Method and system for device movement with autonomous navigation

Abstract

This invention discloses a device mobility method and system with autonomous navigation, relating to the field of device mobility technology. The key technical points include the following steps: acquiring task text information received by the mobile device; retrieving a set of navigation strategies to be tested that matches the task text information from a navigation strategy database; if the task text information contains a single task objective and corresponding basic scene information, performing semantic segmentation and keyword extraction on the task text information to obtain task semantic feature one; evaluating the scene adaptability of task semantic feature one and the set of navigation strategies to be tested to obtain strategy adaptability value one; if the task text information contains multiple related task objectives and cross-scene switching requirements, analyzing the semantic correlation of each related task objective to obtain task semantic feature two; and evaluating the cross-scene switching time and path constraints to obtain strategy adaptability value two; the effect is to enable the device to complete the mobility task using the optimal path.

Description

Technical Field

[0001] This invention relates to the field of equipment mobility technology, and more specifically, to a method and system for moving equipment with autonomous navigation. Background Technology

[0002] In the field of autonomous navigation device mobility, traditional navigation methods often have numerous limitations. With increasingly complex task scenarios, situations arise involving multiple related task objectives and cross-scene switching requirements. Traditional navigation methods struggle to effectively analyze the semantic relationships between tasks and cannot adequately integrate cross-scene time and path constraints for accurate navigation. Consequently, they cannot rationally arrange task order or optimize cross-scene paths, leading to low task completion efficiency and even failure to complete tasks on time. Traditional methods also struggle to accurately judge based on historical task data and real-time scene feedback, resulting in inappropriate navigation strategy selection and impacting the device's mobility. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a method and system for moving equipment with autonomous navigation.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A method for moving a device with autonomous navigation, the method comprising the following steps:

[0006] Obtain the task text information received by the mobile device, and retrieve the set of navigation strategies to be tested that match the task text information from the navigation strategy database;

[0007] If the task text information contains a single task objective and the corresponding basic scene information, then the task text information is semantically segmented and keywords are extracted to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scene adaptation of the navigation strategy set to be tested based on task semantic feature one.

[0008] If the task text information contains multiple related task objectives and cross-scene switching requirements, the semantic correlation of each related task objective is analyzed to obtain task semantic feature two, and the strategy adaptation value two is obtained by combining the time constraint and path constraint of cross-scene switching; wherein, the set of strategy adaptation value one and strategy adaptation value two is a comprehensive strategy adaptation value.

[0009] If there are uncertain semantic expressions or ambiguous keywords in the task text information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scenario prediction, respectively, by using historical task data and real-time scenario feedback information, to obtain the semantic association evaluation value.

[0010] Based on the semantic association evaluation value and the comprehensive strategy adaptation value, the optimal target navigation strategy for the mobile device is selected from the set of navigation strategies to be tested.

[0011] Preferably, the task text information includes task target keywords and scenario description keywords.

[0012] Preferably, if the task text information contains a single task objective and corresponding basic scene information, then semantic segmentation and keyword extraction are performed on the task text information to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scene adaptability of the task semantic feature one and the navigation strategy set to be tested. Specifically, this includes the following steps:

[0013] Basic scene information is determined based on scene description keywords. If the task text information contains a single task objective and corresponding basic scene information, the basic scene information includes scene type, obstacle distribution within the scene, and scene accessibility.

[0014] By retaining task objective keywords and scenario description keywords from the task text information, we obtain task semantic feature one;

[0015] Calculate the scene fit between the navigation strategy set to be tested and the task semantic feature one in the navigation strategy set to be tested; wherein, the scene fit includes path matching rate and scene rule conformity;

[0016] The strategy adaptation value is evaluated based on the weighted calculation result of the scenario adaptation.

[0017] Preferably, the semantic relevance of each associated task objective is analyzed to obtain task semantic feature two, specifically including the following steps:

[0018] The association between keywords of each associated task objective is determined by the semantic association between them, thus obtaining the associated task objectives;

[0019] After extracting scene switching keywords from the task text information, cross-scene switching requirements are determined; wherein, the cross-scene switching requirements include scene switching order, scene switching duration threshold, and switching path priority;

[0020] Semantic relevance scoring is performed on the associated task objectives to obtain the associated semantic score value;

[0021] The second task semantic feature is obtained based on the associated semantic score, cross-scene switching requirements, cross-scene switching time constraints, and cross-scene switching path constraints.

[0022] Preferably, the strategy adaptation value two is obtained by combining the time constraints and path constraints of cross-scene switching, specifically including the following steps:

[0023] Calculate the satisfaction degree of the navigation strategy set to be tested with the second semantic feature of the task. The satisfaction degree includes the efficiency of connecting related tasks and the optimization rate of cross-scene paths. The strategy adaptation value two is evaluated based on the comprehensive calculation result of the satisfaction degree.

[0024] Preferably, if the task text information contains uncertain semantic expressions or ambiguous keywords, then by using historical task data and real-time scene feedback information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scene prediction dimensions to obtain a semantic association evaluation value. Specifically, this includes the following steps:

[0025] If there are uncertain semantic expressions in the task text information, the positive utility value of the navigation strategy set to be tested in the semantic completion dimension is determined by the semantic parsing results of the same uncertain expressions in historical task data.

[0026] If there are ambiguous keywords in the task text information, the scene matching probability is determined based on the real-time scene feedback information, and the negative interference risk of the navigation strategy set to be tested in the scene prediction dimension is determined based on the scene matching probability.

[0027] If the positive utility value is greater than or equal to the negative interference risk, then the statistical comprehensive semantic optimization degree value can obtain a positive evaluation.

[0028] If the positive utility value is less than the negative interference risk, then the statistical comprehensive semantic optimization degree value will be negative evaluation one;

[0029] If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are greater than or equal to the negative interference risk, then the combined utility of statistical semantic completion and scenario prediction is positively evaluated.

[0030] If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are less than the negative interference risk, then the superimposed interference risk of statistical semantic completion and scenario prediction is negatively evaluated.

[0031] Among them, positive evaluation 1, positive evaluation 2, negative evaluation 1 and negative evaluation 2 are combined into a semantic association evaluation value.

[0032] Preferably, the method further includes the following steps:

[0033] Obtain scene environment parameters for real-time location of mobile devices;

[0034] The dynamic correlation value of the scene is obtained by evaluating the dynamic change trend of the scene environment parameters.

[0035] Preferably, the optimal target navigation strategy for the mobile device is selected from the set of navigation strategies to be tested based on the semantic association evaluation value and the comprehensive strategy adaptation value, specifically including the following steps:

[0036] Based on the strategy adaptation value and the scenario description keywords, the adaptation status between the navigation strategy set to be tested and the current task objective is determined to obtain the test trend value.

[0037] Preprocessed trend value 1 is obtained by adjusting the error of the target trend value 1 based on the semantic association evaluation value and the scene dynamic association value.

[0038] Based on the correlation between the strategy adaptation value 2 and the task objective, the overall satisfaction of the navigation strategy set under test with the real-time task is determined to obtain the trend value 2 under test;

[0039] Preprocessed trend value 2 is obtained by adjusting the error of the target trend value 2 based on the semantic association evaluation value and the scene dynamic association value;

[0040] If preprocessing trend one and preprocessing trend two are positive adaptation trends, then the optimal target navigation strategy for the mobile device is obtained from the set of navigation strategies to be tested based on the maximum value of preprocessing trend one or preprocessing trend two.

[0041] If preprocessing trend one and preprocessing trend two belong to negative adaptation trends, then the optimal target navigation strategy for the mobile device is obtained from the set of navigation strategies to be tested based on the minimum value of preprocessing trend one or preprocessing trend two.

[0042] Equipment mobility systems with autonomous navigation include:

[0043] Retrieval module: Acquires task text information received by the mobile device and retrieves the set of navigation strategies to be tested that match the task text information from the navigation strategy database;

[0044] First evaluation module: If the task text information contains a single task objective and corresponding basic scene information, then the task text information is semantically segmented and keywords are extracted to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scenario adaptation of the task semantic feature one and the navigation strategy set to be tested.

[0045] The second evaluation module: If the task text information contains multiple related task objectives and cross-scene switching requirements, the semantic correlation of each related task objective is analyzed to obtain the second task semantic feature, and the strategy adaptation value is obtained by combining the time constraint and path constraint of cross-scene switching; wherein, the set of the first strategy adaptation value and the second strategy adaptation value is a comprehensive strategy adaptation value.

[0046] Processing module: If there are uncertain semantic expressions or ambiguous keywords in the task text information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scene prediction dimensions, respectively, through historical task data and real-time scene feedback information, to obtain the semantic association evaluation value.

[0047] Strategy generation module: Based on semantic association evaluation value and comprehensive strategy adaptation value, it selects the optimal target navigation strategy for the mobile device from the set of navigation strategies to be tested.

[0048] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method for moving a device with autonomous navigation.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] This invention acquires task text information received by a mobile device and retrieves a set of matching navigation strategies from a navigation strategy database. This allows for rapid identification of applicable navigation strategies, avoiding indiscriminate traversal of all strategies and significantly improving the efficiency of subsequent strategy evaluation and selection. Targeted evaluation methods are used for different types of task text information, enhancing the comprehensiveness and accuracy of strategy adaptation. When the task text contains a single task objective and corresponding basic scenario information, semantic segmentation and keyword extraction are performed to obtain task semantic feature one. This is then combined with scenario adaptability to evaluate strategy adaptability value one, considering the strategy's suitability in a single task scenario. When the task involves multiple related task objectives and cross-scenario switching requirements, the semantic correlation of each related task objective is analyzed to obtain task semantic feature two. This is then combined with cross-scenario time and path constraints to evaluate strategy adaptability value two. This allows for accurate evaluation of strategy adaptability in complex multi-task, cross-scenario situations. The optimal target navigation strategy is selected by combining the semantic correlation evaluation value and the comprehensive strategy adaptability value, ensuring the optimality of the final selected strategy. This comprehensive approach, which considers multi-dimensional evaluation results, can select the most suitable navigation strategy for mobile devices in various complex task scenarios, ensuring the efficiency, accuracy, and stability of autonomous navigation, enabling the device to complete the mobile task with the optimal path, and improving the device's autonomous operation capability and practicality. Attached Figure Description

[0051] Figure 1 This is a schematic diagram illustrating the steps of the device movement method with autonomous navigation proposed in this invention;

[0052] Figure 2 This is a schematic diagram of the module of the device mobility system with autonomous navigation proposed in this invention.

[0053] 610. Processor; 620. Communication interface; 630. Memory; 640. Communication bus. Detailed Implementation

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0055] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0056] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0057] Reference Figures 1-2 .

[0058] The embodiments further illustrate the device movement method and system with autonomous navigation proposed in this invention.

[0059] A method for moving a device with autonomous navigation, the method comprising the following steps:

[0060] Obtain the task text information received by the mobile device, and retrieve the set of navigation strategies to be tested that match the task text information from the navigation strategy database;

[0061] If the task text information contains a single task objective and the corresponding basic scene information, then the task text information is semantically segmented and keywords are extracted to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scene adaptation of the navigation strategy set to be tested based on task semantic feature one.

[0062] If the task text information contains multiple related task objectives and cross-scene switching requirements, the semantic correlation of each related task objective is analyzed to obtain the second task semantic feature, and the strategy adaptation value is obtained by combining the time constraint and path constraint of cross-scene switching; where the set of strategy adaptation value one and strategy adaptation value two is the comprehensive strategy adaptation value.

[0063] If there are uncertain semantic expressions or ambiguous keywords in the task text information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scenario prediction, respectively, by using historical task data and real-time scenario feedback information, to obtain the semantic association evaluation value.

[0064] Based on the semantic association evaluation value and the comprehensive strategy adaptation value, the optimal target navigation strategy for the mobile device is selected from the set of navigation strategies to be tested.

[0065] This application first determines basic scene information based on scene description keywords, including scene type, obstacle distribution within the scene, and scene traffic conditions. The task text information is processed, retaining task target keywords and scene description keywords to form task semantic feature one. The scene fit degree between each strategy in the test navigation strategy set and task semantic feature one is calculated. This fit degree includes path matching rate and scene rule compliance. Finally, a strategy fit value one is evaluated based on the weighted calculation result of the scene fit degree.

[0066] If the task text contains multiple related task objectives and cross-scene switching requirements, the semantic relevance of each related task objective is judged to determine the association between task objective keywords, thus obtaining the related task objectives. Simultaneously, scene switching keywords are extracted from the task text to determine cross-scene switching requirements, including scene switching order, scene switching duration threshold, and switching path priority. Semantic relevance scoring is performed on the related task objectives to obtain a related semantic score. This score, combined with the related semantic score, cross-scene switching requirements, cross-scene switching time constraints, and cross-scene switching path constraints, forms the second task semantic feature. Subsequently, the satisfaction level of the navigation strategy set under test with the second task semantic feature is calculated. Satisfaction includes related task connection efficiency and cross-scene path optimization rate. Based on the comprehensive calculation result of the satisfaction level, the second strategy adaptation value is evaluated. The first and second strategy adaptation values ​​together constitute the comprehensive strategy adaptation value.

[0067] If the task text contains uncertain semantic expressions or ambiguous keywords, it is processed using historical task data and real-time scene feedback information. If uncertain semantic expressions exist, the positive utility value of the navigation strategy set under test in the semantic completion dimension is determined based on the semantic parsing results of the same uncertain expressions in historical task data. If ambiguous keywords exist, the scene matching probability is judged based on real-time scene feedback information, thereby determining the negative interference risk of the navigation strategy set under test in the scene prediction dimension. Statistical analysis is performed based on the comparison between positive utility value and negative interference risk. If the positive utility value is greater than or equal to the negative interference risk, a positive evaluation is obtained by statistically analyzing the comprehensive semantic optimization degree; if the positive utility value is less than the negative interference risk, a negative evaluation is obtained by statistically analyzing the comprehensive semantic optimization degree. If both uncertain semantic expressions and ambiguous keywords exist simultaneously, and the positive utility values ​​corresponding to both are greater than or equal to the negative interference risk, the combined utility of statistical semantic completion and scenario prediction yields a positive evaluation II; if the positive utility values ​​corresponding to both are less than the negative interference risk, the combined interference risk of statistical semantic completion and scenario prediction yields a negative evaluation II. These evaluation results together constitute the semantic association evaluation value.

[0068] Obtain scene environment parameters from the real-time location of mobile devices, evaluate the dynamic change trends of these parameters, and obtain scene dynamic correlation values.

[0069] The optimal target navigation strategy is selected based on semantic association evaluation values ​​and comprehensive strategy fit values. Specifically, the fit between the navigation strategy set under test and the current task target is determined based on strategy fit value one and scene description keywords, thus obtaining test trend value one. Error adjustment is performed on test trend value one using semantic association evaluation values ​​and scene dynamic association values, resulting in preprocessed trend value one. The comprehensive satisfaction of the navigation strategy set under test with the real-time task is determined based on the relationship with the task target, resulting in test trend value two. Similarly, error adjustment is performed using semantic association evaluation values ​​and scene dynamic association values, resulting in preprocessed trend value two. If preprocessed trend value one and preprocessed trend value two are positive fit trends, the optimal target navigation strategy is selected from the navigation strategy set under test based on the maximum value; if they are negative fit trends, the minimum value is used for selection.

[0070] The task text information includes task objective keywords and scenario description keywords.

[0071] The task text information includes task target keywords and scenario description keywords, which provide key information for accurately grasping the core objectives of the task and the scenario environment in which the task is located. Subsequent processes such as retrieving matching navigation strategy sets from the navigation strategy database, conducting semantic analysis, and evaluating strategy adaptability all rely on the extraction of these task target keywords and scenario description keywords, so as to plan a navigation strategy that meets the task requirements for mobile devices.

[0072] If the task text information contains a single task objective and its corresponding basic scene information, then semantic segmentation and keyword extraction are performed on the task text information to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scene adaptability of the navigation strategy set to be tested, specifically including the following steps:

[0073] Determine basic scene information based on scene description keywords. If the task text information contains a single task objective and its corresponding basic scene information, the basic scene information includes scene type, obstacle distribution within the scene, and scene accessibility.

[0074] By retaining task objective keywords and scenario description keywords from the task text information, we obtain task semantic feature one;

[0075] Calculate the scene fit between the navigation strategy set to be tested and the task semantic feature 1 in the navigation strategy set to be tested; where scene fit includes path matching rate and scene rule conformity.

[0076] The strategy adaptation value is evaluated based on the weighted calculation result of the scenario adaptation.

[0077] If the task text information contains a single task objective and corresponding basic scene information, the basic scene information is first determined based on scene description keywords. This basic scene information specifically includes the scene type, the distribution of obstacles within the scene, and the scene's accessibility. The scene type includes indoor warehouses, outdoor parks, or other specific scenes. For example, the distribution of obstacles within the scene refers to the placement and quantity of shelves in a warehouse, or the distribution of trees or facilities in a park. The scene's accessibility includes whether the roads are wide and whether there are congested areas. The task text information is processed, retaining the task objective keywords and scene description keywords to obtain task semantic feature one. The scene adaptability of each navigation strategy in the test strategy set with task semantic feature one is calculated. The path matching rate is the degree to which the path planned by the navigation strategy matches the ideal path from the starting position to the task objective position; the scene rule compliance rate is whether it complies with speed limits, avoidance of workstations, or rules for specific areas within the scene. Based on the weighted calculation result of the scene adaptability, a strategy adaptability value one is obtained, which is used to determine the adaptability of the navigation strategy to the current single task objective and basic scene.

[0078] The semantic relevance of each associated task objective is then analyzed to obtain task semantic feature two, which specifically includes the following steps:

[0079] The association between keywords of each associated task objective is determined by the semantic association between them, thus obtaining the associated task objectives;

[0080] After extracting scene switching keywords from the task text information, cross-scene switching requirements are determined; among them, cross-scene switching requirements include scene switching order, scene switching duration threshold, and switching path priority;

[0081] Semantic relevance scoring is performed on the associated task objectives to obtain the associated semantic score value;

[0082] The second task semantic feature is obtained based on the associated semantic score, cross-scene switching requirements, cross-scene switching time constraints, and cross-scene switching path constraints.

[0083] If the task text information involves multiple related task objectives and cross-scenario switching requirements, a deep semantic analysis should first be conducted on each related task objective. Semantic analysis techniques are used to interpret the connotations of each task objective's keywords, exploring the underlying logical connections between them, such as the order of task execution and causal relationships. This clarifies the relationships between the keywords of each task objective, thereby identifying the related task objectives.

[0084] Extract scene switching-related keywords from the task text information, such as expressions like "transfer to" and "arrive," and use these keywords to define cross-scene switching requirements. Cross-scene switching requirements include scene switching order, scene switching duration threshold, and switching path priority. Scene switching order specifies which scene to switch from first and which to next; scene switching duration threshold specifies the maximum allowed time to complete the scene switch; switching path priority determines whether to prioritize the shortest path or the path with the smoothest passage.

[0085] A semantic relevance scoring operation is performed on related task objectives. This scoring mechanism comprehensively considers factors such as the tightness of the logical connection between task objectives and the similarity of semantic content, ultimately deriving a semantic relevance score. A higher score indicates a stronger relevance between task objectives, which better reflects the coherence of the task when planning navigation strategies.

[0086] This paper integrates elements such as associated semantic score, cross-scene switching requirements, cross-scene switching time constraints, and cross-scene switching path constraints. A mathematical model is then constructed to quantify these elements, ultimately generating task semantic feature two. Task semantic feature two comprehensively and meticulously reflects the specific circumstances of multiple associated task objectives and cross-scene switching requirements, providing crucial evidence for subsequent evaluation of the navigation strategy's suitability and ensuring that the navigation strategy can meet the complex needs of multiple tasks and cross-scenes.

[0087] The strategy adaptation value two is obtained by combining the time constraints and path constraints of cross-scene switching, and specifically includes the following steps:

[0088] Calculate the satisfaction degree of the navigation strategy set to be tested with the second semantic feature of the task. The satisfaction degree includes the efficiency of connecting related tasks and the optimization rate of cross-scene paths. The strategy adaptation value two is evaluated based on the comprehensive calculation result of the satisfaction degree.

[0089] Satisfaction primarily includes the efficiency of interconnected task transitions and the cross-scene path optimization rate. The efficiency of interconnected task transitions focuses on the smoothness and efficiency of the navigation strategy in handling multiple related tasks. For example, after completing one task, can the navigation strategy quickly and without redundancy guide the device into the next task flow? Is the transition time within a reasonable range? Are there unnecessary pauses or detours? These factors all affect the efficiency of interconnected task transitions. Secondly, the cross-scene path optimization rate concerns the optimization level of the cross-scene movement path planned by the navigation strategy in terms of distance, travel time, and device energy consumption. This means judging whether the planned path can help the device complete cross-scene movement with a shorter distance, less time, or lower energy consumption, and whether it minimizes the adverse factors of congested road sections and obstacle areas. A comprehensive calculation is performed on the efficiency of interconnected task transitions and the cross-scene path optimization rate. Based on the results of this comprehensive calculation, the strategy adaptation value is evaluated and determined, thus clearly determining the adaptability of the navigation strategy set to tasks containing multiple related task objectives and cross-scene switching requirements.

[0090] If the task text contains uncertain semantic expressions or ambiguous keywords, then by using historical task data and real-time scenario feedback information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scenario prediction, respectively, to obtain the semantic association evaluation value. The specific steps include:

[0091] If there are uncertain semantic expressions in the task text information, the positive utility value of the navigation strategy set to be tested in the semantic completion dimension is determined by the semantic parsing results of the same uncertain expressions in historical task data.

[0092] If there are ambiguous keywords in the task text information, the scene matching probability is determined based on the real-time scene feedback information, and the negative interference risk of the navigation strategy set to be tested in the scene prediction dimension is determined based on the scene matching probability.

[0093] If the positive utility value is greater than or equal to the negative interference risk, then the statistical comprehensive semantic optimization degree value can obtain a positive evaluation.

[0094] If the positive utility value is less than the negative interference risk, then the statistical comprehensive semantic optimization degree value will be negative evaluation one;

[0095] If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are greater than or equal to the negative interference risk, then the combined utility of statistical semantic completion and scenario prediction is positively evaluated.

[0096] If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are less than the negative interference risk, then the superimposed interference risk of statistical semantic completion and scenario prediction is negatively evaluated.

[0097] Among them, positive evaluation 1, positive evaluation 2, negative evaluation 1 and negative evaluation 2 are combined into a semantic association evaluation value.

[0098] If the task text contains ambiguous semantic expressions, the semantic parsing results of the same ambiguous expressions are retrieved from historical task data. This determines the positive utility value of the navigation strategy set under test in the semantic completion dimension. This value reflects the effectiveness of the strategy set in reasonably supplementing ambiguous semantics and making the semantics clearer and more accurate. If the task text contains ambiguous keywords, the matching probability of the keyword with the current scene is judged based on the information from real-time scene feedback. The negative interference risk of the navigation strategy set under test in the scene prediction dimension is determined based on the matching probability, that is, the probability that the strategy set will cause erroneous interference when predicting the scene due to the ambiguity of the keywords.

[0099] The positive utility value and negative interference risk are compared. If the positive utility value is greater than or equal to the negative interference risk, the overall semantic optimization level is statistically analyzed, resulting in a positive evaluation of 1. This indicates that the strategy set is optimized and effective in terms of semantic processing as a whole. If the positive utility value is less than the negative interference risk, the overall semantic optimization level is also statistically analyzed, but a negative evaluation of 1 is obtained, indicating that the strategy set has deficiencies in semantic processing.

[0100] If the task text information contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are greater than or equal to their respective negative interference risks, then the combined utility of semantic completion and scene prediction is statistically analyzed to obtain a second positive evaluation. This means that the strategy set can effectively play a role in both semantic completion and scene prediction, and the combined effect is better. If the positive utility values ​​corresponding to both are less than their respective negative interference risks, then the combined interference risk of semantic completion and scene prediction is statistically analyzed to obtain a second negative evaluation. This indicates that the navigation strategy set under test has interference risks in both aspects, and the negative impact after combination is more significant. Finally, the first positive evaluation, the second positive evaluation, the first negative evaluation, and the second negative evaluation are integrated to form a semantic association evaluation value, which is used to comprehensively measure the performance of the navigation strategy set in terms of semantic association.

[0101] It also includes the following steps:

[0102] Obtain scene environment parameters for real-time location of mobile devices;

[0103] The dynamic correlation value of the scene is obtained by evaluating the dynamic change trend of the scene environment parameters.

[0104] This application acquires scene environment parameters from real-time positioning of mobile devices. These parameters include, but are not limited to, light intensity, temperature, humidity, real-time location and status of obstacles, traffic flow, and ground flatness within the scene. Then, it evaluates the dynamic trends of these scene environment parameters. For example, it analyzes whether light intensity is gradually increasing or decreasing, whether temperature is rising or falling, whether obstacles are moving and their direction and speed, and whether traffic flow is increasing or decreasing. Through comprehensive analysis and calculation of these dynamic trends, a scene dynamic correlation value is finally obtained. This value reflects the correlation of scene environment changes over time, providing a basis for subsequent scene-based decision-making.

[0105] The optimal target navigation strategy for the mobile device is selected from the set of navigation strategies to be tested based on the semantic association evaluation value and the comprehensive strategy adaptation value. The specific steps include:

[0106] Based on the strategy adaptation value and the scenario description keywords, the adaptation status between the navigation strategy set to be tested and the current task objective is determined to obtain the test trend value.

[0107] Preprocessed trend value 1 is obtained by adjusting the error of the target trend value 1 based on the semantic association evaluation value and the scene dynamic association value.

[0108] Based on the correlation between the strategy adaptation value 2 and the task objective, the overall satisfaction of the navigation strategy set under test with the real-time task is determined to obtain the trend value 2 under test;

[0109] Preprocessed trend value 2 is obtained by adjusting the error of the target trend value 2 based on the semantic association evaluation value and the scene dynamic association value;

[0110] If preprocessing trend one and preprocessing trend two are positive adaptation trends, then the optimal target navigation strategy for the mobile device is obtained from the set of navigation strategies to be tested based on the maximum value of preprocessing trend one or preprocessing trend two.

[0111] If preprocessing trend one and preprocessing trend two belong to negative adaptation trends, then the optimal target navigation strategy for the mobile device is obtained from the set of navigation strategies to be tested based on the minimum value of preprocessing trend one or preprocessing trend two.

[0112] This application first uses the strategy adaptation value 1 as a foundation, and then combines it with scenario description keywords to assess the adaptation status of each navigation strategy set to be tested with the current task objective. The strategy adaptation value 1 is derived from the evaluation of a single task objective and basic scenario information, reflecting the degree of adaptation of the strategy under the basic task and scenario, while the scenario description keywords define the specific environment for task execution. The application judges whether the path planned by the strategy conforms to the traffic rules within the scenario and whether it efficiently completes the single task objective. By comprehensively considering these factors, the test trend value 1 is derived, which reflects the adaptation trend of the strategy under the current task objective.

[0113] Semantic association evaluation values ​​and scene dynamic association values ​​are introduced for error adjustment. The semantic association evaluation value comprehensively considers the impact of semantic uncertainty and ambiguity in the task text on the strategy; if the task text contains ambiguous expressions, the semantic association evaluation value can reflect the strategy's ability to handle such problems. The scene dynamic association value reflects the changing trends of scene environmental parameters, such as the narrowing of passageways due to the temporary addition of handling vehicles in the warehouse. The preprocessed trend value is obtained by integrating and analyzing the target trend value with these two key values ​​to correct errors caused by not considering semantic and scene dynamic changes.

[0114] For scenarios involving multiple related task objectives and cross-scene switching requirements, the second strategy adaptation value is used as the core, and the overall satisfaction status of each navigation strategy set under test is judged by combining the relationships between task objectives. The second strategy adaptation value evaluates the strategy's ability to handle complex tasks and cross-scene switching, judging whether the strategy can reasonably arrange the task sequence and efficiently complete cross-scene movement. For example, in the task of transporting goods from the warehouse to the workshop, can it plan the shortest and smoothest cross-scene path, thus deriving the second trend value to be tested.

[0115] Error adjustment is performed on the target trend value 2 based on the semantic association evaluation value and the scene dynamic association value. This step is also to correct the deviation that may be caused by the dynamic changes in semantics and scene, to ensure that the evaluation result is more in line with the actual situation, and finally to obtain the preprocessed trend value 2.

[0116] The decision is made based on the properties of preprocessing trend one and preprocessing trend two. If both preprocessing trend one and preprocessing trend two exhibit positive adaptation trends, it means that these strategies can meet the task requirements to a certain extent. In this case, the navigation strategy set corresponding to the maximum value is selected as the optimal target navigation strategy for the mobile device, because this strategy performs best in the overall evaluation. If both preprocessing trend one and preprocessing trend two exhibit negative adaptation trends, it indicates that these strategies have certain problems in the current task and scenario. In this case, the navigation strategy set corresponding to the minimum value is selected as the optimal target navigation strategy, because this strategy has the least deficiency in meeting the task requirements compared to other strategies.

[0117] Equipment mobility systems with autonomous navigation include:

[0118] Retrieval module: Acquires task text information received by the mobile device and retrieves the set of navigation strategies to be tested that match the task text information from the navigation strategy database;

[0119] First evaluation module: If the task text information contains a single task objective and corresponding basic scene information, then the task text information is semantically segmented and keywords are extracted to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scenario adaptation of the task semantic feature one and the navigation strategy set to be tested.

[0120] The second evaluation module: If the task text information contains multiple related task objectives and cross-scene switching requirements, the semantic correlation of each related task objective is analyzed to obtain the second task semantic feature, and the strategy adaptation value is obtained by combining the time constraints and path constraints of cross-scene switching; where the set of strategy adaptation value one and strategy adaptation value two is the comprehensive strategy adaptation value.

[0121] Processing module: If there are uncertain semantic expressions or ambiguous keywords in the task text information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scene prediction dimensions, respectively, through historical task data and real-time scene feedback information, to obtain the semantic association evaluation value.

[0122] Strategy generation module: Based on semantic association evaluation value and comprehensive strategy adaptation value, it selects the optimal target navigation strategy for the mobile device from the set of navigation strategies to be tested.

[0123] An electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a method for moving the device with autonomous navigation.

[0124] The electronic device may include a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a device movement method with autonomous navigation.

[0125] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0126] On the other hand, the present invention also provides a computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, wherein when the computer program is executed by a processor, the computer is capable of executing a device movement method with autonomous navigation.

[0127] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform a device movement method with autonomous navigation.

[0128] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0129] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for moving equipment with autonomous navigation, characterized in that, The method includes the following steps: Obtain the task text information received by the mobile device, and retrieve the set of navigation strategies to be tested that match the task text information from the navigation strategy database; If the task text information contains a single task objective and the corresponding basic scene information, then the task text information is semantically segmented and keywords are extracted to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scene adaptation of the navigation strategy set to be tested based on task semantic feature one. If the task text information contains multiple related task objectives and cross-scene switching requirements, the semantic correlation of each related task objective is analyzed to obtain task semantic feature two, and the strategy adaptation value two is obtained by combining the time constraint and path constraint of cross-scene switching; wherein, the set of strategy adaptation value one and strategy adaptation value two is a comprehensive strategy adaptation value. If the task text contains uncertain semantic expressions or ambiguous keywords, then by using historical task data and real-time scenario feedback information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scenario prediction, respectively, to obtain the semantic association evaluation value. The specific steps include: If there are uncertain semantic expressions in the task text information, the positive utility value of the navigation strategy set to be tested in the semantic completion dimension is determined by the semantic parsing results of the same uncertain expressions in historical task data. If there are ambiguous keywords in the task text information, the scene matching probability is determined based on the real-time scene feedback information, and the negative interference risk of the navigation strategy set to be tested in the scene prediction dimension is determined based on the scene matching probability. If the positive utility value is greater than or equal to the negative interference risk, then the statistical comprehensive semantic optimization degree value can obtain a positive evaluation. If the positive utility value is less than the negative interference risk, then the statistical comprehensive semantic optimization degree value will be negative evaluation one; If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are greater than or equal to the negative interference risk, then the combined utility of statistical semantic completion and scenario prediction is positively evaluated. If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are less than the negative interference risk, then the superimposed interference risk of statistical semantic completion and scenario prediction is negatively evaluated. Among them, positive evaluation 1, positive evaluation 2, negative evaluation 1 and negative evaluation 2 are combined into a semantic association evaluation value; Based on the semantic association evaluation value and the comprehensive strategy adaptation value, the optimal target navigation strategy for the mobile device is selected from the set of navigation strategies to be tested.

2. The device movement method with autonomous navigation according to claim 1, characterized in that, The task text information includes task objective keywords and scenario description keywords.

3. The device movement method with autonomous navigation according to claim 2, characterized in that, If the task text information contains a single task objective and its corresponding basic scene information, then semantic segmentation and keyword extraction are performed on the task text information to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scene adaptability of the navigation strategy set to be tested, specifically including the following steps: Basic scene information is determined based on scene description keywords. If the task text information contains a single task objective and its corresponding basic scene information, the basic scene information includes scene type, obstacle distribution within the scene, and scene accessibility. By retaining task objective keywords and scenario description keywords from the task text information, we obtain task semantic feature one; Calculate the scene fit between the navigation strategy set to be tested and the task semantic feature one in the navigation strategy set to be tested; wherein, the scene fit includes path matching rate and scene rule conformity; The strategy adaptation value is evaluated based on the weighted calculation result of the scenario adaptation.

4. The device movement method with autonomous navigation according to claim 3, characterized in that, The semantic relevance of each associated task objective is then analyzed to obtain task semantic feature two, which specifically includes the following steps: The association between keywords of each associated task objective is determined by the semantic association between them, thus obtaining the associated task objectives; After extracting scene switching keywords from the task text information, cross-scene switching requirements are determined; wherein, the cross-scene switching requirements include scene switching order, scene switching duration threshold, and switching path priority; Semantic relevance scoring is performed on the associated task objectives to obtain the associated semantic score value; The second task semantic feature is obtained based on the associated semantic score, cross-scene switching requirements, cross-scene switching time constraints, and cross-scene switching path constraints.

5. The device movement method with autonomous navigation according to claim 4, characterized in that, The strategy adaptation value two is obtained by combining the time constraints and path constraints of cross-scene switching, and specifically includes the following steps: Calculate the satisfaction degree of the navigation strategy set to be tested with the second semantic feature of the task. The satisfaction degree includes the efficiency of connecting related tasks and the optimization rate of cross-scene paths. The strategy adaptation value two is evaluated based on the comprehensive calculation result of the satisfaction degree.

6. The device movement method with autonomous navigation according to claim 5, characterized in that, It also includes the following steps: Obtain scene environment parameters for real-time location of mobile devices; The dynamic correlation value of the scene is obtained by evaluating the dynamic change trend of the scene environment parameters.

7. The device movement method with autonomous navigation according to claim 6, characterized in that, The optimal target navigation strategy for the mobile device is selected from the set of navigation strategies to be tested based on the semantic association evaluation value and the comprehensive strategy adaptation value. The specific steps include: Based on the strategy adaptation value and the scenario description keywords, the adaptation status between the navigation strategy set to be tested and the current task objective is determined to obtain the test trend value. Preprocessed trend value 1 is obtained by adjusting the error of the target trend value 1 based on the semantic association evaluation value and the scene dynamic association value. Based on the correlation between the strategy adaptation value 2 and the task objective, the overall satisfaction of the navigation strategy set under test with the real-time task is determined to obtain the trend value 2 under test; Preprocessed trend value 2 is obtained by adjusting the error of the target trend value 2 based on the semantic association evaluation value and the scene dynamic association value; If preprocessing trend one and preprocessing trend two are positive adaptation trends, then the optimal target navigation strategy for the mobile device is obtained from the set of navigation strategies to be tested based on the maximum value of preprocessing trend one or preprocessing trend two. If preprocessing trend one and preprocessing trend two belong to negative adaptation trends, then the optimal target navigation strategy for the mobile device is obtained from the set of navigation strategies to be tested based on the minimum value of preprocessing trend one or preprocessing trend two.

8. A device movement system with autonomous navigation, applied to the device movement method with autonomous navigation as described in any one of claims 1 to 7, characterized in that, include: Retrieval module: Acquires task text information received by the mobile device and retrieves the set of navigation strategies to be tested that match the task text information from the navigation strategy database; First evaluation module: If the task text information contains a single task objective and corresponding basic scene information, then the task text information is semantically segmented and keywords are extracted to obtain task semantic feature one. The strategy adaptation value one is obtained by evaluating the scenario adaptation of the task semantic feature one and the navigation strategy set to be tested. The second evaluation module: If the task text information contains multiple related task objectives and cross-scene switching requirements, the semantic correlation of each related task objective is analyzed to obtain the second task semantic feature, and the strategy adaptation value is obtained by combining the time constraint and path constraint of cross-scene switching; wherein, the set of the first strategy adaptation value and the second strategy adaptation value is a comprehensive strategy adaptation value. Processing module: If the task text contains uncertain semantic expressions or ambiguous keywords, then by using historical task data and real-time scene feedback information, the positive utility value and negative interference risk of the navigation strategy set to be tested are statistically analyzed in terms of semantic completion and scene prediction, respectively, to obtain the semantic association evaluation value. The specific steps include: If there are uncertain semantic expressions in the task text information, the positive utility value of the navigation strategy set to be tested in the semantic completion dimension is determined by the semantic parsing results of the same uncertain expressions in historical task data. If there are ambiguous keywords in the task text information, the scene matching probability is determined based on the real-time scene feedback information, and the negative interference risk of the navigation strategy set to be tested in the scene prediction dimension is determined based on the scene matching probability. If the positive utility value is greater than or equal to the negative interference risk, then the statistical comprehensive semantic optimization degree value can obtain a positive evaluation. If the positive utility value is less than the negative interference risk, then the statistical comprehensive semantic optimization degree value will be negative evaluation one; If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are greater than or equal to the negative interference risk, then the combined utility of statistical semantic completion and scenario prediction is positively evaluated. If the task text contains both uncertain semantic expressions and ambiguous keywords, and the positive utility values ​​corresponding to both are less than the negative interference risk, then the superimposed interference risk of statistical semantic completion and scenario prediction is negatively evaluated. Among them, positive evaluation 1, positive evaluation 2, negative evaluation 1 and negative evaluation 2 are combined into a semantic association evaluation value; Strategy generation module: Based on semantic association evaluation value and comprehensive strategy adaptation value, it selects the optimal target navigation strategy for the mobile device from the set of navigation strategies to be tested.

9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the device movement method with autonomous navigation as described in any one of claims 1 to 7.

Images