Bidirectional dynamic adjustment method and system for networking interception stage of emergency communication sidelink

Abstract

The invention discloses a bidirectional dynamic adjustment method and system for a networking interception stage of an emergency communication sidelink, and belongs to the field of transmission of a wireless communication system. The bidirectional dynamic adjustment method for the networking interception stage of the emergency communication sidelink comprises the following steps: dynamically adjusting a receiving gain threshold value of a receiving end after a node which is not accessed to a network is started; and when the receiving end dynamically adjusts the receiving gain threshold value, the sending power of the sending end is dynamically adjusted based on the MADDPG algorithm. For nodes which are not accessed to the network, the nodes can quickly adapt to the real-time change of a network topology structure by dynamically adjusting a receiving gain threshold value in a process of monitoring a control sub-frame; meanwhile, the network-accessed node dynamically adjusts the sending power according to the real-time change of the current active node so as to optimize the network performance.

Description

Technical Field

[0001] The present invention belongs to the transmission field of wireless communication systems, and in particular relates to a two-way dynamic adjustment method and system in an emergency communication sidelink networking listening phase. Background Art

[0002] Considering the diversity of locations, times and natures of emergencies, emergency communications face many challenges in emergency preparedness and response. In emergency situations, the coverage provided by traditional network infrastructure may be missing or insufficient, so the ability for user equipment (UE) to connect directly to each other becomes critical.

[0003] In the field of emergency communications, sidelink communication technology allows devices to communicate directly without going through a base station, forming a self-organizing network, which mainly serves the information exchange between user terminals. It provides users with a direct communication method without the need for transit through network equipment, and supports multiple data transmission methods such as unicast, multicast and broadcast. A common method is that the UE can reuse the uplink spectrum resources of the authorized frequency band, so as not to occupy additional resources, greatly improve the utilization efficiency and throughput of spectrum resources, and increase network capacity at the same time, so as to ensure that the operation of the communication network is more flexible and efficient, and reduce the load pressure of the base station.

[0004] There are two common modes for implementing sidelink networking technology: direct communication mode and relay communication mode. In direct communication mode, UEs directly establish connections and communicate with each other without forwarding through a central node; in relay communication mode, devices without network coverage use devices in network coverage as a springboard to connect to other devices through relay communication. Compared with sidelink communications under cellular networks or partial cellular network coverage, sidelink communications without cellular network coverage are more suitable for emergency scenarios. In this case, device discovery and session establishment are completed autonomously between devices without the participation of base stations, so the complexity is higher. At the same time, the terminal device has the function of acting as a relay node to forward messages. Devices without cellular network coverage can forward and transmit messages through intermediate devices to achieve the process of accessing the network.

[0005] In the networking stage under the sideline communication link, after the non-networked nodes are started, they need to go through a series of listening steps before they can access the network. These nodes receive the control subframes periodically broadcast by other networked nodes through listening to several wireless frames, and store them locally. After the listening is completed, the non-networked nodes select the parent node according to the networking information field carried by the control subframe to further perform network access operations.

[0006] The current common technology usually adopts a fixed receiving gain threshold in the listening stage as a requirement for whether the non-networked node can receive the control subframe. This technical means is simple and easy to implement, but in actual communication scenarios, due to the randomness of the node position, the setting of a fixed receiving gain threshold often brings difficulties to the network access process of the non-networked node to a certain extent. Specifically, if the starting position of the node is too marginal and far away from other nodes that have been networked or not, setting a larger fixed receiving gain threshold is likely to make the edge node unable to listen to any control subframe from beginning to end, resulting in network access failure. For example, in the high density state of non-networked nodes, the density of networked nodes is relatively low. At this time, although setting a smaller fixed receiving gain threshold can expand the listening range, it may enable the non-networked node to be able to listen to the control subframe sent by the networked node that is actually far away, and then select a non-optimal parent node or increase the complexity of parent node selection due to factors such as the number of communication hops. Considering the dynamic changes in wireless communication conditions, the selection of such a non-optimal parent node may result in poor communication quality or even the risk of communication interruption in the subsequent access process. At the same time, it is more likely that nodes not connected to the network will choose the same parent node, causing communication congestion and mutual interference between nodes.

[0007] In addition, the nodes that have joined the network periodically and continuously broadcast control subframes with a fixed transmission power (for example, in order to allow more users to listen, the maximum power is used), which also hinders the optimization of the network topology to a certain extent. Especially when the nodes that have not joined the network are densely distributed, this fixed transmission power broadcasting method is also one of the reasons why the nodes that have not joined the network can easily listen to non-preferred parent nodes, thereby aggravating the situation where the nodes that have not joined the network concentrate on selecting the same parent node, causing communication congestion and interference between nodes. Summary of the invention

[0008] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a method and system for bidirectional dynamic adjustment of emergency communication sidelink networking listening phase, which solves the problems in the prior art.

[0009] The purpose of the present invention can be achieved through the following technical solutions:

[0010] The method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase comprises the following steps:

[0011] After the node not connected to the network is powered on, the receiving gain threshold of the receiving end is dynamically adjusted;

[0012] While the receiving end dynamically adjusts the receiving gain threshold, the transmitting end's transmitting power is dynamically adjusted based on the MADDPG algorithm.

[0013] Furthermore, the step of dynamically adjusting the receiving gain threshold of the receiving end includes:

[0014] S11, set an initial receiving gain threshold Thres init ;

[0015] S12, after the node not connected to the network is powered on, the initial receiving gain threshold Threshold init Enter the initial listening process, set and start the listening timer, continue to receive control subframes until the timer overflows, and evaluate the existing local networking information after the end;

[0016] S13, in the listening and waiting process, a relative listening and waiting time is introduced to dynamically reduce the receiving gain threshold.

[0017] Furthermore, when the relative listening waiting time is introduced to reduce the receiving gain threshold, a linear adjustment or an exponential adjustment method is adopted;

[0018] The expression for linear adjustment of the receive gain threshold is:

[0019] Thres listen =Thres init +l·ΔT waiting

[0020] Among them, Thres listen is the receiving gain threshold that changes dynamically during the listening waiting process, ΔT waiting is the relative listening waiting time, which is positively correlated with the number of cycles of the listening waiting period, and l is a negative adjustment coefficient;

[0021] The expression for adjusting the receiving gain threshold in exponential form is:

[0022]

[0023] Among them, θ is the positive adjustment coefficient.

[0024] Furthermore, if during the dynamic adjustment process, the number of listening and waiting cycles exceeds the limit N Timer_Limit Or the receiving gain threshold is adjusted below the lower limit Thres Lower_limit , the dynamic adjustment is terminated, and the non-networked nodes that failed to successfully access the network will be used as the initial nodes to initiate networking.

[0025] Furthermore, when the transmitter dynamically adjusts the transmission power, in the MADDPG algorithm, first, the central node and each relay node are defined as an agent respectively; π={π1,π2,...,π N} to represent the N strategies that the agent can adopt, π i ={a i |o i} indicates that agent i is observing o iThe action a obtained i , with x={o1,o2,...,o N} to represent the information of the entire environment; each interaction of the agent will be stored in the experience replay pool D, and data will be obtained in batches from it to update and learn the network parameters;

[0026] Each of the above intelligent agents uses a policy-based Actor network and a value-based Critic network; the Actor network collects state information from the environment and uses the policy π i Perform the corresponding action a i , the Critic network uses the state-action value function Q to evaluate the action a selected by the Actor network i The degree of quality.

[0027] Furthermore, the step of dynamically adjusting the transmit power of the transmitter based on the MADDPG algorithm includes:

[0028] S21, initialize the experience replay pool;

[0029] S22, initialize the random process N noise ;

[0030] S23, initialize the network parameters of Actor and Critic,

[0031] S24, setting the experience learning batch size, number of iterations and the step size of each iteration;

[0032] S25, in each iteration, initialize the state s of the current agent t , select a t , the agent is based on the current state s t and action a t Perform state transition and reach the next state s t+1 , and calculate the reward r t ;

[0033] S26, the current data {s t ,a t ,r t ,s t+1}Store it in the experience replay pool, and take out N groups of data from it to update the network parameters;

[0034] S27, updates the Actor network gradient and the Critic network's Loss function, thereby updating the Actor and Critic network parameters and

[0035] S28, soft update target Actor network parameters and target critic network parameters End this round of iteration;

[0036] S29, repeat S25 to S28 until convergence.

[0037] The two-way dynamic adjustment system of the emergency communication side link networking listening phase includes:

[0038] Receiver gain adjustment module: After the node not connected to the network is powered on, dynamically adjust the receiver gain threshold;

[0039] And, the transmitting end power adjustment module: while the receiving end dynamically adjusts the receiving gain threshold, the transmitting end's transmitting power is dynamically adjusted based on the MADDPG algorithm.

[0040] A computer storage medium stores a readable program, which can execute the above-mentioned two-way dynamic adjustment method of the emergency communication side link networking listening phase when the program is running.

[0041] An electronic device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other through the communication bus;

[0042] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned two-way dynamic adjustment method in the emergency communication side link networking listening phase.

[0043] A computer program product includes computer instructions, wherein the computer instructions instruct a computing device to execute operations corresponding to the above-mentioned method for bidirectional dynamic adjustment of the emergency communication side link networking listening phase.

[0044] Beneficial effects of the present invention:

[0045] 1. The present invention designs a bidirectional dynamic adjustment method for the sidelink networking listening stage in emergency communication. For nodes that have not yet joined the network, the receiving gain threshold is dynamically adjusted during the process of listening to the control subframe, so that they can quickly adapt to the real-time changes of the network topology. At the same time, the nodes that have joined the network dynamically adjust the transmission power according to the real-time changes of the current active nodes to optimize the network performance.

[0046] 2. Through the two-way dynamic adjustment method of the present invention, after the listening is completed, the node that has not joined the network can be evaluated based on the optimization information stored locally, so as to select the best access node and achieve fast and efficient network access. This can not only ensure the communication quality, but also greatly improve the flexibility and robustness of node network access, providing a more reliable and efficient solution for emergency communication networking. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1 It is a schematic diagram of the network interception related process of the present invention;

[0049] Figure 2 is a schematic diagram of a communication scenario related to the initial receiving gain threshold setting of the present invention;

[0050] Figure 3 It is a flow chart of initial listening of networking of the present invention;

[0051] Figure 4 It is a signaling interaction diagram between a non-networked node and a networked node of the present invention;

[0052] Figure 5 It is a listening and waiting flow chart of the present invention;

[0053] Figure 6 It is the Actor and Critic neural network structure diagram of the present invention;

[0054] Figure 7 It is a schematic diagram of the structure of the MADDPG algorithm of the present invention;

[0055] Figure 8 It is a schematic diagram of the transmission power of the networked nodes before and after the dynamic adjustment of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0057] like Figure 1 As shown in the figure, after a non-networked node is powered on, it first executes the initial listening process; if the broadcast control subframe is successfully listened to, key information such as the broadcast node ID, signal strength and access hop count are extracted and stored; if the initial listening process does not obtain any result, it enters the listening waiting process and continues to listen until a signal is received; after the listening is completed, the non-networked node evaluates the local information and selects the parent node to send subsequent information such as the network access request to ensure rapid integration into the current network infrastructure.

[0058] The method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase comprises the following steps:

[0059] S1, after the node not connected to the network is powered on, dynamically adjust the receiving gain threshold of the receiving end;

[0060] The steps of dynamically adjusting the receiving gain threshold of the node not in the network specifically include:

[0061] S11, set an initial receiving gain threshold Thres init ;

[0062] like Figure 2 As shown in the figure, the relevant communication scenarios of the initial receiving gain threshold setting are shown; among them, the black nodes represent the central nodes, the gray nodes represent the nodes that have been connected to the network and can be relayed, and the white nodes represent the nodes that are not connected to the network; Figure 2 As shown in (a) in the figure, for nodes that have not yet joined the network, if the initial listening threshold is set too low, although it can increase the possibility of the node receiving the control subframe in the initial listening period, the node may capture a weak signal, and the parent node corresponding to such a signal is more likely to cause unstable communication quality or even direct interruption during the subsequent access process due to distance, environmental noise or interference from surrounding nodes. On the contrary, if Figure 2 As shown in (b) in Figure 1, if the initial threshold is set too high, the node may not be able to receive the signal at the appropriate distance, thereby increasing the number of unnecessary relays and increasing the complexity and time consumption of communication. Taking the above factors into consideration, in order to ensure the stability and efficiency of communication, a moderate initial receiving gain threshold Thres is selected. init It is of great necessity to ensure the smooth progress of the initial listening process.

[0063] S12, the node not in the network receives the initial gain threshold Thres init Enter the initial listening process;

[0064] like Figure 3 As shown, the initial listening process specifically includes the following steps:

[0065] S121, setting and starting the initial listening timer;

[0066] S122, continue to receive the control subframe, and if the receiving gain of the control subframe exceeds the current threshold, store the relevant information locally;

[0067] S123, if the timer overflows, the initial listening ends. If there is a networking information field locally, it is evaluated, and the networked node corresponding to the broadcast node ID of the field is selected as the parent node, and a network access request is sent for subsequent network access procedures; if there is no information, enter S13, that is, the listening and waiting process.

[0068] S13, in the listening and waiting process, a relative listening and waiting time is introduced to dynamically reduce the receiving gain threshold;

[0069] like Figure 5 As shown in the figure, in the listening waiting process, the listening waiting timer is set in the hope that the non-networked nodes can screen the control subframes that are intercepted. If the waiting time is long, it means that the range of the receiver's detection signal is not ideal at this time, and it is difficult to receive the broadcast of other networked nodes. It should be dynamically adjusted based on the initial receiving gain threshold. Introduce an item related to the listening waiting time to reduce the receiving gain threshold;

[0070] Two calculation formulas for reducing the receiving gain threshold are provided below:

[0071] a) Linear adjustment threshold:

[0072] Thres listen =Thres init +l·ΔT waiting (1)

[0073] Among them, Thres listen is the receiving gain threshold that changes dynamically during the listening waiting process, ΔT waiting is the relative listening waiting time, which is positively correlated with the number of cycles of the listening waiting period, and l is a negative adjustment coefficient used to adjust ΔT waiting The degree of linear impact on the reduction of the receive gain threshold.

[0074] The linear adjustment threshold provides a stable and gradual adjustment strategy. At the same time, the strategy is simple, convenient for system design and implementation, and also easy for debugging.

[0075] b) Adjust the threshold exponentially:

[0076]

[0077] Where e is the base of the natural logarithm and θ is the positive adjustment coefficient.

[0078] The exponential adjustment strategy provides faster adaptability and can quickly adapt to environmental changes when network quality deteriorates. At the same time, the threshold adjustment is more detailed, ensuring that the network recovers to a controllable communication quality level in a short time. However, although the linear adjustment strategy is inferior to the exponential adjustment strategy in flexibility and dynamism, it is simple, stable, low-cost and easy to control, which makes the linear adjustment strategy still have advantages.

[0079] like Figure 5 As shown in the figure, the specific steps of the listening and waiting process are:

[0080] S131, setting a listening waiting timer;

[0081] S132, linearly or exponentially adjust the receiving gain threshold. If the threshold adjustment lower limit Threshold is not reached, Lower_limit , then go to step S133; if the threshold adjustment lower limit Thres is reached Lower_limit , then the listening and waiting ends, and the node that has not joined the network initiates networking as the initial node;

[0082] S133, continue to receive the control subframe, and if the receiving gain of the control subframe exceeds the current threshold, store the relevant information locally;

[0083] S134, if the timer overflows, determine whether there is any networking related information stored locally. If so, the listening and waiting ends, the stored information is evaluated, and the non-networked node performs the subsequent network access process; if not, proceed to step S135;

[0084] S135, determine whether the current listening waiting cycle limit N is exceeded Timer_Limit If yes, the listening and waiting ends, and the node not in the network initiates networking as the initial node; if no, repeat S131-S134 and adjust the dynamic threshold until the control subframe is received or the listening and waiting end condition is reached.

[0085] Figure 4 The signaling interaction in the networking listening stage of the present invention is demonstrated. The networked nodes periodically broadcast control subframes to the surroundings. If the non-networked nodes receive the control subframes during the initial listening stage, a network access request is sent to the corresponding networked nodes according to the instructions of the control subframes. If the non-networked nodes do not successfully receive any control subframes during the initial listening stage, the listening wait stage is entered. If a control subframe is received during this stage, a network access request is sent to the corresponding networked nodes as described above. The receiving gain threshold is dynamically adjusted to quickly adapt to network changes, so that the non-networked nodes can access the appropriate parent node as quickly as possible to carry out the subsequent networking process, thereby achieving rapid networking. The receiving gain threshold setting of each non-networked node after powering on has personalized characteristics, and the receiving gain thresholds between different nodes are independent of each other to adapt to different environments.

[0086] S2, while the receiving end dynamically adjusts the receiving gain threshold, the transmitting end's transmit power is dynamically adjusted based on the MADDPG algorithm;

[0087] While the receiving end dynamically adjusts the gain threshold, the transmitting end adjusts the transmission power in real time to ensure the stability and efficiency of communication. In order to realize the continuous power adjustment strategy, the present invention uses the Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm in the dynamic power adjustment. Through centralized training and distributed execution, in the actual communication scenario, the transmitting end can achieve the optimal adjustment of the transmission power based on the local state information.

[0088] For a system with N agents, define its state space, action space, and reward function.

[0089] State space: The total state space at step t is:

[0090]

[0091] In the formula, Represents the state of the i-th agent in the t-step system.

[0092] For each agent,

[0093]

[0094] In the formula, is the number of currently active nodes with i as parent node, α is the path loss factor, d i represents the distance between parent node i and its child nodes, h i Indicates the channel gain.

[0095] Nodes that have sent network access requests to their parent nodes and have been received by their parent nodes but have not yet officially joined the network are considered active nodes; these active nodes will continue to occupy the communication resources of their parent nodes. When a parent node handles multiple active nodes at the same time, communication congestion is unavoidable. At the same time, other relay nodes may be in a relatively idle state. This imbalance in resource allocation not only affects communication efficiency, but also prolongs networking time.

[0096] Action Space:

[0097] Considering the continuity of power, the action space uses a continuous space.

[0098] a={a1,a2,...,a N} (11)

[0099] a i =μ i (o i )+N noise (12)

[0100] In the formula, a i represents the continuous actions of the ith agent; μ i () represents the current strategy of the i-th agent, which is used to adjust the transmission power; o i represents the current observation of the i-th agent; N noise Represents superimposed noise, which follows a normal distribution and is used to increase the randomness of action selection.

[0101] When each agent takes its own action, its state changes accordingly.

[0102] Reward function:

[0103]

[0104]

[0105] In the formula, represents the signal-to-interference-to-noise ratio at time slot t, P i t represents the received power of agent i at time slot t, represents the interference from other active nodes when nodes i and j communicate, σ 2 is Gaussian white noise.

[0106] r t It represents the sum rate of the system in the tth step, which is both the reward function of the agent and the common optimization goal of the agent. By dynamically adjusting the transmission power of the transmitter, the cumulative reward is increased to obtain a better strategy.

[0107] In the MADDPG algorithm, first, the central node and each relay node are defined as an agent; π={π1,π2,...,π N} to represent the N strategies that the agent can adopt, π i ={a i |o i} indicates that agent i is observing o i The action a obtained i , with x={o1,o2,...,o N} to represent the information of the entire environment; each interaction of the agent will be stored in the experience replay pool D, and data will be obtained in batches from it to update and learn the network parameters.

[0108] Each agent uses an independent Actor-Critic architecture, namely a policy-based Actor network and a value-based Critic network. The network structure is as follows: Figure 6 As shown; the Actor network collects state information from the environment and uses the strategy π iPerform the corresponding action a i , the Critic network uses the state-action value function Q to evaluate the action a selected by the Actor network i The degree of quality.

[0109] Q π (s,a)=E[r t +βQ π (s t+1 ,a t+1 )] (3)

[0110] Where β is the discount factor.

[0111] MADDPG algorithm uses a deterministic strategy μ i , implemented through a neural network, so that the Actor network updates parameters through this policy gradient:

[0112]

[0113] Where D is the experience replay pool, is the i-th evaluation function, θ i is the network parameter of the ith agent.

[0114] The Critic network optimizes the update parameters by minimizing the loss function of each agent:

[0115]

[0116]

[0117] In the formula, represents the action function of the target network, θ i ' is the parameter of the target network of the i-th agent, r i Represents the return reward of agent i.

[0118] The target actor network and the target critic network both use soft update to update network parameters, namely:

[0119]

[0120]

[0121] Where τ is the soft update parameter, which are and Parameters of the current Actor network and the target Actor network; and are the parameters of the current Critic network and the target Critic network.

[0122] like Figure 7 As shown, the specific steps of dynamic power adjustment based on the MADDPG algorithm are as follows:

[0123] S21, initialize the experience replay pool;

[0124] S22, initialize the random process N noise ;

[0125] S23, initialize the network parameters of Actor and Critic,

[0126] S24, setting the experience learning batch size, number of iterations and the step size of each iteration;

[0127] S25, in each iteration, initialize the state s of the current agent t , according to formula (12) select a t , the agent is based on the current state s t and action a t Perform state transition and reach the next state s t+1 , and calculate the reward r t .

[0128] S26, the current data {s t ,a t ,r t ,s t+1}Store it in the experience replay pool, and take out N groups of data from it to update the network parameters.

[0129] S27, updates the Actor network gradient and the Critic network's Loss function, thereby updating the Actor and Critic network parameters and

[0130] S28, soft update target Actor network parameters and target critic network parameters End this round of iteration;

[0131] S29, repeat S25 to S28 until convergence.

[0132] Before performing the dynamic adjustment scheme of the present invention, Figure 8 As shown in (a) of FIG. 1 , when the communication load of relay node A is heavy, the nodes not connected to the network still preferentially select node A to connect to the network, resulting in communication congestion at node A, while other relay nodes are relatively idle, thus causing network delay. After the dynamic adjustment scheme of the present invention is implemented, Figure 8As shown in (b), relay node A automatically reduces the broadcast transmission power in real time when the communication load is heavy. At the same time, the non-networked nodes dynamically adjust the receiving gain threshold, receive the control subframe broadcast from the central node, and select the central node for direct connection, thereby optimizing the network topology and improving the networking efficiency. As a result, the networked nodes with heavy active node loads dynamically adjust the broadcast transmission power, filter out the non-networked nodes, and reduce the possibility of receiving unnecessary network access requests. The non-networked nodes avoid the nodes with heavy loads in the next hop selection, balance the network load, and dynamically adjust the receiving gain threshold to optimize the network topology.

[0133] The method of the present invention may be implemented in hardware, firmware, or as software or computer code that may be stored in a recording medium (such as a CDROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code that is originally stored in a remote recording medium or a non-temporary machine-readable medium downloaded over a network and will be stored in a local recording medium, so that the method described herein may be stored in such software processing on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It is understood that a computer, processor, microprocessor controller, or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code, and when the software or computer code is accessed and executed by a computer, processor, or hardware, the method described herein is implemented. In addition, when a general-purpose computer accesses the code for implementing the method shown herein, the execution of the code converts the general-purpose computer into a dedicated computer for executing the method shown herein.

[0134] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A two-way dynamic adjustment method for the emergency communication sidelink networking listening phase, characterized in that: The following steps are involved: After the node not connected to the network is powered on, the receiving gain threshold of the receiving end is dynamically adjusted; While the receiving end dynamically adjusts the receiving gain threshold, the transmitting end's transmitting power is dynamically adjusted based on the MADDPG algorithm.

2. The method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase according to claim 1 is characterized in that: The steps of dynamically adjusting the receiving gain threshold of the receiving end include: S11, set an initial receiving gain threshold Thres init ; S12, after the node not connected to the network is powered on, the initial receiving gain threshold Threshold init Enter the initial listening process, set and start the listening timer, continue to receive control subframes until the timer overflows, and evaluate the existing local networking information after the end; S13, in the listening and waiting process, a relative listening and waiting time is introduced to dynamically reduce the receiving gain threshold.

3. The method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase according to claim 2 is characterized in that: When the relative listening waiting time is introduced to reduce the receiving gain threshold, a linear adjustment or an exponential adjustment method is adopted; The expression for linear adjustment of the receive gain threshold is: Tres listen =Three init +l·ΔT waiting Among them, Thres listen is the receiving gain threshold that changes dynamically during the listening waiting process, ΔT waiting is the relative listening waiting time, which is positively correlated with the number of cycles of the listening waiting period, and l is a negative adjustment coefficient; The expression for adjusting the receiving gain threshold in exponential form is: Among them, θ is the positive adjustment coefficient.

4. The method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase according to claim 2 is characterized in that: If during the dynamic adjustment process, the number of listening waiting cycles exceeds the limit N Timer_Limit Or the receiving gain threshold is adjusted below the lower limit Thres Lower_limit , the dynamic adjustment is terminated, and the nodes that have not successfully accessed the network will be used as the initial nodes to initiate networking.

5. The method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase according to claim 1 is characterized in that: When the transmitter dynamically adjusts the transmission power, in the MADDPG algorithm, first, the central node and each relay node are defined as an agent; use π = {π1, π2, ..., π N } to represent the N strategies that the agent can adopt, π i ={a i |o i } indicates that agent i is observing o i The action a obtained i , with x={o1,o2,...,o N } to represent the information of the entire environment; each interaction of the agent will be stored in the experience replay pool D, and data will be obtained in batches from it to update and learn the network parameters; Each of the above intelligent agents uses a policy-based Actor network and a value-based Critic network; the Actor network collects state information from the environment and uses the policy π i Perform the corresponding action a i , the Critic network uses the state-action value function Q to evaluate the action a selected by the Actor network i The degree of quality.

6. The method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase according to claim 5 is characterized in that: The steps of dynamically adjusting the transmit power of the transmitter based on the MADDPG algorithm include: S21, initialize the experience replay pool; S22, initialize the random process N noise ; S23, initialize the network parameters of Actor and Critic, S24, setting the experience learning batch size, number of iterations and the step size of each iteration; S25, in each iteration, initialize the state s of the current agent t , select a t , the agent is based on the current state s t and action a t Perform state transition and reach the next state s t+1 , and calculate the reward r t ; S26, the current data {s t ,a t ,r t ,s t+1 }Store it in the experience replay pool, and take out N groups of data from it to update the network parameters; S27, updates the Actor network gradient and the Critic network's Loss function, thereby updating the Actor and Critic network parameters and S28, soft update target Actor network parameters and target critic network parameters End this round of iteration; S29, repeat S25 to S28 until convergence.

7. A two-way dynamic adjustment system for the emergency communication sidelink networking listening phase, characterized in that: include: Receiver gain adjustment module: After the node not connected to the network is powered on, dynamically adjust the receiver gain threshold; And, the transmitting end power adjustment module: while the receiving end dynamically adjusts the receiving gain threshold, the transmitting end's transmitting power is dynamically adjusted based on the MADDPG algorithm.

8. A computer storage medium storing a readable program, characterized in that: When the program is running, it can execute the bidirectional dynamic adjustment method in the emergency communication sidelink networking listening phase described in any one of claims 1-6.

9. An electronic device, comprising: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the method for bidirectional dynamic adjustment of the emergency communication side link networking listening phase as described in any one of claims 1-6.

10. A computer program product, comprising computer instructions, wherein the computer instructions instruct a computing device to execute operations corresponding to the method for bidirectional dynamic adjustment of the emergency communication sidelink networking listening phase as described in any one of claims 1-6.

Images