A method for seamless signal switching across operators based on AI multi-network aggregation cards.

Abstract

This invention discloses a method for seamless cross-carrier signal handover based on an AI multi-network aggregation card, belonging to the field of robot communication technology. The method includes: acquiring the user's service range and robot information; configuring a cloud AI model in the cloud based on the robot information and service range; uploading the robot's task information to the cloud for analysis to obtain a cloud information map of the robot; storing the cloud information map in the robot's database; configuring a local AI model, a multi-network aggregation card, and a data acquisition module on the robot; performing signal handover preparation processing based on the cloud information map in the robot's database; acquiring line signal data in real time through the data acquisition module and generating a local information map based on the line signal data; determining the local signal priority sequence of each local equivalent area based on the local AI model and the local information map; performing signal handover preparation processing based on the local information map; and performing the corresponding signal handover when carrier signal handover is required.

Description

Technical Field

[0001] This invention belongs to the field of robot communication technology, specifically a method for seamless cross-carrier signal switching based on an AI multi-network aggregation card. Background Technology

[0002] With the rapid development of robotics technology, its applications have widely penetrated key areas such as industrial inspection, emergency rescue, logistics and transportation, and medical assistance, placing stringent demands on the stability, reliability, and real-time performance of communications. However, existing single-operator networks have revealed many insurmountable defects when supporting robot operations, becoming a core bottleneck restricting the intelligent development of robots.

[0003] Traditional single-operator networks rely on their own base stations for signal transmission, and their coverage is constrained by factors such as geographical environment, base station density, and frequency band characteristics. In complex scenarios, such as densely built-up urban canyons, underground pipe networks, remote mountainous areas, or disaster sites, single-operator base station deployments often suffer from blind spots or severe signal attenuation; even due to exclusivity agreements, only one operator may be allowed to enter the same area. Furthermore, single-operator networks lack flexibility in responding to dynamic environmental changes. During robot operations, real-time changes in their position, posture, and movement trajectory cause rapid fluctuations in signal strength. In existing technologies, some robots attempt to improve reliability by adding redundant communication modules (such as dual SIM dual standby), but such solutions are still limited by the coverage range of a single operator.

[0004] In order to solve the above problems, this invention provides a method for seamless signal switching across operators based on an AI multi-network aggregation card. Summary of the Invention

[0005] To address the problems of the above solutions, this invention provides a method for seamless signal switching across operators based on an AI multi-network aggregation card.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A method for seamless cross-carrier signal switching based on an AI-powered multi-network aggregation card, the method including:

[0008] Step 1: Obtain the user's service scope and robot information, and configure the corresponding cloud AI model in the cloud based on the robot information and service scope;

[0009] Furthermore, the cloud AI model can be one or more. When there are multiple cloud AI models, each cloud AI model is labeled with a corresponding applicable robot and task.

[0010] Step 2: Acquire task information of each robot at the user's location in real time, upload the task information to the cloud for analysis, and obtain the cloud information map of the robot. The cloud information map includes the signal priority sequence of each information-equivalent area on the robot's task route; store the cloud information map in the robot's database.

[0011] Furthermore, the task information is uploaded to the cloud for analysis, including:

[0012] The received task information is identified in the cloud to obtain the corresponding robot information and task information. The robot's task route is identified based on the task information, and a route map is generated based on the task route.

[0013] The system collects cloud signal data from various locations on the route map in real time and marks the cloud signal data accordingly on the route map. Based on robot information and task information, the system calls the corresponding cloud AI model and analyzes the cloud signal data from various locations to obtain the priority of each operator's information number in the location area. Based on the priority of each operator's information number, the system generates a signal priority sequence for the location area.

[0014] Merge adjacent locations with the same signal priority sequence to obtain the corresponding merged region, and mark the merged region as an information equivalent region; mark the signal equivalent region and the signal priority sequence accordingly on the line diagram; mark the current line image as a cloud information map.

[0015] Furthermore, the cloud information map is dynamically updated based on the real-time collected cloud signal data, and the updated cloud information map is synchronously updated to the robot's database.

[0016] Furthermore, the completed route of the robot is acquired in real time, and the completed route is marked accordingly in the route map, without collecting cloud signal data corresponding to the completed route in the route map.

[0017] Furthermore, a cloud-based information map is generated for the first time before the robot starts the task, and the generation of the cloud-based information map stops when it is determined that the robot has completed the task.

[0018] Step 3: Configure the corresponding local AI model, multi-network aggregation card and acquisition module on the robot. The acquisition module is used to collect the corresponding line signal data in real time; and perform signal switching preparation processing based on the cloud information map in the robot database.

[0019] Furthermore, signal handover preparation is carried out based on the cloud-based information map, including:

[0020] Step SA1: Mark the current position of the robot in real time in the cloud information map, identify the information equivalent area that the robot has in the remaining task line; obtain the robot's status information, and determine the switching preparation analysis area in real time based on the status information;

[0021] Step SA2: Identify the signal priority sequence and cloud signal data corresponding to each switching preparation analysis area; obtain the operator signal currently connected to the robot;

[0022] When the operator's signal ranks first in the signal priority sequence of the corresponding handover preparation analysis area, no signal handover preparation is required.

[0023] When the operator signal is not ranked first in the signal priority sequence of the corresponding handover preparation analysis area, the application signal data of the operator signal in each handover preparation analysis area is obtained; the application signal data is evaluated to see if it meets the preset robot task requirements.

[0024] When the applied signal data meets the robot's task requirements, no signal switching preparation is needed during the evaluation.

[0025] When the application signal data does not meet the robot's task requirements, an assessment is made to prepare for signal switching, and the operator's signal ranked first in the corresponding signal priority sequence is used as the backup switching signal.

[0026] Step SA3: Mark the handover preparation analysis area that needs to be prepared for signal handover in the cloud information map, and process it according to the preset signal handover preparation method.

[0027] Furthermore, assess whether the applied signal data meets the preset robot task requirements, including:

[0028] Establish a task calibration model, the expression of which is:

[0029] ;

[0030] In the formula: (s, KB) is the input data, s is the application signal data, and KB is the robot task requirement; s→KB indicates that the application signal data meets the robot task requirement; the output data is the task calibration value WR(s, KB), and the task calibration value is 1 or 0.

[0031] The relevant application signal data and robot task requirements are integrated into the input data and input into the task calibration model for analysis to obtain the corresponding task calibration values.

[0032] When the task calibration value is 1, the evaluation application signal data meets the robot task requirements;

[0033] When the task calibration value is 0, the evaluation application signal data does not meet the robot task requirements.

[0034] Step 4: Collect line signal data in real time through the acquisition module, supplement the line signal data in the preset route map, and generate a local information map. The local information map includes local equivalent areas and the line signal data corresponding to the local equivalent areas. Analyze the line signal data in the local information map through a local AI model to obtain the local signal priority sequence of each local equivalent area. Supplement the local signal priority sequence to the local information map.

[0035] Step 5: Prepare for signal handover based on the local information map. When it is necessary to switch carrier signals, perform the corresponding signal handover.

[0036] Furthermore, when the result of signal handover preparation processing based on the local information map differs from the result of signal handover preparation processing based on the cloud information map, the result of signal handover preparation processing based on the local information map shall be used as the standard.

[0037] Furthermore, when the signal switching analysis at the robot is abnormal, the cloud signal data of each signal equivalent area in the cloud information map is identified, and the cloud signal data is analyzed by the local AI model to obtain a new signal priority sequence; signal switching preparation is carried out according to the adjusted cloud information map.

[0038] When the local AI model cannot meet the analysis requirements, the signal switching preparation process based on the cloud information graph in step three is carried out.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] By utilizing cloud-based information maps to generate information about the robot, the robot can learn about the operator information in various areas along its route in advance, facilitating early handover preparation and improving signal handover efficiency. Simultaneously, the cloud-based information map enables the robot to continue signal handover analysis and processing during the task even when signal handover analysis anomalies occur, addressing corresponding emergencies and providing time for subsequent emergency response. This invention's handover preparation process avoids relying solely on AI assessment for frequent signal handovers. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0043] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0044] like Figure 1 As shown, a method for seamless cross-carrier signal switching based on an AI multi-network aggregation card is described, including:

[0045] Step 1: Determine the service area of ​​the user's robots based on historical user data. The service area covers all possible task route areas of the user's robots; alternatively, the service area can be determined based on the user's business needs. Obtain information about the user's robots, such as model, type, and purpose. Configure the corresponding cloud AI model based on the robot information and service area. The cloud AI model is used to analyze the input data at each location to determine the priority of each operator for a specific robot and task at that location. The input data includes signal quality parameters of each operator, robot status data, task requirements, environmental characteristic data, operator network status, and other relevant data.

[0046] In one embodiment, the cloud AI model is built based on existing AI technologies, and can be selected from reinforcement learning (RL), graph neural network (GNN), multimodal fusion model (combining CNN (processing signal time series data), Transformer (processing task text description) and LSTM (processing robot motion trajectory) to achieve cross-modal feature fusion), specifically using existing AI technologies to build the cloud AI model.

[0047] In one embodiment, a unified cloud AI model can be established for analysis. When the unified cloud AI model does not meet the requirements for analysis of robots or tasks, a targeted cloud AI model suitable for it can be established, and the corresponding cloud AI model can be called for analysis according to the actual situation.

[0048] Step 2: Obtain task information for each robot at the user's location in real time, upload the task information to the cloud for analysis, and obtain the cloud information map of the corresponding robot. The cloud information map includes the operator signal priority of each area along the robot's task route; store the cloud information map in the robot's database.

[0049] In one embodiment, uploading task information to the cloud for analysis includes:

[0050] The system identifies received task information in the cloud to obtain corresponding robot and task information. Based on the task information, it identifies the robot's task route and generates a route map. For example, it can use the route as the center and expand outwards to a certain distance as the route area, generating a route map based on the route area. Alternatively, it can generate a route map in other ways, meaning the route map is not a single line. It also pre-connects to the information channels needed for cloud AI model analysis, such as operator signal quality parameters, environmental characteristic data, and operator network status. Because the cloud information map is used to provide preliminary materials for robot signal switching, it is difficult to analyze data collected on-site by the robot. Therefore, data is generally collected through other information channels. Current collection technologies can collect relevant data, but robot status and task conditions can be uploaded by the robot. If uploading is not possible, the most recent data is analyzed or extrapolated, and the data collected from various channels is integrated and marked as cloud signal data.

[0051] The system collects real-time cloud signal data from various locations on the route map. Locations with identical or identical cloud signal data are grouped together to reduce the amount of data to be analyzed. Clustering algorithms can also be used to create different location areas based on cloud signal data. The cloud signal data is then marked on the route map. Based on robot and task information, the system calls the appropriate cloud AI model to analyze the cloud signal data in each location area, obtaining the priority of each operator's information number within that area. A signal priority sequence is then generated for that location area based on the priority of each operator's information. Location areas with identical and adjacent signal priority sequences are merged to obtain merged areas, which are then marked as information-equivalent areas. Finally, the signal-equivalent areas and signal priority sequences are marked on the route map accordingly.

[0052] Mark the current line image as a cloud infographic.

[0053] Subsequently, the cloud information map is dynamically updated based on changes in cloud signal data, and the updated cloud information map is synchronously updated to the robot's database.

[0054] In one embodiment, collecting and analyzing data for areas where the robot has already completed a route would be a waste of resources. Therefore, the completed route of the robot can be obtained in real time and marked accordingly on the route map, without collecting the cloud signal data corresponding to the completed route on the route map.

[0055] In one embodiment, a cloud-based information map is generated for the first time before the robot starts a task. When it is determined that the robot has completed the task, the generation or updating of the cloud-based information map stops. If a return trip or other similar event is performed, it is considered as starting a new task.

[0056] By utilizing cloud-based information maps generated for robots, robots can learn about operator information in various areas along their routes in advance, facilitating early handover preparations and improving signal handover efficiency.

[0057] Step 3: Configure the corresponding AI model, multi-network aggregation card, and acquisition module at the robot. To distinguish it from the cloud AI model, mark the AI ​​model at the robot as the local AI model. To distinguish the input data of the local AI model from the input data of the cloud AI model (cloud signal data), mark the input data of the local AI model as line signal data. The acquisition module is used to collect the corresponding line signal data in real time. Prepare for signal switching based on the cloud information map.

[0058] The multi-network aggregation card supports 4G / 5G networks of multiple operators (China Mobile / China Unicom / China Telecom) and has hardware-level multi-link aggregation capabilities. It can also support 3G and future 6G networks as needed.

[0059] In one embodiment, the selection and deployment of the multi-network aggregation card are configured using existing methods to enable subsequent switching of the corresponding operator signal based on the analysis results; for example:

[0060] Embedded integration: The multi-network aggregation card is directly embedded into the robot's motherboard and connected to the main control hardware via a 5*6 surface-mount chip or a standard SIM card slot. It communicates with the main control AP via the standard ISO7618 protocol, making it suitable for space-constrained scenarios (such as inspection robots and drones).

[0061] External device connection: Connect to an independent multi-card aggregation router (such as Zhongyi 4G aggregation router or 4G aggregation router onboard) via WIFI or Ethernet interface, suitable for scenarios that require flexible replacement or upgrade of the network (such as logistics vehicles, emergency command vehicles).

[0062] In one embodiment, the local AI model is also built using existing AI technologies. The difference between the local AI model and the cloud-based AI model is that the local AI model is built for a specific robot and optional tasks.

[0063] In one embodiment, signal handover preparation processing based on a cloud-based information map includes:

[0064] Step SA1: Mark the current robot position in real time on the cloud information map, identify the information equivalent regions the robot has in the remaining task path, including the information equivalent region where the current position is located; obtain the robot's status information, mainly the speed-related status, to determine which information equivalent regions to switch to for analysis, such as a preset time period, estimate the corresponding distance based on the status information and the preset time period, determine the information equivalent regions passed through based on the distance, and mark the corresponding information equivalent regions as switching preparation analysis regions; this is to avoid invalid or excessive analysis due to changes in information equivalent regions later; alternatively, a fixed distance can be preset to determine the switching preparation analysis regions.

[0065] Step SA2: Identify the signal priority sequence and cloud signal data corresponding to each switching preparation analysis area; obtain the operator signal currently connected to the robot;

[0066] When the operator's signal ranks first in the signal priority sequence of the corresponding handover preparation analysis area, no signal handover preparation is required in the assessment.

[0067] When the operator's signal is not ranked first in the signal priority sequence of the corresponding handover preparation analysis area, the cloud signal data of the current operator's signal in each handover preparation analysis area is obtained. Since it is part of the overall cloud signal information, it is marked as application signal data, that is, the data corresponding to the operator information in the cloud signal data. The application signal data is evaluated to determine whether it meets the preset robot task requirements, such as the requirements for signal strength, signal-to-noise ratio, latency, and packet loss rate. The application signal data is compared with these requirements to determine whether they meet the robot task requirements. The robot task requirements are set in advance by the staff according to the task needs before the robot departs. For example, if the latency is >200ms, real-time control commands (such as robot turning) may fail. Therefore, for tasks that require real-time control, the latency should not exceed 200ms.

[0068] When the applied signal data meets the robot's task requirements, no signal switching preparation is needed during the evaluation.

[0069] When the application signal data does not meet the robot's task requirements, an assessment is made to prepare for signal switching, and the operator's signal ranked first in the corresponding signal priority sequence is used as the backup switching signal.

[0070] Step SA3: Mark the handover preparation analysis area that needs to be prepared for signal handover in the cloud information map, and process it according to the preset signal handover preparation method; the specific signal handover preparation method can be preset manually or through other methods.

[0071] By iterating through steps SA1 to SA2, the switching preparation analysis area is determined in real time based on the robot's position. Based on the current operator signal of the robot, it is determined whether each switching preparation analysis area needs to be prepared for signal switching, and corresponding markings are made on the cloud information map. When the next area needs to be switched for signal switching, the corresponding switching preparation is carried out, and the corresponding signal switching is carried out. At this time, the robot switches to the new operator signal and re-evaluates; thus realizing real-time preparation processing.

[0072] In one embodiment, assessing whether the application signal data meets the preset robot task requirements can be done by making a direct judgment based on existing methods.

[0073] In one embodiment, evaluating whether the applied signal data meets the preset robot task requirements includes:

[0074] Establish a task calibration model, the expression of which is:

[0075] ;

[0076] In the formula: (s, KB) is the input data, s is the application signal data, and KB is the robot task requirement; s→KB indicates that the application signal data meets the robot task requirement; the output data is the task calibration value WR(s, KB), and the task calibration value is 1 or 0; the failure to meet the robot task requirement is considered an abnormal situation, and the corresponding training set is set up for training based on the historical data of whether the corresponding historical task signal is qualified.

[0077] The relevant application signal data and robot task requirements are integrated into the input data and input into the task calibration model for analysis to obtain the corresponding task calibration values.

[0078] When the task calibration value is 1, the evaluation application signal data meets the robot task requirements;

[0079] When the task calibration value is 0, the evaluation application signal data does not meet the robot task requirements.

[0080] Step 4: Collect line signal data in real time through the acquisition module, supplement the line signal data in the preset route map, and generate a local information map. That is, form corresponding signal equivalent areas according to the line signal data, and mark them as local equivalent areas for differentiation; analyze the local information map through the local AI model to obtain the priority of the operator's signal in each local equivalent area in the local information map, and integrate them into a local signal priority sequence; supplement the local signal priority sequence into the local information map.

[0081] Step 5: Prepare for signal handover based on the local information map. When it is necessary to switch carrier signals, perform the corresponding signal handover.

[0082] For example, when the robot approaches the boundary of the switching preparation analysis area (e.g., 500 meters from the tunnel entrance), the switching preparation is automatically triggered, for example:

[0083] Activate backup links in advance (such as switching from 4G to 5G or satellite communication).

[0084] Increase the main link signal monitoring frequency (from once per second to 10 times per second).

[0085] The corresponding signal is switched during the switching process.

[0086] In one embodiment, signal handover preparation processing based on a local information map is performed in a manner similar to signal handover preparation processing based on a cloud information map; and when the result of signal handover preparation processing based on a local information map differs from the result of signal handover preparation processing based on a cloud information map, the result of signal handover preparation processing based on a local information map shall be taken as the standard.

[0087] The process of preparing for signal switching based on cloud-based information maps is merely preliminary processing; the actual processing still relies on real-world data.

[0088] The switching preparation process of this invention avoids frequent signal switching based solely on AI evaluation.

[0089] In one embodiment, when the robot is unable to perform local analysis due to various special reasons during the task, such as abnormal acquisition module issues, the following switching process can be adopted: identify cloud signal data of each signal equivalent area in the cloud information map, analyze the cloud signal data through the local AI model, adjust the signal priority sequence of the corresponding signal equivalent area, and perform signal switching preparation processing based on the adjusted cloud information map;

[0090] When the local AI model cannot meet the analysis requirements, i.e., due to reasons such as local AI model abnormality, the analysis requirements cannot be met, the result of signal switching preparation based on the cloud information map in step three shall be processed.

[0091] By combining cloud-based information maps, the robot can still perform signal switching analysis and processing during the task even when signal switching analysis anomalies occur, thus responding to corresponding emergencies and providing time for subsequent emergency handling.

[0092] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0093] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for seamless cross-carrier signal handover based on an AI multi-network aggregation card, characterized in that, The methods include: Step 1: Obtain the user's service scope and robot information, and configure the corresponding cloud AI model in the cloud based on the robot information and service scope; Step 2: Acquire task information of each robot at the user's location in real time, upload the task information to the cloud for analysis, and obtain the cloud information map of the robot. The cloud information map includes the signal priority sequence of each signal-equivalent area on the robot's task route; store the cloud information map in the robot's database. Step 3: Configure the corresponding local AI model, multi-network aggregation card and acquisition module on the robot. The acquisition module is used to collect the corresponding line signal data in real time; perform signal switching preparation processing based on the cloud information map in the robot database; Step 4: Real-time acquisition of line signal data is performed using the acquisition module. This data is then added to the preset route map to generate a local information map. This involves creating corresponding signal-equivalent regions based on the line signal data, which are then marked as local equivalent regions for differentiation. The local information map includes these local equivalent regions and the corresponding line signal data. A local AI model is used to analyze the line signal data in the local information map to obtain the local signal priority sequence for each local equivalent region. This local signal priority sequence is then added to the local information map. Step 5: Prepare for signal handover based on the local information map. When it is necessary to switch carrier signals, perform the corresponding signal handover. Upload task information to the cloud for analysis, including: The received task information is identified in the cloud to obtain the corresponding robot information and task information. The robot's task route is identified based on the task information, and a route map is generated based on the task route. The system collects cloud signal data from various locations on the route map in real time and marks the cloud signal data accordingly on the route map. Based on robot information and task information, the system calls the corresponding cloud AI model, analyzes the cloud signal data from various locations using the cloud AI model, obtains the priority of each operator's information in the location area, and generates a signal priority sequence for the location area based on the priority of each operator's information. Merge adjacent locations with the same signal priority sequence to obtain the corresponding merged region, and mark the merged region as a signal equivalent region; mark the signal equivalent region and the signal priority sequence accordingly on the route diagram; mark the current route diagram as a cloud information diagram.

2. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 1, characterized in that, The cloud AI model can be one or more. When there are multiple cloud AI models, each cloud AI model is labeled with the appropriate robot and task.

3. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 1, characterized in that, The cloud information map is dynamically updated based on real-time collected cloud signal data, and the updated cloud information map is synchronously updated to the robot's database.

4. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 1, characterized in that, The completed route of the robot is acquired in real time, and the completed route is marked on the route map accordingly. The cloud signal data corresponding to the completed route on the route map is no longer collected.

5. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 1, characterized in that, The cloud-based information map is generated for the first time before the robot starts the task, and stops being generated when the robot has completed the task.

6. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 1, characterized in that, Based on the cloud-based information map, prepare for signal switching, including: Step SA1: Mark the current position of the robot in real time in the cloud information map, identify the signal equivalent area of ​​the robot in the remaining task line; obtain the robot's status information, and determine the switching preparation analysis area in real time based on the status information; Step SA2: Identify the signal priority sequence and cloud signal data corresponding to each switching preparation analysis area; obtain the operator signal currently connected to the robot; When the operator's signal ranks first in the signal priority sequence of the corresponding handover preparation analysis area, no signal handover preparation is required. When the operator signal is not ranked first in the signal priority sequence of the corresponding handover preparation analysis area, the application signal data of the operator signal in each handover preparation analysis area is obtained; the application signal data is evaluated to see if it meets the preset robot task requirements. When the applied signal data meets the robot's task requirements, no signal switching preparation is needed during the evaluation. When the application signal data does not meet the robot's task requirements, an assessment is made to prepare for signal switching, and the operator's signal ranked first in the corresponding signal priority sequence is used as the backup switching signal. Step SA3: Mark the handover preparation analysis area that needs to be prepared for signal handover in the cloud information map, and process it according to the preset signal handover preparation method.

7. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 6, characterized in that, Evaluate whether the applied signal data meets the preset robot task requirements, including: Establish a task calibration model, the expression of which is: ； In the formula: (s, KB) is the input data, s is the application signal data, and KB is the robot task requirement; s→KB indicates that the application signal data meets the robot task requirement; the output data is the task calibration value WR(s, KB), and the task calibration value is 1 or 0. The relevant application signal data and robot task requirements are integrated into the input data and input into the task calibration model for analysis to obtain the corresponding task calibration values. When the task calibration value is 1, the evaluation application signal data meets the robot task requirements; When the task calibration value is 0, the evaluation application signal data does not meet the robot task requirements.

8. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 1, characterized in that, When the result of signal handover preparation processing based on the local information map differs from the result of signal handover preparation processing based on the cloud information map, the result of signal handover preparation processing based on the local information map shall be taken as the standard.

9. The method for seamless cross-carrier signal handover based on an AI multi-network aggregation card according to claim 1, characterized in that, When the signal switching analysis at the robot is abnormal, the cloud signal data of each signal equivalent area in the cloud information map is identified, and the cloud signal data is analyzed by the local AI model to obtain a new signal priority sequence; signal switching preparation is carried out according to the adjusted cloud information map. When the local AI model cannot meet the analysis requirements, the signal switching preparation process based on the cloud information graph in step three is carried out.

Images