Fire-fighting robot network interface dynamic management method based on dual-mode configuration

Abstract

The invention belongs to the technical field of fire-fighting robots, and particularly relates to a fire-fighting robot network interface dynamic management method based on dual-mode configuration, and the method comprises the steps: network state full-level monitoring: through a communication module hardware interface, a network protocol stack statistical function and video stream analysis; intelligent link switching decision making: adopting a strategy of combining threshold triggering type switching and load balancing type switching; switching execution and fault-tolerant processing: seamless link switching is realized through a pre-connection mechanism and data cache resending; dynamically scheduling communication resources: allocating bandwidth resources according to data service priorities; and network interface security guarantee: data encryption, role-based access control and security audit monitoring are adopted. According to the method, the optimal communication link can be automatically selected in a complex environment, communication failure caused by single network interruption is avoided, the communication reliability of the fire-fighting robot is remarkably improved, communication requirements of different rescue scenes can be met, and the method can be used according to priorities and real-time network states of different services.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of fire-fighting robots, and particularly relates to a dynamic management method for a network interface of a fire-fighting robot based on a dual-mode configuration. BACKGROUND

[0002] In modern fire rescue work, a fire-fighting robot as an important rescue equipment can enter a dangerous environment such as high temperature, poison and oxygen deficiency to perform tasks such as reconnaissance, fire extinguishing and rescue, and effectively protect the life safety of firefighters. However, the fire scene environment is complex and changeable, which often leads to unstable or even interrupted communication signals, affecting the normal work and remote control of the fire-fighting robot. At present, the fire-fighting robot usually adopts a single network communication mode, such as using only 5G network or only MESH self-organizing network. This is prone to communication interruption or quality decline in the case of signal shielding and interference. In addition, the existing management method for the network interface of the fire-fighting robot lacks real-time monitoring and intelligent switching capability of the network state, and cannot dynamically adjust the communication strategy according to the environmental changes, resulting in low communication reliability in complex environments. At the same time, with the continuous enhancement of the functions of the fire-fighting robot, the requirements for communication bandwidth, delay and reliability are also increasing. The traditional single network communication mode is difficult to meet the differentiated needs of various businesses such as high-definition video transmission and real-time control instruction transmission, limiting the application of the fire-fighting robot in complex rescue scenes. Therefore, there is an urgent need for a fire-fighting robot network interface management method that can adapt to complex environments, has high reliability and flexibility, to ensure that the fire-fighting robot can maintain stable communication connection under various adverse conditions, and improve rescue efficiency and success rate. SUMMARY

[0003] The purpose of the present application is to provide a dynamic management method for the network interface of a fire-fighting robot based on a dual-mode configuration, which can automatically select the optimal communication link in complex environments, avoid communication failure caused by single network interruption, significantly improve the communication reliability of the fire-fighting robot, and adapt to the communication needs of different rescue scenes. It can reasonably allocate network resources according to the priority of different businesses and real-time network state, improve bandwidth utilization, and ensure the communication quality of key businesses.

[0004] The technical solution adopted by the present application is as follows: The dynamic management method for the network interface of a fire-fighting robot based on a dual-mode configuration comprises the following steps: S1: Network state full-level monitoring: through the communication module hardware interface, network protocol stack statistical function and video stream analysis, the communication indicators of the physical layer, link layer and application layer are obtained respectively. S2: Intelligent link switching decision: adopt the strategy of combining threshold triggered switching and load balancing switching, determine whether to perform link switching and traffic distribution scheme based on monitored network indicators and fire scene environment parameters; S3: Switching execution and fault handling: realize seamless link switching through pre-connection mechanism and data cache retransmission, and cope with switching failure and system failure through automatic rollback, multi-level alarm and dual system hot backup; S4: Dynamic scheduling of communication resources: allocate bandwidth resources according to data service priority, and dynamically adjust the transmission power of the communication module according to the physical layer signal strength; S5: Network interface security guarantee: adopt data encryption, role-based access control and security audit monitoring to ensure the security and integrity of data transmission between the fire-fighting robot and the command center.

[0005] The physical layer indicators in S1 include signal strength, signal-to-noise ratio and bit error rate; The signal strength is obtained by 5G module AT instruction receiving level, and its calculation formula is:

[0006] Among them, the is the signal strength, the 113 represents the reference value corresponding to the theoretical minimum receivable signal strength threshold of a specific communication module, and the is the receiving level; The signal-to-noise ratio is obtained by measuring signal power and noise power, and its calculation formula is:

[0007] Among them, the is the signal-to-noise ratio, the is the signal power, and the is the noise power; The bit error rate is obtained by comparing the CRC check codes of the sent data and the received data through the cyclic redundancy check algorithm; The link layer indicators include end-to-end delay, packet loss rate and throughput, and the calculation formula of the end-to-end delay is:

[0008] Among them, the is the end-to-end delay, the is the data packet receiving timestamp, and the is the data packet sending timestamp; The calculation formula of the packet loss rate is:

[0009] Among them, the is the packet loss rate, the is the number of sent data packets, the is the number of received data packets; The calculation of the throughput is:

[0010] wherein the is the packet loss rate, the is the unit time of data transmission, the In the total amount of data actually transmitted between the robot and the command center within the corresponding time interval; The application layer indicators include the video frame rate and the control instruction response time; The calculation formula of the frame rate is:

[0011] wherein the is the video frame rate, the is the time interval between adjacent two frames of video; The calculation formula of the control instruction response time is:

[0012] wherein the is the control instruction response time, the is the response receiving time, the is the instruction sending time.

[0013] The network state full-level monitoring in the S1 adopts a combination of periodic polling and event triggering, the periodic polling is realized through the timer interrupt mechanism, and the event triggering is triggered through monitoring the abnormal situation that the signal strength changes by more than 10 dB within 100 ms, and after triggering, a comprehensive network state detection is performed once.

[0014] The threshold triggered switching in the S2 includes basic threshold setting and dynamic threshold adjustment based on reinforcement learning; The basic threshold setting is to automatically switch to the backup link when the main link delay is greater than 50 ms or the packet loss rate is greater than 5%; The reinforcement learning adopts the Q-learning algorithm, takes the network state and the fire scene environmental factors as the state input, takes the switching decision as the action, takes the number of communication interruptions and the data transmission success rate as the reward function, dynamically adjusts the switching threshold through the Q value update formula, and the formula is:

[0015] wherein the is the learning rate, is a current reward, is a discount factor, is a current state, is a current action.

[0016] The load balancing switching in S2 is realized by combining a biogeography optimization algorithm with an IPMP protocol, the biogeography optimization algorithm regards network links as habitats and data flow as biological populations, and assigns flow through a fitness function; The calculation formula of the fitness function is:

[0017] Among them, the , and are weight coefficients.

[0018] The biogeography optimization algorithm introduces a dynamic weight adjustment mechanism, when the network is congested, the delay and packet loss rate weights are increased, and the bandwidth utilization rate weight is reduced, and when the network is idle, the bandwidth utilization rate weight is increased.

[0019] The seamless switching technology in S3 includes a pre-connection mechanism and data cache retransmission; The pre-connection mechanism makes the 5G and MESH modules simultaneously maintain a registered state through the double-link registration function of the fire-fighting robot communication module, and the switching time is less than 50ms; the data cache retransmission opens a 20-minute key data cache area in the robot memory, stores unsent data according to data priority, and retransmits in turn after switching is completed.

[0020] The fault tolerance processing in S3 includes automatic rollback and dual-system hot backup; The automatic rollback rolls back to the main link and triggers multi-level alarms when the standby link switching fails, the dual-system hot backup sets two independent operating systems in the robot controller, the standby system monitors the state of the main system in real time, and when the main system fails, the standby system takes control and calls the BIM data of the TEE security zone backup.

[0021] The technical effects obtained by the application are: The application can automatically select the optimal communication link in a complex environment through a dual-mode configuration and an intelligent switching strategy, avoid communication failure caused by single network interruption, significantly improve the communication reliability of the fire-fighting robot, support various network combination configurations, adapt to the communication needs of different rescue scenes, such as using optical fiber and wireless combination in high-rise building fire, using cellular network and satellite communication combination in mountain fire, and through dynamic resource scheduling, can reasonably allocate network resources according to the priority of different services and real-time network state, improve the bandwidth utilization, and ensure the communication quality of key services. BRIEF DESCRIPTION OF DRAWINGS

[0022] Fig. 1 is the whole dynamic management method flowchart of the application; Fig. 2 is the network state full-level monitoring sub-flowchart of the application. DETAILED DESCRIPTION

[0023] In order to make the purpose and advantages of the application more clear and obvious, the application will be specifically described below in combination with examples. It should be understood that the following text is only used to describe one or several specific embodiments of the application, and does not strictly limit the specific protection scope requested by the application.

[0024] As shown in Figs. 1-2 , the network interface dynamic management method of the fire-fighting robot based on dual-mode configuration includes the following steps: S1: Network state full-level monitoring: through the communication module hardware interface, network protocol stack statistical function and video stream analysis, the communication indicators of the physical layer, link layer and application layer are obtained respectively; The physical layer indicators include signal strength (RSSI), signal-to-noise ratio (SNR) and bit error rate (BER); The signal strength is obtained by 5G module AT instruction receiving level, and its calculation formula is:

[0025] Among them, is the signal strength, -113 represents the reference value corresponding to the theoretical minimum receivable signal strength threshold of a specific communication module, is the receiving level; The signal-to-noise ratio is obtained by measuring the signal power and noise power, and its calculation formula is:

[0026] Among them, is the signal-to-noise ratio, is the signal power, is the noise power; The bit error rate can be calculated by the bit error detection algorithm built in the communication module, such as the cyclic redundancy check (CRC) algorithm, by comparing the CRC check codes of the sent data and the received data. The link layer indicators include the end-to-end delay, packet loss rate and throughput obtained by using the statistical functions in the network protocol stack. For the 5G network, the end-to-end delay can be calculated by recording the sending and receiving timestamps of the data packets at the sending end and the receiving end, and the calculation formula of the end-to-end delay is:

[0027] wherein, is the end-to-end delay, is the data packet receiving timestamp, is the data packet sending timestamp; The calculation formula of the packet loss rate is:

[0028] wherein, is the packet loss rate, is the number of sent data packets, is the number of received data packets; The calculation formula of the throughput is:

[0029] wherein, is the packet loss rate, is the unit time of data transmission, is the total amount of data actually transmitted between the robot and the command center in the corresponding time interval. The application layer indicators include the video frame rate and the control instruction response time. The calculation formula of the frame rate is:

[0030] wherein, is the video frame rate, is the time interval between two adjacent video frames; The control instruction response time can be calculated by recording the sending time when the control instruction is sent and the receiving time when the response is received, and the difference between the two is the response time. The calculation formula of the control instruction response time is:

[0031] wherein, is the control instruction response time, is the response receiving time, is the instruction sending time; ​The network state full-level monitoring adopts a combination of periodic polling and event triggering. The periodic polling is achieved through a timer interrupt mechanism and adopts a combination of periodic polling and event triggering. The periodic polling is set to occur every 100 ms, and is achieved through a timer interrupt mechanism. The event triggering is achieved by monitoring signal mutations and other abnormal conditions. For example, when the signal strength (RSSI) changes by more than a certain threshold (such as 10 dB) within a short period of time, a comprehensive network state detection is triggered immediately. This approach can ensure real-time performance while effectively reducing energy consumption. S2: Intelligent link switching decision: a combination of threshold-triggered switching and load-balanced switching strategy, based on the monitored network indicators and fire scene environmental parameters, to determine whether to perform link switching and traffic distribution scheme; The threshold-triggered switching includes basic threshold setting and dynamic threshold adjustment based on reinforcement learning. The basic threshold setting is to automatically switch to the backup link when the main link delay is greater than 50 ms or the packet loss rate is greater than 5%, and then determine the basic requirements for real-time performance and reliability based on the fire scene, for example, in a multi-story building fire rescue, when the 5G main link delay exceeds 50 ms, it indicates that the network transmission is congested or interfered, at this time, the MESH backup link should be switched to in time to ensure that the control command can be transmitted to the fire robot in time; The reinforcement learning adopts the Q-learning algorithm, taking the network state and fire scene environmental factors (such as temperature, smoke concentration, etc.) as state input, the switching decision as action, the number of communication interruptions and the data transmission success rate as reward function, and through continuous training in actual scenes, the algorithm can dynamically adjust the threshold value according to different environmental conditions, and through the Q value update formula, the switching threshold value is dynamically adjusted, and the formula is:

[0032] wherein, is the learning rate, is the current reward, is the discount factor, is the current state, is the current action, and through continuous iteration of the Q value, the algorithm can learn the optimal switching threshold value in different environments.

[0033] The load balancing switching is realized by combining the biogeography-based optimization algorithm with the IPMP protocol. The biogeography-based optimization algorithm regards network links as habitats and data traffic as biological populations. The fitness function (such as bandwidth utilization, delay, etc.) of each link is calculated to determine the allocation of traffic. For example, video streams are allocated to 5G links because 5G links have high bandwidth characteristics and can meet the large data transmission requirements of video streams; control instructions are allocated to MESH links because MESH links have good stability in complex environments and can ensure reliable transmission of control instructions. The calculation formula of the fitness function is:

[0034] wherein, , and are weight coefficients, which are adjusted according to the requirements of different data types. Through continuous iteration and optimization, the traffic allocation reaches an optimal state, maximizing bandwidth utilization. The biogeography-based optimization algorithm introduces a dynamic weight adjustment mechanism. When the network is congested, the weights of delay and packet loss rate are increased, and the weight of bandwidth utilization is reduced. When the network is idle, the weight of bandwidth utilization is increased. The IPMP protocol monitors the state of the network interface. When a certain interface fails, it automatically switches the traffic to other normal interfaces. In the fire-fighting robot network, by configuring the IPMP protocol, parallel transmission of multiple network interfaces such as 5G and MESH is realized. For example, when the 5G link is interrupted, the IPMP protocol can automatically switch part of the data traffic to the MESH link, ensuring the continuity of data transmission. At the same time, the IPMP protocol can also transmit data through redundant paths to improve the reliability of data transmission. The biogeography-based optimization algorithm simulates the distribution and migration process of biological populations in different habitats. In the BBO algorithm, each habitat represents a potential solution, and the habitat suitability index (HSI) represents the quality of the solution. The algorithm gradually optimizes the solution by simulating the migration and variation process of species between habitats. Different network links (such as 5G, MESH) are regarded as different habitats, and data traffic is regarded as biological populations. The fitness function of each link, such as bandwidth utilization, delay, and packet loss rate, is calculated to determine the allocation scheme of traffic. For example, video streams are allocated to 5G links because 5G links have high bandwidth characteristics and can meet the large data transmission requirements of video streams; control instructions are allocated to MESH links because MESH links have good stability in complex environments and can ensure reliable transmission of control instructions. To adapt to the characteristics of the fire-fighting robot network, the standard BBO algorithm is improved by introducing a dynamic weight adjustment mechanism. The weights of the parameters in the fitness function are adjusted in real time according to the changes in the network state. When the network is congested, the weights of delay and packet loss rate are increased, and the weight of bandwidth utilization is reduced to prioritize low-delay transmission. When the network is idle, the weight of bandwidth utilization is increased to improve the utilization efficiency of network resources. S3: Switching execution and fault tolerance processing: When the main link is normal, the connection of the standby link is established in advance. For example, for 5G and MESH networks, a dual-link registration function is set in the communication module of the fire-fighting robot to keep the 5G and MESH modules registered at the same time. When switching is needed, the established connection can be directly used, and the switching time is less than 50ms. This mechanism can effectively reduce the switching delay and ensure the continuity of communication. Seamless link switching is achieved through pre-connection mechanism and data cache retransmission. At the same time, automatic rollback, multi-level alarm and dual-system hot backup are used to deal with switching failure and system failure. Seamless switching technology includes pre-connection mechanism and data cache retransmission. The pre-connection mechanism uses the dual-link registration function of the fire-fighting robot communication module to keep the 5G and MESH modules registered at the same time, with a switching time of less than 50ms. A 20-minute key data cache area is set up. During network switching, the unsent data is temporarily stored in the cache area. After switching is completed, the data in the cache is retransmitted according to the priority of the data to ensure the continuity of the command. For high-priority data such as control commands, retransmission is given priority to ensure the normal operation of the fire-fighting robot.

[0035] Fault tolerance processing includes automatic rollback and dual-system hot backup. Automatic rollback is used when the standby link fails to switch. After detecting the failure signal, it returns to the main link and triggers multi-level alarm. Dual-system hot backup sets up two independent operating systems in the robot controller. The standby system monitors the status of the main system in real time. When the main system fails, the standby system takes control and calls the BIM data backed up in the TEE security area. If the standby link fails to switch, the hardware or software monitoring mechanism detects the switching failure signal and automatically returns to the main link, triggering multi-level alarm. For example, the buzzer on the fire-fighting robot emits an alarm sound, and the LED light is turned on for local prompt. Then the wireless communication module sends a notification to the remote command center to inform the switching failure, so that timely measures can be taken. Adopt dual-system hot backup technology, such as setting two independent operating systems in the controller of the fire robot, one as the main system, the other as the standby system to monitor the main system in real time, when the main system fails, the standby system can take control immediately, ensure the core control logic uninterrupted, the BIM data stored in the TEE security area as the key data backup, in the system switching process, ensure the integrity and security of data; S4: Dynamic scheduling of communication resources: allocate bandwidth resources according to data service priority, and dynamically adjust the transmission power of the communication module according to the physical layer signal strength; The priority of control instructions such as fire monitor steering, water spray control, etc. is set to the highest, and the bandwidth reservation is not less than 30%, because the real-time and reliability of control instructions are directly related to the effect of fire rescue action, for example, in the fire scene, the accurate steering of fire monitor and the control of water spray are crucial to fire extinguishing work, and these instructions must be ensured to be transmitted to the fire robot in time and accurately; And the priority of video stream is second, using H.265 encoding compression technology to save bandwidth, H.265 encoding can save about 50% of bandwidth compared with H.264 encoding under the same picture quality; For example, in the on-site video transmitted by the fire robot, the video data is compressed by H.265 encoding, which reduces the occupation of network bandwidth under the premise of meeting the rescue demand of video quality, so that more bandwidth can be used to transmit critical data such as control instructions; Voice communication priority: based on DMR digital trunking protocol to realize low delay transmission of voice communication, less than 100ms, DMR protocol uses time division multiple access (TDMA) technology, which allocates each time slot to different users to realize efficient voice communication; And in fire rescue, the voice communication between rescue personnel needs to be timely and clear, which can be ensured by DMR protocol to realize low delay and high reliability of voice communication; And can dynamically adjust the transmission power according to the signal strength, for example, for 5G module, when the signal strength is-70dBm, the transmission power is set to 23dBm, when the signal strength decreases to-90dBm, the transmission power is increased to 30dBm, in this way, the battery endurance of the fire robot is prolonged under the premise of ensuring the communication quality, the adjustment of transmission power can be realized through the power control interface of the communication module, and the specific adjustment strategy can be optimized according to the characteristics of different communication modules; The fire-fighting robot is locally deployed with an edge computing module, such as an RK3576 controller, which has strong computing power and can process sensor data in real time, for example, fire source positioning data collected by a dual-spectrum thermal imaging camera. The edge computing module can analyze the data in real time, quickly determine the location of the fire source, and reduce the pressure of data transmission to the cloud. At the same time, the edge computing module can also preprocess other sensor data (such as temperature, smoke concentration, etc.) and extract key features to support subsequent intelligent decision-making. And use LSTM model to predict network state: Long Short-Term Memory (LSTM) model can effectively process time series data and has good application prospects in fire-fighting robot network state prediction. The historical data of network state such as delay, packet loss rate, signal strength are used as input, and the time series features in the data are learned through LSTM model to predict the communication interruption risk 5 seconds in advance and trigger pre-switching. The core structure of LSTM model includes input gate, forget gate, output gate and memory cell. At each time step , the input gate decides which information in the input data will be written into the memory cell , the forget gate decides which information in the memory cell will be retained, and the output gate decides which information in the memory cell will be output. The calculation formula is as follows:

[0036] Where, is the Sigmoid function, is the element-wise multiplication, is the weight matrix, is the bias vector. By training the LSTM model, it can accurately predict the network state and take switching measures in advance to avoid communication interruption. YOLOv7 algorithm is used to analyze the video stream collected by the fire-fighting robot in real time to identify smoke concentration and obstacles. YOLOv7 algorithm has high accuracy and real-time performance in target detection tasks. By analyzing the images in the video stream, the communication strategy can be dynamically adjusted. When detecting a smoky environment, MESH network is preferred because MESH network has better signal transmission stability than other networks in smoky environment. At the same time, according to the identification result of the obstacles, the moving path of the robot can be adjusted to ensure the safe passage of the robot in complex environment. S5: Network interface security assurance: data encryption, role-based access control and security audit monitoring are adopted to ensure the security and integrity of data transmission between the fire-fighting robot and the command center.

[0037] The secure transmission layer protocol (TLS) and data encryption technology are adopted to ensure the security and integrity of data in the transmission process, the control instructions and video data are encrypted using the AES-256 algorithm to prevent data from being stolen or tampered with, and at the same time, digital certificates are used for two-way authentication to ensure the authenticity of the identity of the communication parties; A role-based access control (RBAC) system is established to limit access and operation of the fire-fighting robot according to the roles and permissions of different users, the senior commanders of the command center have the highest permission and can control all functions of the robot, the on-site rescue personnel have partial permission and can only control the basic functions of the robot, and the ordinary users can only view the video and data transmitted by the robot; Security audit and monitoring of the network interface of the fire-fighting robot are realized, all access and operation behaviors are recorded, security events are discovered and handled in time, network traffic is monitored in real time through the intrusion detection system (IDS) to identify abnormal behaviors and potential attacks, and all operations and events are recorded through the security log system to provide a basis for security audit.

[0038] The following effects are further realized: Through the dual-mode configuration and intelligent switching strategy, the optimal communication link can be automatically selected in complex environments, communication failure caused by single network interruption is avoided, the communication reliability of the fire-fighting robot is significantly improved, and multiple network combination configurations are supported, which can adapt to the communication needs of different rescue scenes, such as using optical fiber and wireless combination in high-rise building fire, using cellular network and satellite communication combination in mountain fire, and through dynamic resource scheduling, network resources can be reasonably allocated according to the priority of different services and real-time network state, bandwidth utilization is improved, communication quality of key services is ensured, and then transmission power is dynamically adjusted according to signal strength, battery endurance time of the fire-fighting robot is prolonged while communication quality is ensured, work efficiency is improved, multiple services such as high-definition video transmission, control instruction transmission and voice communication can be supported at the same time, and the diversified needs of modern fire rescue are met.

[0039] The above only describes the preferred embodiments of the present application, it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, these improvements and refinements should also be considered as the protection scope of the present application. The structures, devices and operation methods not specifically described and explained in the present application, such as without special description and limitation, are implemented according to the conventional means in the art.

Claims

1. A dynamic management method for the network interface of a firefighting robot based on dual-mode configuration, characterized in that, Includes the following steps: S1: Full-level network status monitoring: Through the communication module hardware interface, network protocol stack statistics function and video stream parsing, communication indicators of the physical layer, link layer and application layer are obtained respectively; S2: Intelligent Link Switching Decision: Adopting a strategy that combines threshold-triggered switching and load balancing switching, based on monitored network indicators and fire scene environmental parameters, it determines whether to execute link switching and traffic allocation schemes. S3: Switching Execution and Fault Tolerance: Seamless link switching is achieved through pre-connection mechanism and data caching and resending, while automatic rollback, multi-level alarms and dual-system hot backup are used to deal with switching failures and system failures; S4: Dynamic scheduling of communication resources: allocate bandwidth resources according to data service priority, and dynamically adjust the transmission power of communication modules according to the physical layer signal strength; S5: Network Interface Security: Employs data encryption, role-based access control, and security audit monitoring to ensure the security and integrity of data transmission between the firefighting robot and the command center.

2. The dynamic management method for the network interface of a fire-fighting robot based on dual-mode configuration according to claim 1, characterized in that: The physical layer metrics in S1 include signal strength, signal-to-noise ratio, and bit error rate; The signal strength is obtained by using AT commands from the 5G module to obtain the received level, and its calculation formula is as follows: RSSI = -113 + RxLev Wherein, RSSI is the signal strength, -113 represents the reference value corresponding to the theoretical minimum receivable signal strength threshold of a specific communication module, and RxLev is the received level; The signal-to-noise ratio is obtained by measuring the signal power and noise power, and its calculation formula is: Wherein, SNR is the signal-to-noise ratio, Psignal is the signal power, and Pnoise is the noise power; The bit error rate is calculated by comparing the CRC checksums of the transmitted and received data using a cyclic redundancy check algorithm. The link layer metrics include end-to-end latency, packet loss rate, and throughput. The formula for calculating end-to-end latency is: Delay=T receive -T send Wherein, Delay is the end-to-end delay, and T receive The T is the timestamp for receiving data packets. send For the timestamp of the data packet transmission; The formula for calculating the packet loss rate is: Wherein, the Packet Loss Rate is the packet loss rate, and N send N represents the number of data packets to be sent. receive The number of data packets received; The formula for calculating the throughput is: Wherein, Packet Loss Rate is the packet loss rate, Time is the unit time for data transmission, and DataVolume is the total amount of data actually transmitted between the robot and the command center within the time interval corresponding to Time; The application layer metrics include video frame rate and control command response time; The formula for calculating the frame rate is: Wherein, Frame Rate is the video frame rate, and T frame The time interval between two adjacent video frames; The formula for calculating the control command response time is: Response Time=T response -T command Wherein, Response Time is the control command response time, and T response In response to the reception time, the T command This refers to the time the instruction was sent.

3. The dynamic management method for the network interface of a fire-fighting robot based on dual-mode configuration according to claim 2, characterized in that: The network status monitoring at all levels in S1 adopts a combination of periodic polling and event triggering. The periodic polling is implemented through a timer interrupt mechanism, and the event triggering is triggered by an abnormal situation where the signal strength changes by more than 10dB within 100ms. After triggering, a comprehensive network status detection is performed.

4. The dynamic management method for the network interface of a fire-fighting robot based on dual-mode configuration according to claim 3, characterized in that: The threshold-triggered switching in S2 includes basic threshold setting and dynamic threshold adjustment based on reinforcement learning; The basic threshold is set to automatically switch to the backup link when the main link latency is greater than 50ms or the packet loss rate is greater than 5%. The reinforcement learning employs the Q-learning algorithm, taking network state and fire scene environmental factors as state inputs, switching decisions as actions, and communication interruption counts and data transmission success rates as reward functions. The switching threshold is dynamically adjusted through a Q-value update formula, which is as follows: Q(s t ,a t )=Q(s t ,a t )+α[r t +γ·max a Q(s t+1 ,a)-Q(s t ,a t )] Where 'a' is the learning rate, and 'r' is the learning rate. t s is the current reward, γ is the discount factor, and s t As the current state, a t This is the current action.

5. The dynamic management method for the network interface of a fire-fighting robot based on dual-mode configuration according to claim 4, characterized in that: The load balancing switching in S2 is achieved by combining a biogeographical optimization algorithm with the IPMP protocol. The biogeographical optimization algorithm treats network links as habitats and data traffic as biological populations, and allocates traffic through a fitness function. The fitness function is calculated as follows: H = ω1 × bandwidth utilization + ω2 × latency + ω3 × packet loss rate Wherein, ω1, ω2 and ω3 are weighting coefficients.

6. The dynamic management method for the network interface of a fire-fighting robot based on dual-mode configuration according to claim 5, characterized in that: The biogeographic optimization algorithm introduces a dynamic weight adjustment mechanism. When the network is congested, the weights of latency and packet loss rate are increased, while the weight of bandwidth utilization is decreased. When the network is idle, the weight of bandwidth utilization is increased.

7. The dynamic management method for the network interface of a fire-fighting robot based on dual-mode configuration according to claim 6, characterized in that: The seamless handover technology described in S3 includes a pre-connection mechanism and data caching and resending. The pre-connection mechanism uses the dual-link registration function of the fire robot's communication module to keep the 5G and MESH modules registered simultaneously, with a switching time of less than 50ms. The data cache retransmission opens a 20-minute critical data cache area in the robot's memory, stores unsent data according to data priority, and retransmits it sequentially after the switch is completed.

8. The dynamic management method for the network interface of a fire-fighting robot based on dual-mode configuration according to claim 7, characterized in that: The fault-tolerant processing described in S3 includes automatic rollback and dual-system hot backup; The automatic rollback, upon detecting a failure signal when the backup link fails to switch, rolls back to the main link and triggers multi-level alarms. The dual-system hot backup sets up two independent operating systems in the robot controller. The backup system monitors the status of the main system in real time. When the main system fails, the backup system takes over control and calls the BIM data backed up in the TEE security zone.

Images