[0079] When 0 <P * m,k <P max And P * m,B <P * m,k When , the power allocation feasible region is below the baseline;

[0080] When P * k <P max And P * m,B <P max When , the feasible region of power allocation is located on both sides of the baseline;

[0081] In other cases except the above three cases, the power allocation feasible domain has no solution.

[0082] Step S2: For each communication time slot, calculate the optimal power allocation solution corresponding to each power allocation feasible domain.

[0083] Specifically, in each communication time slot, the optimal power allocation solution corresponding to each power allocation feasible domain is calculated according to the channel state information. In one embodiment of the present invention, when the power allocation feasible domain is above the baseline, the optimal power allocation solution (P k ,P m )=(P * k,B ,P max); When the power allocation feasible region is below the baseline, the power allocation optimal solution (P k ,P m )= (P max ,P * m,k ); When the power allocation feasible region is located on both sides of the baseline, the optimal solution for power allocation (P k ,P m )=arg max(Γ), where Γ={C(P * k ,P max ),C(P max ,P * m,B )}, C(·) represents the V2X system throughput function, P k represents the transmission power of the kth V2V link, P m represents the transmit power of the mth V2I link.

[0084] Therefore, the above power allocation algorithm maximizes the improvement of system throughput while ensuring the QoS constraints of each link.

[0085] Step S3: normalize the channel state information and the corresponding optimal power allocation solution.

[0086] Specifically, normalization is performed on the collected channel state information and the corresponding optimal power allocation solution, and subsequent model training and verification are performed based on the normalized data.

[0087] In one embodiment of the present invention, when the channel state information and the corresponding optimal power allocation solution are normalized, (C i -C min ) / (C max -C min ) maps the value to the interval (0,1), where C i The i-th value, C min is the minimum value of the interval, C max The maximum value of the interval.

[0088] Step S4: training a DNN model based on the normalized channel state information of a preset number of communication time slots and the corresponding optimal power allocation solution, so as to obtain the optimal solution for V2X power allocation under the characteristics of the channel state information to be predicted based on the trained DNN model.

[0089] Specifically, DNN (Deep Neural Networks), or deep neural network, is a technology in the field of machine learning. DNN uses a large number of nonlinear layers to model high-level abstractions, and produces a powerful representation ability for the mapping relationship between random vector units, which requires very little manual calculation and provides possibilities for many complex problems. The outstanding performance of DNN stems from its ability to use statistical learning methods to extract high-level features from raw sensory data and obtain effective representations of the input space in a large amount of data. This is different from the previous methods of manually extracting features or expert designing rules. Since the channel characteristics of V2X are fast-varying and there is delayed channel state information feedback, the traditional power allocation algorithm based on instantaneous channel state information feedback is no longer applicable. However, the statistical channel state information of past time slots has a high guiding significance for the power allocation method at the current moment. Therefore, the present invention uses the statistical learning method of DNN to process the channel state information of past image time slots, which is very applicable.

[0090] The normalized channel state information and the corresponding optimal power allocation solution are divided into a training set and a validation set, and the DNN model is trained and validated to obtain a trained DNN model to achieve the optimal power allocation in V2X communication.

[0091] When training the DNN model, the channel state information of the previous multiple communication time slots is used as the input of the DNN model, which can be expressed as x(t)={A(t-nt0),A(t-(n-1)t0),...,A(t-t0)}, where t0 represents a time slot, A(t-nt0) represents the channel state information of the previous n communication time slots, which can be expressed as The optimal power allocation solution y(t) = {P k ,P m} as the output of the DNN model, and train the DNN model until convergence. When verifying the DNN model, the channel state information of the previous multiple time slots is used as the input of the trained DNN model, and the output of the trained DNN model is compared with the target output to obtain the model prediction accuracy.

[0092] It should be noted that too many input parameters may easily lead to overfitting of the DNN model, so the number of communication time slots is not greater than 10. Preferably, 6 communication time slots are used. At the same time, the transmission power of the V2V and V2I links is set to an integer value, that is, {0,1,...,23}dBm. Preferably, as Figure 4As shown, the input layer of the DNN model includes 4*6 neurons, the three hidden layers are set to 40, 40, and 50 neurons respectively, and the output layer includes 24*2 neurons. The number of neurons set in this way is more suitable as the weight of the DNN model, which is conducive to the convergence of the DNN model. By adopting the DNN model for prediction, it is effectively avoided to directly rely on instantaneous CSI feedback, and with the support of historical data, the current optimal power allocation solution can be predicted more accurately. Therefore, the V2X power allocation method of the present invention solves the problem of CSI delayed feedback in V2X communication.

[0093] The V2X power allocation method of the present invention is further described below through specific embodiments.

[0094] In this embodiment, first, according to the protocol standard of 3GPP 37.885, a V2X communication simulation model on a highway is built. There are 20 V2I communication links and 20 V2V communication links. Each V2V link is allowed to reuse the spectrum of at most one V2I link. Next, considering the vehicle communication situation within 1000 time slots, the vehicle speed is set to 144km / h. Then, the channel state information under each spectrum reuse pair is obtained, including the signal channel and the interference channel, that is, {h m,B 、h k 、h k,B 、h m,k According to the channel state information, the corresponding maximum receiving / interference power {P max h m,B , P max h k , P max h k,B , P max h m,k} and bring it into the mathematical model of the problem (P k ,P m )=argmax[log2(1+γ k )+log2(1+γ m )], where γ k and γ m are the signal-to-noise ratios of the kth V2V link and the mth V2I link, respectively. Thus, the maximum allowed interference power / minimum allowed receiving power {P * k,B P * m,k P * m,B P * kThen, the classification of the power allocation feasible domain is determined by the maximum allowed interference power / minimum allowed received power, and the optimal power allocation solution is derived for each classification. The above steps are completed for one time slot, that is, the above process is performed for each communication time slot to obtain a dynamic power allocation solution.

[0095] Subsequently, the DNN model is trained and tested based on the channel state information and the corresponding optimal power allocation solution. The channel state information in the first 6 communication time slots is used as input, and the optimal power allocation solution is used as the output of the DNN model. 80,000 sets of such data are used for model training, and the number of iterations is set to 5,000 until the DNN model converges. After the DNN model is trained, the optimal V2X power allocation solution under the channel state information characteristics to be predicted can be completed.

[0096] Finally, the trained DNN model is used in the actual V2X scenario for real-time power allocation. A highway V2X communication scenario model under the 3GPP37.885 standard is built, and the relevant parameters are the same as above. Afterwards, as the vehicle moves, in each communication time slot, the channel state information of the first 6 image time slots is input into the DNN model, and the QoS constraint satisfaction and system throughput are calculated according to the corresponding output power allocation scheme. In the actual system, V2X communication is always ongoing. As long as there is previous channel state information, the corresponding optimal power solution can be calculated through the classification algorithm, and the DNN model can be trained to fit it based on these data, and then the trained DNN prediction model can be continuously applied to the actual V2X system.

[0097] Specifically, the present invention uses MATLAB system-level simulation tools to evaluate the performance of improving system throughput while ensuring QoS constraints. Figure 5 The figure shows the simulation comparison of V2X throughput of the power allocation algorithm of the present invention and the power allocation algorithm in the prior art. Figure 5 It can be seen that the power allocation algorithm proposed in the present invention not only meets the QoS constraints of each link, but also greatly improves the total throughput of the V2X system. Compared with other popular algorithms, it shows better performance.

[0098] Figure 6 The figure shows the simulation comparison of the V2X system throughput and the optimal system throughput obtained by the DNN model of the present invention under delayed CSI feedback. Each simulation figure is implemented under 1000 time slots. Figure 6 It can be seen that the DNN model of the present invention appropriately solves the challenges brought by V2X delayed CSI feedback and achieves throughput performance comparable to the ideal optimal solution.

[0099] like Figure 7 As shown, in one embodiment, the V2X power allocation system of the present invention includes an acquisition module 71, a calculation module 72, a normalization module 73 and an allocation module 74.

[0100] The acquisition module 71 is used to map the relationship between the throughput, QoS constraints and power allocation of V2X from a three-dimensional space to a two-dimensional plane, and obtain a feasible domain of power allocation that ensures maximum throughput under QoS constraints.

[0101] The calculation module 72 is connected to the acquisition module 71 and is used to calculate the optimal power allocation solution corresponding to each power allocation feasible domain for each communication time slot.

[0102] The normalization module 73 is connected to the calculation module 72 and is used to normalize the channel state information and the corresponding optimal power allocation solution.

[0103] The allocation module 74 is connected to the normalization module 73, and is used to train the DNN model based on the channel state information of a preset number of communication time slots processed by normalization and the corresponding optimal power allocation solution, so as to obtain the V2X power allocation optimal solution under the channel state information characteristics to be predicted based on the trained DNN model.

[0104] Among them, the structures and principles of the acquisition module 71, the calculation module 72, the normalization module 73 and the allocation module 74 correspond one to one with the steps in the above-mentioned V2X power allocation method, so they are not repeated here.

[0105] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the x module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0106] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more digital singnal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0107] The storage medium of the present invention stores a computer program, which implements the above-mentioned V2X power allocation method when executed by a processor. The storage medium includes: ROM, RAM, disk, USB flash drive, memory card or optical disk, etc., which can store program codes.

[0108] like Figure 8 As shown, in one embodiment, the terminal of the present invention includes: a processor 81 and a memory 82 .

[0109] The memory 82 is used to store computer programs.

[0110] The memory 82 includes: ROM, RAM, disk, USB flash drive, memory card or CD and other media that can store program codes.

[0111] The processor 81 is connected to the memory 82 and is used to execute the computer program stored in the memory 82 so that the terminal executes the above-mentioned V2X power allocation method.

[0112] Preferably, the processor 81 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0113] In summary, the V2X power allocation method, system, storage medium and terminal of the present invention realize V2X power allocation based on deep neural network and classification algorithm, effectively reduce the mutual interference between V2V and V2I links, optimize the throughput of V2V and V2I links under the premise of ensuring QoS constraints; can predict power allocation of V2X links based on previous CSI without obtaining current CSI, effectively guarantee the reliability of Internet of Vehicles; meet the communication delay of V2V, save the power of the transmitting end, the calculation complexity is not high, and the practicality is strong. Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value.

[0114] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A V2X power allocation method, characterized in that: The following steps are involved: The relationship between V2X throughput, QoS constraints and power allocation is mapped from three-dimensional space to two-dimensional plane to obtain the feasible domain of power allocation that maximizes throughput under QoS constraints. For each communication time slot, calculate the optimal power allocation solution corresponding to each power allocation feasible domain; Normalizing the channel state information and the corresponding optimal power allocation solution; Based on the normalized channel state information of a preset number of communication time slots and the corresponding optimal power allocation solution, a DNN model is trained to obtain the optimal solution of V2X power allocation under the characteristics of the channel state information to be predicted based on the trained DNN model; Obtaining the feasible region of power allocation that maximizes throughput under QoS constraints includes the following steps: Assume that the V2I link set is M = {1, ..., M}, the V2V link set is K = {1, ..., K}, and the transmitter uses the maximum transmission power P max , according to the channel gain {h m,B 、h k 、h k,B 、h m,k }, calculate the corresponding maximum receiving / interference power {P max h m,B , P max h k , P max h k,B , P max h m,k }; where h m,B is the signal gain of the mth V2I link, h k is the signal gain of the kth V2V link, h k,B is the interference gain of the mth V2I link, h m,k is the interference gain of the kth V2V link, P max h m,B is the maximum received power of the mth V2I link, P max h k is the maximum received power of the kth V2V link, P max h k,B is the maximum interference power of the mth V2I link, P max h m,k is the maximum interference power of the kth V2V link, where m∈M and k∈K; According to the maximum receiving / interference power, the maximum allowed interference power / minimum allowed receiving power {P * k,B , P * m,k , P * m,B , P * k }; Among them, P * k,B is the maximum allowed interference power corresponding to the mth V2I link, P * m,k is the maximum allowed interference power corresponding to the kth V2V link, P * m,B is the minimum allowed receiving power corresponding to the mth V2I link, P * k is the minimum allowed receiving power corresponding to the kth V2V link; Acquire the power allocation feasible domain according to the maximum allowed interference power / minimum allowed received power; When 0 <P * k,B <P max And P * k <P * k,B When , the power allocation feasible region is above the baseline; When 0 <P * m,k <P max And P * m,B <P * m,k When , the power allocation feasible region is below the baseline; When P * k <P max And P * m,B <P max When , the feasible region of power allocation is located on both sides of the baseline; In other cases, the power allocation feasible domain has no solution; When the power allocation feasible region is above the baseline, the power allocation optimal solution (P k ,P m )=(P * k,B ,P max ); When the power allocation feasible region is below the baseline, the power allocation optimal solution (P k ,P m )=(P max ,P * m,k ); When the power allocation feasible region is located on both sides of the baseline, the optimal power allocation solution (P k ,P m )=arg max(Γ), where Γ={C(P * k ,P max ),C(P max ,P * m,B )}, C(·) represents the V2X system throughput function, P k represents the transmission power of the kth V2V link, P m represents the transmit power of the mth V2I link.

2. The V2X power allocation method according to claim 1, characterized in that: When normalizing the channel state information and the corresponding optimal power allocation solution, (C i -C min ) / (C max -C min ) maps the value to the interval (0,1), where C i The i-th value, C min is the minimum value of the interval, C max The maximum value of the interval.

3. The V2X power allocation method according to claim 1, characterized in that: The preset number is no more than 10.

4. The V2X power allocation method according to claim 1, characterized in that: The input layer of the DNN model includes 4*6 neurons, the three hidden layers are respectively provided with 40, 40, and 50 neurons, and the output layer includes 24*2 neurons.

5. A V2X power distribution system, characterized in that: It includes an acquisition module, a calculation module, a normalization module and an allocation module; The acquisition module is used to map the relationship between the throughput, QoS constraints and power allocation of V2X from a three-dimensional space to a two-dimensional plane, and obtain a feasible domain of power allocation that guarantees maximum throughput under QoS constraints; The calculation module is used to calculate the optimal power allocation solution corresponding to each power allocation feasible domain for each communication time slot; The normalization module is used to normalize the channel state information and the corresponding optimal power allocation solution; The allocation module is used to train a DNN model based on the normalized channel state information under a preset number of communication time slots and the corresponding optimal power allocation solution, so as to obtain the optimal solution of V2X power allocation under the characteristics of the channel state information to be predicted based on the trained DNN model; Obtaining the feasible region of power allocation that maximizes throughput under QoS constraints includes the following steps: Assume that the V2I link set is M = {1, ..., M}, the V2V link set is K = {1, ..., K}, and the transmitter uses the maximum transmission power P max , according to the channel gain {h m,B 、h k 、h k,B 、h m,k }, calculate the corresponding maximum receiving / interference power {P max h m,B , P max h k , P max h k,B , P max h m,k }; where h m,B is the signal gain of the mth V2I link, h k is the signal gain of the kth V2V link, h k,B is the interference gain of the mth V2I link, h m,k is the interference gain of the kth V2V link, P max h m,B is the maximum received power of the mth V2I link, P max h k is the maximum received power of the kth V2V link, P max h k,B is the maximum interference power of the mth V2I link, P max h m,k is the maximum interference power of the kth V2V link, where m∈M and k∈K; According to the maximum receiving / interference power, the maximum allowed interference power / minimum allowed receiving power {P * k,B , P * m,k , P * m,B , P * k }; Among them, P * k,B is the maximum allowed interference power corresponding to the mth V2I link, P * m,k is the maximum allowed interference power corresponding to the kth V2V link, P * m,B is the minimum allowed receiving power corresponding to the mth V2I link, P * k is the minimum allowed receiving power corresponding to the kth V2V link; Acquire the power allocation feasible domain according to the maximum allowed interference power / minimum allowed received power; When 0 <P * k,B <P max And P * k <P * k,B When , the power allocation feasible region is above the baseline; When 0 <P * m,k <P max And P * m,B <P * m,k When , the power allocation feasible region is below the baseline; When P * k <P max And P * m,B <P max When , the feasible region of power allocation is located on both sides of the baseline; In other cases, the power allocation feasible domain has no solution; When the power allocation feasible region is above the baseline, the power allocation optimal solution (P k ,P m )=(P * k,B ,P max ); When the power allocation feasible region is below the baseline, the power allocation optimal solution (P k ,P m )=(P max ,P * m,k ); When the power allocation feasible region is located on both sides of the baseline, the optimal power allocation solution (P k ,P m )=arg max(Γ), where Γ={C(P * k ,P max ),C(P max ,P * m,B )}, C(·) represents the V2X system throughput function, P k represents the transmission power of the kth V2V link, P m represents the transmit power of the mth V2I link.

6. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the V2X power allocation method according to any one of claims 1 to 4 is implemented.

7. A terminal, characterized in that: include: Processor and memory; The memory is used to store computer programs; The processor is used to execute the computer program stored in the memory so that the terminal performs the V2X power allocation method according to any one of claims 1 to 4.

Images