Multi-mode communication agile adaptation method and system

Abstract

The invention discloses a multi-mode communication agile adaptation method and system, and relates to the technical field of automatic driving, and the method comprises the steps: obtaining driving scene information through a multi-mode communication node, and generating multi-mode communication network element information based on the driving scene information; obtaining travel information through the multimode communication vehicle node, and generating communication exchange content information based on the multimode communication network element information and the travel information; the multimode communication vehicle node determines a multimode communication terminal node set based on the communication exchange content information, and determines a service quality parameter set according to the multimode communication terminal node set; generating a communication mode decision variable based on the service quality parameter set and the communication exchange content information, and optimizing the communication mode decision variable to obtain an optimal service quality parameter; communication between multimode communication nodes is realized through the optimal service quality parameters; according to the invention, the problems of low communication efficiency and high communication time delay of the existing multi-mode communication adaptation method are solved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a multi-mode communication agile adaptation method and system. Background Art

[0002] With the development of vehicle networking communication technology, corresponding progress has been made in road safety and driving energy efficiency. In vehicle networking communication applications, the choice of communication mode directly affects the rate, delay, and reliability of data transmission, and thus affects the communication efficiency of the overall communication unit of the vehicle networking. The high dynamics and spatiotemporal complexity of the network topology caused by the high-speed movement of vehicles, the complex and fast-changing wireless propagation environment, the new scientific problems of low latency and high reliability communication under high node density and high communication frequency, are subject to various factors such as driving scenarios, driving intentions, and business needs. Different communication modes have their own advantages and limitations. Therefore, the on-board unit needs to quickly and adaptively select the corresponding communication mode or combination of communication modes according to the driving scenario, driving intention, and business needs.

[0003] At present, the existing multi-mode communication agile adaptation methods cannot deeply analyze the communication terminals and data quality requirements that need to be transmitted by communication data services based on the three major factors of driving scenarios, driving intentions, and business needs, and it is difficult to combine the various communication modes and corresponding communication capabilities of the communication initiator to build a multi-mode communication agile adaptation mechanism for the Internet of Vehicles. It can be seen that the existing multi-mode communication adaptation methods have the problems of low communication efficiency and high communication delay. Summary of the invention

[0004] The present invention provides a multi-mode communication agile adaptation method and system to solve the problems of low communication efficiency and high communication delay in the existing multi-mode communication adaptation method.

[0005] In order to achieve the above object, the present invention is implemented by the following technical solutions:

[0006] In a first aspect, the present invention provides a multimode communication agile adaptation method, the method comprising:

[0007] Determining a multimode communication node in a driving environment, wherein the multimode communication node includes: a multimode communication vehicle node and a multimode communication non-vehicle node;

[0008] Acquiring driving scene information through a multimode communication node, and generating multimode communication network element information based on the driving scene information;

[0009] Acquiring travel information through a multimode communication vehicle node, and generating communication exchange content information based on the multimode communication network element information and the travel information;

[0010] The multimode communication vehicle node determines a set of multimode communication terminal nodes based on the communication exchange content information, and determines a set of quality of service parameters according to the set of multimode communication terminal nodes;

[0011] Generate communication mode decision variables based on the service quality parameter set and the communication exchange content information, and optimize the communication mode decision variables to obtain optimal service quality parameters;

[0012] Communication between multi-mode communication nodes is achieved through optimal quality of service parameters.

[0013] Optionally, determining the multimode communication node in the driving environment includes:

[0014] Determine a multimode communication node in the driving environment, and then the multimode communication node determines all other multimode communication nodes related to the vehicle driving;

[0015] The multi-mode communication nodes are divided into multi-mode communication vehicle nodes and multi-mode communication non-vehicle nodes. The relative position information between the multi-mode communication nodes is determined by acquiring relevant position information with respect to the driving environment.

[0016] Optionally, the acquiring driving scene information through the multi-mode communication node includes:

[0017] Collect the traffic conditions, vehicle status and network status in the current driving environment through communication equipment, perception equipment or information fusion center, and integrate the traffic conditions, vehicle status and network status into driving scene information;

[0018] Generating multi-mode communication network element information based on the driving scenario information includes:

[0019] The multimode communication node determines all communication modes adapted to the multimode communication terminal node set, constructs a communication mode set from all communication modes adapted to the multimode communication node, and uses the communication mode set as multimode communication network element information.

[0020] Optionally, the obtaining travel information through a multi-mode communication vehicle node includes:

[0021] The multi-mode communication vehicle node collects the vehicle driving information at the current moment according to the vehicle communication equipment, the vehicle sensing equipment and the information fusion center, and uses the vehicle driving information as the travel information.

[0022] Optionally, the multimode communication vehicle node determines a set of multimode communication terminal nodes based on the communication exchange content information, including:

[0023] The multimode communication vehicle node determines a set of multimode communication terminal nodes that need to communicate with the multimode communication vehicle node based on the communication exchange content information, multimode communication network element information and driving scenario information. The multimode communication terminal node set includes all multimode communication nodes that need to communicate with the multimode communication vehicle node in the current driving environment.

[0024] Optionally, determining a set of quality of service parameters according to the set of multimode communication terminal nodes includes:

[0025] The delay requirements and reliability requirements when a multi-mode communication vehicle node communicates with a set of multi-mode communication terminal nodes are determined based on the communication exchange content information, and the delay requirements and reliability requirements are integrated into a set of service quality parameters.

[0026] Optionally, the generating a communication mode decision variable based on the service quality parameter set and the communication exchange content information includes:

[0027] The communication mode decision variables are generated according to the multi-mode communication network element information, the service quality parameter set and the communication exchange content information.

[0028] Optionally, optimizing the communication mode decision variable to obtain an optimal quality of service parameter includes:

[0029] The communication mode decision variables are optimized through the corresponding optimization criteria, and the optimization criteria satisfy the following learning model:

[0030]

[0031] Among them, the performance criterion function f(D(i,j),R(i,j),S(i,j,c),F(i,j,c),B(i,j,c),a(i,j,c)) represents the specific performance criterion designed according to the requirements, such as satisfying the transmission delay of all multi-mode communication nodes.

[0032] Optionally, the learning model construction process is as follows:

[0033] The Markov decision model in the multi-mode communication agile adaptation method for the Internet of Vehicles is constructed, including three key elements: state, action, and reward, which are defined as follows:

[0034] State: The state in the Markov decision model reflects the set of supportable communication network elements between the multimode communication node i as an intelligent agent and other multimode communication nodes in the current period. The state set of the multimode communication node i at the current moment is represented by S i :

[0035] S i={(S(i,j,c),F(i,j,c),B(i,j,c))|i,j∈V,c∈C};

[0036] Action: In the constructed Markov decision model, the multi-mode communication node i decides to use a certain communication mode for communication or not to use a certain communication mode for communication at the current moment. A communication mode strategy of the communication node set is an action, and the action space is expressed as:

[0037] A i (s)={a(i,j,c)i,j∈V,c∈C,a(i,j,c)∈{0,1},a(i,j,c)=0};

[0038] Among them, s represents the network status of multi-mode communication vehicle node i at the current moment, A i (s) represents the action space of multi-mode communication vehicle node i in network state s at the current moment;

[0039] Reward: Based on the current State and Action, the reinforcement learning algorithm obtains reward R through evaluation i (s,a),R i (s,a) represents the reward value of multimodal communication node i at the current moment, and a represents the decision action of multimodal communication node i at the current moment, that is, the decision variable a(i,j,c):

[0040] R i (s,a)=f(D(i,j),R(i,j),S(i,j,c),F(i,j,c),B(i,j,c),a(i,j,c)),i∈V;

[0041] It represents the reward obtained by the multimode communication node i according to the corresponding QoS formulation after adopting the current decision action a(i,j,c). The performance criterion function f(D(i,j),R(i,j),S(i,j,c),F(i,j,c),B(i,j,c),a(i,j,c)) is used to quantify the performance of a specific decision action in the communication network;

[0042] The state transition probability P of multimode communication node i i (s′|s,a), its specific definition is expressed as:

[0043] P i (s′|s,a)=P((S′(i,j,c),F′(i,j,c),B′(i,j,c)|S(i,j,c),F(i,j,c),B(i,j,c)),a(i,j,c));

[0044] State transition probability P i(s′|s,a) represents the probability distribution of the state set S at the next moment after the multimode communication node i decides to adopt a certain communication mode at the current moment, S′(i,j,c), F′(i,j,c), B′(i,j,c) respectively represent the support state of the communication mode c available to the multimode communication node i and the multimode communication node j at the next moment, the size of the data to be sent in the buffer of the multimode communication node i and the multimode communication node j under the communication mode c at the next moment, and the bandwidth size of the communication mode available to the multimode communication node i and the multimode communication node j at the next moment;

[0045] The strategy function π of multimodal communication node i i (a|s), its specific definition is expressed as:

[0046] π i (a|s)=Pπ(a(i,j,c)|S(i,j,c),F(i,j,c),B(i,j,c));

[0047] Indicates that when the environment state is s, the multi-mode communication node i selects the mapping of action a, P π Represents the probability of selecting an action. It is used to determine the next decision-making method of the communication mode c between the multimode communication node i and the multimode communication node j according to the state transition probability at the current moment;

[0048] The state value function V of the communication service quality QoS constraint of multi-mode communication node i i (s), whose specific definition is as follows:

[0049]

[0050] It represents the expected return of multi-mode communication node i in state s′, which is used to measure the agile adaptation strategy π when the environment state is s. i (a|s). γ∈[0,1] represents the reward decay coefficient, V i (s) represents the expected reward when the environment state is s′;

[0051] The state-action value function Q(s,a) is specifically defined as:

[0052]

[0053] Q(s,a) estimates the expected reward after taking action a in state s. Its training goal is to complete the exchange of content D while satisfying the QoS constraint. i transmission, and maximize the cumulative reward. a′ represents the decision action of the multi-mode communication node i at the next moment.

[0054] In a second aspect, an embodiment of the present application provides a vehicle multi-mode communication agile adaptation system, including a processor and a memory;

[0055] Memory, used to store computer programs;

[0056] The processor is used to implement any method step described in the first aspect when executing a program stored in the memory.

[0057] Beneficial effects:

[0058] The multimode communication agile adaptation method provided by the present invention obtains driving scene information through a multimode communication node, and generates multimode communication network element information based on the driving scene information; obtains travel information through a multimode communication vehicle node, and generates communication exchange content information based on the multimode communication network element information and the travel information; the multimode communication vehicle node determines a multimode communication terminal node set based on the communication exchange content information, and determines a service quality parameter set according to the multimode communication terminal node set; generates a communication mode decision variable based on the service quality parameter set and the communication exchange content information, and optimizes the communication mode decision variable to obtain an optimal service quality parameter; realizes communication between multimode communication nodes through the optimal service quality parameter; thereby, an agile adaptation mechanism for multimode communication equipment in a vehicle networking environment is specifically constructed based on driving scene-driving intention-business demand elements, and by matching the communication exchange content information, there is no need to overly rely on real-time collection and analysis of scene element information of specific multimode communication nodes, which is conducive to quickly generating service quality parameters and corresponding objective functions, so that it can meet the reasonable update of scene element information for frequently changing external scene information;

[0059] It is worth noting that this method is applicable to the Internet of Vehicles scenario based on vehicle business demand factors as input variables, effectively providing the prediction capability of communication modes based on short-term future decisions beyond traditional algorithms, while avoiding the long-term solution problem of deep learning algorithms on unstable business demand factors. Through agile analysis of driving scenarios, driving intentions, and business needs, the flexibility and real-time performance of the vehicle-mounted multi-mode gateway communication process are improved, so that it can better meet the ever-changing Internet of Vehicles needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A flowchart of a multi-mode communication agile adaptation method according to a preferred embodiment of the present invention;

[0061] Figure 2 A schematic diagram of a multi-mode communication agile adaptation method according to a preferred embodiment of the present invention;

[0062] Figure 3 Schematic diagram of traffic status and vehicle status according to a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0063] The technical solution of the present invention is described clearly and completely below. Obviously, the described embodiments are only a 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 are within the scope of protection of the present invention.

[0064] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the usual meanings understood by persons with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "one" do not indicate quantity restrictions, but indicate the existence of at least one. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship also changes accordingly.

[0065] See also Figure 1-3 , an embodiment of the present application provides a multi-mode communication agile adaptation method, the method comprising:

[0066] Determining a multimode communication node in a driving environment, wherein the multimode communication node includes: a multimode communication vehicle node and a multimode communication non-vehicle node;

[0067] Acquiring driving scene information through a multimode communication node, and generating multimode communication network element information based on the driving scene information;

[0068] Acquiring travel information through a multimode communication vehicle node, and generating communication exchange content information based on the multimode communication network element information and the travel information;

[0069] The multimode communication vehicle node determines a set of multimode communication terminal nodes based on the communication exchange content information, and determines a set of quality of service parameters according to the set of multimode communication terminal nodes;

[0070] Generate communication mode decision variables based on the service quality parameter set and the communication exchange content information, and optimize the communication mode decision variables to obtain optimal service quality parameters;

[0071] Communication between multi-mode communication nodes is achieved through optimal quality of service parameters.

[0072] In the above embodiment, driving scene information is obtained through a multimode communication node, and multimode communication network element information is generated based on the driving scene information; travel information is obtained through a multimode communication vehicle node, and communication exchange content information is generated based on the multimode communication network element information and the travel information; the multimode communication vehicle node determines a set of multimode communication terminal nodes based on the communication exchange content information, and determines a set of service quality parameters based on the set of multimode communication terminal nodes; communication mode decision variables are generated based on the set of service quality parameters and the communication exchange content information, and the communication mode decision variables are optimized to obtain optimal service quality parameters; communication between multimode communication nodes is achieved through the optimal service quality parameters; thereby, an agile adaptation mechanism for multimode communication devices in a vehicle networking environment is specifically constructed based on driving scene-driving intention-business demand elements. By matching the communication exchange content information, there is no need to rely too much on real-time collection and analysis of scene element information of specific multimode communication nodes, which is conducive to quickly generating service quality parameters and corresponding objective functions, so that the scene element information can be reasonably updated for the frequently changing scene information in the outside world.

[0073] Taking a specific example to illustrate, the method specifically includes the following steps:

[0074] Step 1: Construct the communication network element generation module.

[0075] For example, the traffic conditions include: the number of lanes in the current driving scene, the number of multi-mode communication nodes, the vehicle location, the vehicle driving status, etc. Figure 3As shown, the total number of lanes in the current driving scene is 3, and the set of multimode communication nodes that can be collected is i = {1, 2, 3, 4, 5, 6, 7}, among which multimode communication nodes 1 to 5 are vehicle nodes, and all multimode communication vehicle nodes are in a moving (driving) state, and multimode communication nodes 6 and 7 are non-vehicle nodes. Multimode communication vehicle nodes 1 and 2 are located in lane 2, multimode communication vehicle nodes 3 and 4 are located in lane 3, and multimode communication vehicle node 5 is located in lane 1. Multimode communication non-vehicle nodes 6 and 7 are located on both sides of the road (non-lane), and the driving state and lane state of the non-lane multimode communication node information are specified to be 0. The position information G(1), G(2), G(3), G(4), G(5), G(6), G(7) of multimode communication nodes 1, 2, 3, 4, 5, 6, 7 are obtained in real time through GPS, laser radar or camera, and are: (39.8996, 116.4069), (39.8993, 116.4068), (39.8998, 116.4069), (39.8995, 116.4069), (39.8998, 116.4069), (39.9000, 116.4071), (39.8996, 116.4063). For example, Table 1 lists a set of 5 vehicle driving states, represented by numbers 0 to 4. Among them, 0 means the vehicle is stationary, 1 means the vehicle is in normal (constant speed) driving state, 2 means the vehicle is accelerating, 3 means the vehicle is decelerating, and 4 means the vehicle is turning right (or changing lanes to the right). The detailed information of the communication network elements is shown in Table 1:

[0076] Table 1: Traffic status information table of multi-mode communication nodes at the current moment

[0077] i G(i) Driving status Lane 1 (39.8996,116.4069) {4} 2 2 (39.8993,116.4068) {4} 2 3 (39.8998,116.4069) {2} 3 4 (39.8995,116.4069) {1} 3 5 (39.8998,116.4069) {3} 1 6 (39.9000,116.4071) {0} / 7 (39.8996,116.4063) {0} /

[0078] The network status includes: the available communication modes between the multimode communication node i and other nodes, the size of the data to be sent in the buffer at the current moment, and the bandwidth.

[0079] Exemplarily, Table 2 lists a communication set C={1,2,3,4} consisting of four communication modes, where c=1 represents the communication protocol LTE, c=2 represents the communication protocol DSRC, c=3 represents the communication protocol Wi-Fi, and c=4 represents the communication protocol PC5.

[0080] Taking multimode communication vehicle node 1 and multimode communication vehicle node 2 as an example, the communication protocols supported between multimode communication vehicle node 1 and multimode communication vehicle node 2 at the current moment are LTE, DSRC, Wi-Fi, and PC5, and the available communication modes between multimode communication vehicle node 1 and multimode communication vehicle node 2 are expressed as S(1,2,1)=1, S(1,2,2)=1, S(1,2,3)=1, and S(1,2,4)=1. Based on the communication protocols supported between multimode communication vehicle node 1 and multimode communication vehicle node 2, the network status also includes the available buffer and bandwidth at the current moment. For example, the communication protocol LTE supports a channel bandwidth of 20M. Since the LTE communication mode needs to be forwarded through the base station, the bandwidth used for transmission is B(1,2,1)=20 / 2=10M; the communication protocol DSRC does not pass through the base station forwarding, and the supported bandwidth is B(1,2,2)=10M; the communication protocol Wi-Fi does not pass through the base station forwarding, and the supported bandwidth is B(1,2,3)=20M; the communication protocol PC5 does not pass through the base station forwarding, and the supported bandwidth is B(1,2,4)=40M;

[0081] At the current moment, the data sizes to be sent in the buffers of the communication protocols LTE, DSRC, Wi-Fi and PC5 of the multi-mode communication vehicle node 1 are F(1,j,1)=5Kbits, F(1,j,2)=10Kbits, F(1,j,3)=20Kbits, and F(1,j,4)=45Kbits respectively. Taking the multi-mode communication vehicle nodes 1 and 5 as an example, the communication protocols supported between the multi-mode communication vehicle node 1 and the multi-mode communication vehicle node 5 at the current moment are DSRC, Wi-Fi and PC5, then the available communication mode between the multi-mode communication vehicle node 1 and the multi-mode communication vehicle node 2 is represented by S(1,5,2)=1, and the unavailable communication mode between the multi-mode communication vehicle node 1 and the multi-mode communication vehicle node 5 is represented by S(1,5,1)=0. The communication mode support between the multi-mode communication vehicle node 1 and other multi-mode communication nodes is shown in Table 2:

[0082] Table 2: Network status information table of multi-mode communication nodes at the current moment

[0083]

[0084] Step 2: Construct a communication exchange content generation module.

[0085] Taking multi-mode communication vehicle node 1 as an example, based on the information collected by the vehicle communication device or vehicle sensing device at the current moment and the travel information of the multi-mode communication vehicle node, the communication exchange content information of the multi-mode communication vehicle node at the current moment is generated. The communication exchange content constructed by multi-mode communication vehicle node 1 is {change lane to the right}, that is, D1 = {4}.

[0086] Step 3: Construct a communication node set generation module.

[0087] Exemplarily, multimode communication vehicle node 1 determines the multimode communication node set {J(1,j)} through the communication exchange content information D1={4} at the current moment. According to the position of the multimode communication vehicle node 1 at the current moment and the characteristics of the exchange content, the multimode communication vehicle node 1 communicates with nearby multimode communication vehicle nodes and multimode communication non-vehicle nodes. Then J(1,2)=1, J(1,3)=1, J(1,4)=1, J(1,5)=1, J(1,6)=1, J(1,7)=1, indicating that at the current moment, the multimode communication vehicle node 1 needs to send the communication exchange content information D1={4} to multimode communication vehicle nodes 2, 3, 4, 5 and multimode communication non-vehicle nodes 6, 7.

[0088] Step 4: Construct the communication service quality generation module.

[0089] Exemplarily, the multimode communication vehicle node 1 is based on the communication exchange content information D1={4} and the determined communication node set information {J(1,2), J(1,3), J(1,4), J(1,5), J(1,6), J(1,7)}. At the current moment, the communication exchange content generated by the multimode communication vehicle node 1 mainly focuses on the delay requirement of the communication process. Therefore, only the delay is used as the satisfaction target to generate the corresponding QoS:

[0090] Exemplarily, D(1,2)=7ms, D(1,3)=5ms, D(1,4)=5ms, D(1,5)=6ms, D(1,6)=4ms, and D(1,7)=6ms, respectively representing the maximum acceptable transmission delays of the decision-making communications of the multi-mode communication vehicle nodes (1,2), (1,3), (1,4), (1,5), (1,6), and (1,7) at the current moment;

[0091] Step 5: Build a communication mode agile adaptation module.

[0092] Exemplarily, the multimode communication vehicle node 1 constructs the state space information S according to the contents of Table 1 and Table 2 in step 1. In the agile adaptation scenario, the space information is specifically reflected as the delay under a certain communication mode. Exemplarily, the objective function f(D(i,j), R(i,j), S(i,j,c), F(i,j,c), B(i,j,c), a(i,j,c)) is specifically to meet the transmission delay QoS minimization, then the cumulative delay τ of the data size to be sent in the buffer is calculated by formula (1), as shown below:

[0093]

[0094] Among them, λ cIndicates the number of paths that a communication node passes through when transmitting information in different communication modes. When the communication mode is LTE or NR, that is, cellular communication, λ c =2, other communication modes λ c =1. For example, for multimode communication vehicle node 1, there is only data size Db(1,j)=5Kbits corresponding to the exchange content information D1 sent to multimode communication node j. Multimode communication vehicle node 1 selects different communication modes to communicate with other nodes as actions. The accumulated delays of different actions and the accumulated delays of the data size to be sent in the buffer of multimode communication vehicle node 1 under different communication modes are based on formula (1). The specific values ​​are shown in Table 3:

[0095] Table 3: Cache size, bandwidth and initialization delay of different communication modes of multi-mode communication node 1

[0096] Communication Mode Index Communication mode cache size bandwidth Cache accumulated delay The accumulated delay of exchanging content 1 5K 10M 0.5ms 0.5ms 2 10K 10M 1ms 0.5ms 3 20K 20M 1ms 0.25ms 4 45K 40M 1.125ms 0.125ms

[0097] Based on the agile adaptation scenario requirements and constraints, construct the reward function R i (s,a):

[0098]

[0099] In formula (2), T i Indicates that the communication service D is completed between the multi-mode communication vehicle node i and the multi-mode communication node j i After the delay, when T i When the delay is greater than the time constraint, the reward is negative infinity. At this time, the communication mode maintains the previous state and no longer updates the subsequent state. i When the delay is less than the time constraint, the reward is the maximum delay sum max(T i ) is different from the next state action delay T i The difference of ′.

[0100]

[0101] The objective of the model is the same as that shown in formula (3), which is to minimize the number of multi-mode communication vehicle nodes Final delay in completing the business , The size of the multi-mode communication vehicle node With other multi-mode communication nodes The maximum value of the delay to complete the service is determined by

[0102]

[0103] According to the description of the basic elements of the above model, the initialization state space is shown in state 0 of Table 4. The state values ​​of different communication modes of multi-mode communication vehicle node 1 are all 0. At this time, the multi-mode communication vehicle node is selected for communication. First, a random strategy is used to select the node. The adaptation strategy from state 1 to state 6 is shown in formula (4):

[0104]

[0105] Among them, Cn in the communication node set represents the number of multi-mode communication nodes that have not been agilely adapted, and does not include nodes that are not supported by the communication mode. i (s) Calculate the state value of communication mode 1 (state 1), as shown in formula (5):

[0106]

[0107] Among them, the adaptation strategy is random selection, then there is only one strategy, taking multi-mode communication vehicle node 2 and multi-mode communication vehicle node 5 as an example, there are Initial Phase V i (s′)=0, and the discount factor γ=1. According to steps 1 to 3, the nodes that need to communicate and the supported communication modes are determined. Therefore, the scenario of agile adaptation is known, and the constructed state transition probability matrix P i (s′|s,a)=1. Taking communication mode c=1 as an example, any multimode communication node that selects communication mode 1 in state 1 will receive a reward R i (s,a)=1.125-1=0.125. Similarly, the state values ​​of state 1 of communication mode 2, communication mode 3 and communication mode 4 can be calculated, which are -0.375, -0.125 and -0.125 respectively. The state is updated according to the calculation process shown in formula (5), as shown in Table 4:

[0108] Based on state 1, update state 2 of mode 1 and randomly remove a node that has completed communication, then we have formula (6):

[0109]

[0110] Among them, V i (s′) is updated in state 1 and applied to the calculation process of formula (5). According to the above process, states 3 to 6 are updated as shown in Table 4:

[0111] Table 4: State value corresponding to state space

[0112] Communication Modec c＝1 c＝2 c＝3 c＝4 State 0 0 0 0 0 State 1 0.125 -0.375 -0.125 -0.125 State 2 -0.1 -0.6 -0.1 0.025 State 3 -0.194 -0.694 0.056 0.306 State 4 -0.175 -0.675 0.325 0.7 State 5 0.088 -0.413 0.838 1.3375 State 6 1.8505 1.3505 2.8505 3.4755

[0113] After calculation, the state value of multimode communication node 1 in all states is obtained, but the strategy is still consistent with the traditional adaptation scheme at this time, and the best matching scheme is not obtained. In order to obtain the best agile adaptation scheme adjustment strategy, the Q(s,a) values ​​corresponding to different actions can be easily calculated based on the state-action value function. The optimal agile adaptation scheme can be obtained by using the action with the largest Q(s,a) value in each state through the greedy strategy. Taking state 0 as an example, the value Q(s,a) of action a in state 0 is calculated. Based on the greedy strategy, node 6 with the smallest delay constraint and supporting communication mode 1 is first considered, and its state-action value is calculated as shown in formulas (7) to (10):

[0114]

[0115]

[0116]

[0117]

[0118] The value of action a(1,6,1) is the largest among the above state-action values, so the updated agile adaptation strategy is {a(1,6,1)=1}. At this time, mode 1 is updated to state 1, and modes 2, 3, and 4 remain in state 0. The maximum delay max(T i )=1.5ms, and the total delay is 1ms. Based on state 1, multimode communication node 3 and multimode communication node 4 with the same constraints are selected according to the delay constraint, and one of them is randomly selected to calculate different state-action values, as shown in formulas (11) to (14):

[0119]

[0120]

[0121]

[0122]

[0123] At this time, the reward of mode 3 is the largest, so mode 3 is selected. At the same time, mode 3 enters state 1, mode 1 maintains state 1, and modes 2 and 4 maintain state 0. At this time, the total delay is 1.25ms, and the updated agile adaptation strategy is {a(1,6,1)=1,a(1,3,3)=1}. According to the above process, the state-action value is calculated in turn, as shown in Table 5:

[0124] Table 5: State-Action Value Table

[0125] Communication Modec c＝1 c＝2 c＝3 c＝4 Optimal a(i,j,c) Total delay (ms) State 0 0 0 0 0 / / State 1 0.125 -0.375 -0.125 -0.125 a(1,6,1)＝1 1 State 2 0.125 0.125 0.675 0.5 a(1,3,3)＝1 1.25 State 3 0.8 0.8 0.8 1.05 a(1,4,4)＝1 1.25 State 4 1.85 1.85 1.85 1.975 a(1,5,4)＝1 1.375 State 5 1.975 1.975 1.475 1.975 a(1,7,4)＝1 1.5 State 6 1.975 1.975 1.975 1.85 a(1,2,2)＝1 1.5

[0126] When the state action values ​​in Table 5 are the same, they are randomly selected according to the communication mode supported by the node. Combining Table 5, it can be obtained that {a(1,6,1)=1,a(1,3,3)=1,a(1,4,4)=1,a(1,5,4)=1,a(1,7,4)=1,a(1,2,2)=1} is the optimal agile adaptation strategy with a total delay of 1.5ms. Compared with the random strategy, the maximum delay can be reduced by 3.5ms. The detailed strategy is shown in Table 6:

[0127] Table 6: Best Agile Adaptation Strategy Table

[0128] Communication Modec c＝1 c＝2 c＝3 c＝4 Optimal a(i,j,c) Multimode communication vehicle node 2 0 1 0 0 a(1,2,2)＝1 Multimode communication vehicle node 3 0 0 1 0 a(1,3,3)＝1 Multimode communication vehicle node 4 0 0 0 1 a(1,4,4)＝1 Multimode communication vehicle node 5 0 0 0 1 a(1,5,4)＝1 Multimode communication non-vehicle node 6 1 0 0 0 a(1,6,1)＝1 Multimode communication non-vehicle node 7 0 0 0 1 a(1,7,4)＝1

[0129] It can be seen from Table 6 that the decision communication mode of multimode communication vehicle node 1 and multimode communication vehicle node 2 is DSRC, the decision communication mode of multimode communication vehicle node 1 and multimode communication vehicle node 3 is Wi-Fi, the decision communication mode of multimode communication vehicle node 1 and multimode communication vehicle node 4 is PC5, the decision communication mode of multimode communication vehicle node 1 and multimode communication vehicle node 5 is PC5, the decision communication mode of multimode communication vehicle node 1 and multimode communication non-vehicle node 6 is LTE, and the decision communication mode of multimode communication vehicle node 1 and multimode communication non-vehicle node 7 is PC5.

[0130] The embodiment of the present application also provides a vehicle multi-mode communication agile adaptation system, including a processor and a memory;

[0131] Memory, used to store computer programs;

[0132] The processor is used to implement any method step in the multi-mode communication agile adaptation method when executing the program stored in the memory.

[0133] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A multi-mode communication agile adaptation method, characterized in that: The method comprises: Determining a multimode communication node in a driving environment, wherein the multimode communication node includes: a multimode communication vehicle node and a multimode communication non-vehicle node; Acquiring driving scene information through a multimode communication node, and generating multimode communication network element information based on the driving scene information; Acquiring travel information through a multimode communication vehicle node, and generating communication exchange content information based on the multimode communication network element information and the travel information; The multimode communication vehicle node determines a set of multimode communication terminal nodes based on the communication exchange content information, and determines a set of quality of service parameters according to the set of multimode communication terminal nodes; Generate communication mode decision variables based on the service quality parameter set and the communication exchange content information, and optimize the communication mode decision variables to obtain optimal service quality parameters; Communication between multi-mode communication nodes is achieved through optimal quality of service parameters.

2. The multi-mode communication agile adaptation method according to claim 1, characterized in that: The determining of the multi-mode communication nodes in the driving environment comprises: Determine all communication devices related to vehicle driving in the driving environment, and determine all communication devices as multi-mode communication nodes; By acquiring relevant position information of the multimode communication nodes and the driving environment, the multimode communication nodes are divided into multimode communication vehicle nodes and multimode communication non-vehicle nodes.

3. The multi-mode communication agile adaptation method according to claim 1, characterized in that: The obtaining of driving scene information through the multi-mode communication node includes: Collect the traffic conditions, vehicle status and network status in the current driving environment through communication equipment, perception equipment or information fusion center, and integrate the traffic conditions, vehicle status and network status into driving scene information; Generating multi-mode communication network element information based on the driving scenario information includes: The multimode communication node determines all communication modes adapted to other multimode communication terminal nodes, constructs a communication mode set from all communication modes adapted to the multimode communication node, and uses the communication mode set as multimode communication network element information.

4. The multi-mode communication agile adaptation method according to claim 1, characterized in that: The obtaining of travel information through a multi-mode communication vehicle node includes: The multi-mode communication vehicle node collects the vehicle driving information at the current moment according to the vehicle communication equipment, the vehicle sensing equipment and the information fusion center, and uses the vehicle driving information as the travel information.

5. The multi-mode communication agile adaptation method according to claim 1, characterized in that: The multimode communication vehicle node determines a set of multimode communication terminal nodes based on the communication exchange content information, including: The multimode communication vehicle node determines a set of multimode communication terminal nodes that need to communicate with the multimode communication vehicle node based on the communication exchange content information, multimode communication network element information and driving scenario information. The multimode communication terminal node set includes all multimode communication nodes that need to communicate with the multimode communication vehicle node in the current driving environment.

6. The multi-mode communication agile adaptation method according to claim 1, characterized in that: The determining of a set of quality of service parameters according to the set of multimode communication terminal nodes comprises: The delay requirements and reliability requirements when a multi-mode communication vehicle node communicates with a set of multi-mode communication terminal nodes are determined based on the communication exchange content information, and the delay requirements and reliability requirements are integrated into a set of service quality parameters.

7. The multi-mode communication agile adaptation method according to claim 1, characterized in that: The generating of the communication mode decision variable based on the service quality parameter set and the communication exchange content information comprises: The communication mode decision variables are generated according to the multi-mode communication network element information, the service quality parameter set and the communication exchange content information.

8. The multi-mode communication agile adaptation method according to claim 1, characterized in that: The step of optimizing the communication mode decision variables to obtain optimal service quality parameters includes: The communication mode decision variables are optimized through the corresponding optimization criteria, and the optimization criteria satisfy the following learning model: f(D(i,j),R(i,j),S(i,j,c),F(i,j,c),B(i,j,c),a(i,j,c)) a(i,j,c)∈{0,1} Among them, the performance criterion function f(D(i,j),R(i,j),S(i,j,c),F(i,j,c),B(i,j,c),a(i,j,c)) represents the specific performance criterion designed according to the requirements, such as satisfying the transmission delay of all multi-mode communication nodes.

9. The multi-mode communication agile adaptation method according to claim 8, characterized in that: The construction process of the learning model is as follows: The Markov decision model in the multi-mode communication agile adaptation method for the Internet of Vehicles is constructed, including three key elements: state, action, and reward, which are defined as follows: State: The state in the Markov decision model reflects the set of supportable communication network elements between the multimode communication node i as an intelligent agent and other multimode communication nodes in the current period. The state set of the multimode communication node i at the current moment is represented by S i : S i ={(S(i,j,c),F(i,j,c),B(i,j,c))|i,j∈V,c∈C}; Action: In the constructed Markov decision model, the multi-mode communication node i decides to use a certain communication mode for communication or not to use a certain communication mode for communication at the current moment. A communication mode strategy of the communication node set is an action, and the action space is expressed as: A i (s)={a(i,j,c)i,j∈V,c∈C,a(i,j,c)∈{0,1},a(i,j,c)=0}; Among them, s represents the network status of multi-mode communication vehicle node i at the current moment, A i (s) represents the action space of multi-mode communication vehicle node i in network state s at the current moment; Reward: Based on the current State and Action, the reinforcement learning algorithm obtains reward R through evaluation i (s,a),R i (s,a) represents the reward value of multimodal communication node i at the current moment, and a represents the decision action of multimodal communication node i at the current moment, that is, the decision variable a(i,j,c): R i (s,a)=f(D(i,j),R(i,j),S(i,j,c),F(i,j,c),B(i,j,c),a(i,j,c)),i∈V; It represents the reward obtained by the multimode communication node i according to the corresponding QoS formulation after adopting the current decision action a(i,j,c). The performance criterion function f(D(i,j),R(i,j),S(i,j,c),F(i,j,c),B(i,j,c),a(i,j,c)) is used to quantify the performance of a specific decision action in the communication network; The state transition probability P of multimode communication node i i (s′|s,a), its specific definition is expressed as: P i (s′|s,a)=P((S′(i,j,c),F′(i,j,c),B′(i,j,c)|S(i,j,c),F(i,j,c),B(i,j,c)),a(i,j,c)); State transition probability P i (s′|s,a) represents the probability distribution of the state set S at the next moment after the multimode communication node i decides to adopt a certain communication mode at the current moment, S′(i,j,c), F′(i,j,c), B′(i,j,c) respectively represent the support state of the communication mode c available to the multimode communication node i and the multimode communication node j at the next moment, the size of the data to be sent in the buffer of the multimode communication node i and the multimode communication node j under the communication mode c at the next moment, and the bandwidth size of the communication mode available to the multimode communication node i and the multimode communication node j at the next moment; The strategy function π of multimodal communication node i i (a|s), its specific definition is expressed as: π i (a|s)=Pπ(a(i,j,c)|S(i,j,c),F(i,j,c),B(i,j,c)); Indicates that when the environment state is s, the multi-mode communication node i selects the mapping of action a, P π Represents the probability of selecting an action. It is used to determine the next decision-making method of the communication mode c between the multimode communication node i and the multimode communication node j according to the state transition probability at the current moment; The state value function V of the communication service quality QoS constraint of multi-mode communication node i i (s), whose specific definition is as follows: It represents the expected return of multi-mode communication node i in state s′, which is used to measure the agile adaptation strategy π when the environment state is s. i (a|s). γ∈[0,1] represents the reward decay coefficient, V i (s′) represents the expected reward when the environment state is s′; The state-action value function Q(s,a) is specifically defined as: Q(s,a) estimates the expected reward after taking action a in state s. Its training goal is to complete the exchange of content D while satisfying the QoS constraint. i transmission, and maximize the cumulative reward. a′ represents the decision action of multi-mode communication node i at the next moment.

10. A multi-mode communication agile adaptation system, characterized in that: Including processor and memory; Memory, used to store computer programs; A processor, for implementing the method steps described in any one of claims 1 to 9 when executing a program stored in a memory.

Images