Collision-free global path planning method for mobile robot

Abstract

The invention provides a collision-free global path planning method for a mobile robot, belongs to the field of mobile robot path planning, and is used for solving the problems of intention misjudgment, low role allocation efficiency and high collision risk in a dynamic scene caused by dependence on statistical association in related technologies. According to the method, agent intention reasoning is achieved by constructing a mental atlas containing a causal relationship, efficient role allocation is performed based on a quantum role superposition state, path search is optimized in combination with a conflict risk field, role collapse is triggered when a conflict risk exceeds a threshold value, and cooperation of adjacent agents is guaranteed through a quantum entanglement cooperation state. According to the scheme, intention prediction accuracy and role allocation efficiency are improved, and collision-free and efficient global path planning in a dynamic environment is realized.

Description

Technical Field

[0001] The present application relates to the field of mobile robot path planning, and in particular to a collision-free global path planning method for a mobile robot. Background Art

[0002] Existing multi-agent path planning methods for mobile robots often rely on the statistical correlation of trajectory data for decision-making, which makes it difficult to distinguish between true causal relationships and random associations, and can easily lead to misjudgment of intentions; role allocation often adopts a serial trial-and-error mode, which is inefficient in high-density dynamic scenarios and has a high conflict rate, making it difficult to meet the needs of collision-free global path planning in complex environments. Summary of the Invention

[0003] This application provides a collision-free global path planning method for a mobile robot, which can achieve efficient collision-free path planning for mobile robots in complex scenarios by combining causal cognition with quantum parallel decision-making.

[0004] In the first aspect, the present application provides a collision-free global path planning method for a mobile robot, including: constructing a causal mental map to realize intention reasoning between intelligent agents, the causal mental map contains the causal relationship, target belief set, behavior parameter set and collaborative credit score between intelligent agents; assigning roles to intelligent agents based on quantum role superposition state, the quantum role superposition state contains probability amplitudes of multiple roles; combining the results of the intention reasoning with the results of the role assignment to generate a collision-free global path.

[0005] By adopting the above technical solutions, we can use causal mind maps to achieve accurate reasoning of the agent's intentions, overcome the limitations of traditional statistical correlation, and use the parallelism of quantum role superposition states to improve the efficiency of role allocation. The combination of the two provides a reliable foundation for generating collision-free global paths, effectively addressing the path planning challenges in complex scenarios.

[0006] Furthermore, the construction of the causal mental map includes: extracting causal event pairs from the historical trajectory data of the intelligent agent, the causal event pairs including cause events and result events; performing statistical significance verification on the causal event pairs to screen out valid causal relationships that meet a preset significance threshold; and updating the causal mental map based on the valid causal relationships.

[0007] By adopting the above technical solutions, the authenticity and validity of the causal relationship in the causal mental map can be ensured, the interference of false associations on intention reasoning can be reduced, the accuracy of intention reasoning can be improved, and a more reliable basis can be provided for subsequent path planning.

[0008] Furthermore, extracting causal event pairs from the historical trajectory data of the intelligent agent includes: ensuring that the timestamp of the cause event is earlier than the timestamp of the result event through a time synchronization mechanism, and the synchronization accuracy of the time synchronization mechanism is not lower than a preset synchronization time threshold.

[0009] By adopting the above technical solution, the temporal relationship of causal events is strictly guaranteed, the causal reversal caused by time synchronization errors is avoided, the accuracy of the extracted causal event pairs is ensured, and the foundation is laid for the effective construction of causal relationships.

[0010] Furthermore, the intention reasoning includes: generating a counterfactual trajectory of the agent based on counterfactual reasoning, wherein the counterfactual trajectory is a possible predicted trajectory of the agent assuming that a specific cause event has not occurred; and calculating a target probability distribution by combining the counterfactual trajectory with the actual trajectory of the agent, wherein the target probability distribution is used to characterize the potential target of the agent and the corresponding confidence level.

[0011] By adopting the above technical solution, counterfactual reasoning is used to make up for the lack of consideration of events that have not occurred in traditional intention reasoning, and to more comprehensively analyze the intention of the intelligent agent, the calculation of the target probability distribution is made more accurate, thereby improving the reliability of intention prediction.

[0012] Furthermore, the probability amplitude of each role in the quantum role superposition state is determined in the following way: the probability amplitude of each role is calculated based on the marginal contribution value, role fitness and causal effect of the intelligent agent, the marginal contribution value is obtained by calculating the Shapley value, and the role fitness is used to characterize the degree of matching between the behavioral parameters of the intelligent agent and the role.

[0013] By adopting the above technical solution, the role probability amplitude is determined by comprehensively considering multiple factors, so that the quantum role superposition state is more in line with the actual situation and collaboration needs of the intelligent body, and the rationality and effectiveness of role allocation are improved.

[0014] Furthermore, it also includes: when the conflict risk value between intelligent agents exceeds a preset risk threshold, measuring the quantum role superposition state to cause the quantum role superposition state to collapse into a determined role.

[0015] By adopting the above technical solutions, roles can be quickly determined when there is a high risk of conflict, collisions caused by role uncertainty can be avoided, role exploration and collision avoidance can be balanced, and the real-time and safety of path planning can be improved.

[0016] Furthermore, the conflict risk value is obtained by calculating a conflict risk field, which comprehensively considers the spatial distance between agents and the target distribution entropy, and the target distribution entropy is used to characterize the uncertainty of the potential targets of the agents.

[0017] By adopting the above technical solutions, the conflict risk can be comprehensively assessed, taking into account both the spatial distance factor and the risk brought by target uncertainty, making the calculation of the conflict risk value more comprehensive and accurate, and providing a reliable basis for role collapse decision-making.

[0018] Furthermore, it also includes: constructing a quantum entangled cooperative state of adjacent intelligent agents, wherein the quantum entangled cooperative state includes a joint probability amplitude of a complementary role pair, and the complementary role pair is a role combination that can achieve collaborative work.

[0019] By adopting the above technical solutions, the collaborative work of adjacent intelligent agents is promoted, conflicts are avoided at the role level, the collaboration efficiency between intelligent agents is improved, and the coordination cost in path planning is reduced.

[0020] Furthermore, the construction of the causal mental map also includes: dynamically updating the causal mental map through a federated learning framework, the federated learning framework includes edge nodes and cloud nodes, the edge nodes are responsible for the calculation of local causal relationships, and the cloud nodes are responsible for the aggregation of global causal relationships.

[0021] By adopting the above technical solutions, efficient dynamic updating of the causal mental map can be achieved, taking into account the timeliness of local calculations and the comprehensiveness of global aggregation, thereby improving the adaptability of the causal mental map to environmental changes.

[0022] Furthermore, generating a collision-free global path includes: performing a path search based on an A* algorithm, wherein the heuristic function of the A* algorithm combines the path length and the integral value of the conflict risk field, and the integral value of the conflict risk field is the integral of the conflict risk value within a preset time interval.

[0023] By adopting the above technical solution, path search not only considers the path length, but also fully incorporates the conflict risk factor. The generated path can more effectively avoid collisions while ensuring efficiency, thereby improving the safety and rationality of the path.

[0024] In summary, this application has at least the following beneficial effects:

[0025] 1. A collision-free path planning method for mobile robots that combines causal cognition with quantum parallel decision-making improves planning efficiency and safety.

[0026] 2. Enhanced collaboration and intention understanding between agents through precise causal reasoning and efficient role assignment;

[0027] 3. Multi-dimensional optimization of path generation further improves path reliability and adaptability.

[0028] It should be understood that the contents described in the Summary of the Invention are not intended to limit the key or important features of the embodiments of the present application, nor are they intended to limit the scope of the present application. Other features of the present application will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0030] Figure 1 A schematic diagram of an exemplary operating environment in which embodiments of the present application can be implemented is shown.

[0031] Figure 2 A flowchart of a collision-free global path planning method for a mobile robot in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0032] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0033] In this document, the term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document generally indicates that the related objects are in an "or" relationship.

[0034] This application provides a collision-free global path planning method for mobile robots, which effectively improves the path planning efficiency and collision-free performance of robot collaboration in complex scenarios by integrating causal cognition and quantum parallel decision-making technology.

[0035] The embodiments of the present application disclose a collision-free global path planning method for a mobile robot.

[0036] Figure 1 A schematic diagram of an exemplary operating environment in which embodiments of the present application can be implemented is shown.

[0037] Reference Figure 1 The operating environment includes mobile robot clusters, environmental perception networks, computing and processing systems, communication transmission networks, and physical environment bases. Each part forms a closed-loop support system through data interaction links.

[0038] A mobile robot cluster consists of multiple robot bodies with autonomous movement capabilities. Each robot is equipped with a drive chassis (for executing the planned path) and a central control unit (for running the core logic of the path planning algorithm) as the execution carrier of the path planning.

[0039] The environmental perception network consists of a lidar, an event camera, and an inertial measurement unit (IMU). The lidar and event camera are deployed on the top of the robot, used to collect high-density environmental point clouds and dynamic event data, respectively. The IMU is integrated into the robot chassis to obtain motion state parameters. The three are connected to the central control unit through an internal data bus to provide environmental and self-state inputs for path planning.

[0040] The computing and processing system adopts an edge-cloud collaborative architecture. The edge computing nodes (deployed at key locations in the physical environment) are connected to the robot control unit through wired links, responsible for parallel processing of causal reasoning and quantum role state calculations. The cloud server interacts with the edge nodes through the communication network for aggregation and updating of the global causal relationship library.

[0041] The communication transmission network includes 5G-Advanced wireless communication modules (integrated in robots and edge nodes) and a fiber-optic synchronization network. The former is used to transmit real-time trajectory data and character state information, and the latter achieves nanosecond-level time synchronization across devices to ensure the accuracy of the timing of causal events.

[0042] The physical environment must conform to predefined spatial partitioning (e.g., a 50m x 50m grid). A UWB positioning base station must be deployed to provide a global coordinate reference. The density of dynamic entities within the environment must not exceed 50 per square meter, and static obstacles must possess physical outlines recognizable by LiDAR. This provides a spatial foundation for modeling the path planning algorithm. These components interact through data to form a complete "perception-computation-decision-execution" chain, supporting the implementation of collision-free global path planning for mobile robots.

[0043] Figure 2 The flowchart of a collision-free global path planning method for a mobile robot in an embodiment of the present application is shown. Figure 1 Execute in the runtime environment.

[0044] Reference Figure 2 , the method specifically comprises the following steps:

[0045] S1: Construct a causal mind map to realize intention reasoning between intelligent agents. The causal mind map includes the causal relationship between intelligent agents, the target belief set, the behavior parameter set and the collaborative credit score.

[0046] In the method of this step, constructing a causal mental map includes: extracting causal event pairs from the historical trajectory data of the intelligent agent, wherein the causal event pairs include cause events and result events; performing statistical significance verification on the causal event pairs to screen out valid causal relationships that meet a preset significance threshold; and updating the causal mental map based on the valid causal relationships.

[0047] The extraction of causal event pairs from the historical trajectory data of the intelligent agent includes: ensuring that the timestamp of the cause event is earlier than the timestamp of the result event through a time synchronization mechanism, and the synchronization accuracy of the time synchronization mechanism is not less than a preset synchronization time threshold (such as 50ns, based on the IEEE1588PTPv2 protocol). Specifically, the causal event pair is defined as a triple ,in Cause event (such as "battery level ”), is a result event (such as "turn to charging area"), The time difference between two events , must meet , too short may be due to measurement error, too long may be due to other intervening factors). By fusing the dynamic change sequence captured by the event camera with the position data of the lidar, the and The spatial coordinates and timestamps ensure Calculation error .

[0048] When testing the statistical significance of the causal event pair, we use the hypothesis testing framework: the null hypothesis for" and No causal relationship", alternative hypothesis for" lead to By calculating the causal effect value ,in Indicates the occurrence of active forced event E (excluding interference from other confounding factors), Indicates that E is forced not to occur, is the causal effect value, which quantifies the difference in the probability of B occurring when "intervention E occurs" and "intervention E does not occur" (reflecting the causal influence of E on B); And the corresponding value When, refuse , determined to be a valid causal relationship. Preset significance threshold Usually 0.4 is used (calibrated based on 100,000 sets of historical data to ensure False positive rate at confidence level ).

[0049] Target belief set The construction of the system needs to combine the counterfactual reasoning results, and its core is the computational agent. right The probability distribution of potential targets ,Right now For intelligent agents For intelligent agents The potential target probability (i.e. by is the probability of the actual target). For each candidate target (The first candidate targets (such as target points / areas in a map)), which are trained by a variational autoencoder (VAE) Historical trajectory Encoding to obtain the latent vector , the decoder outputs the initial target probability ; Combined with the counterfactual correction term, the final target probability is:

[0050]

[0051] in, The determination method includes statistical baseline probability, which is calculated by variational autoencoder (VAE): using the VAE encoder to Historical trajectory Encode and get the latent vector (key features of the compressed trajectory), output by the VAE decoder (measure and ), and then through Convert to probability; ICE To quantify the “forced by For the target's trajectory The degree of impact (reflecting the goal causal relationship with the trajectory), is the magnification factor, let , used to balance the statistical correlation ( historical associations reflected in the ICE) and causal determinism (intervention effects reflected in the ICE); For traversing all candidate target dummy variables (with the numerator belong to the same target set, used to normalize the probability).

[0052] Behavior Parameter Set Contains agents Motion characteristic parameters, such as velocity distribution , turning curvature , these parameters are extracted from historical trajectories by maximum likelihood estimation and updated every 30 minutes based on new data. The update formula is is the smoothing coefficient, is the mean velocity of the new trajectory), ensuring that the parameters are dynamically adjusted as the behavior pattern changes.

[0053] Collaborative Credit Scoring The initial value is set to 0.5 and is dynamically updated with the interaction results: The actual behavior and When the predictions are consistent, ; When unforeseen conflicts or collaboration failures occur, This score directly affects the model aggregation weight in subsequent federated learning. The local causal relationships of agents with high credit have a higher proportion in the global aggregation.

[0054] In this step, the method of constructing the causal mental map also includes: dynamically updating the causal mental map through a federated learning framework, wherein the federated learning framework includes edge nodes and cloud nodes, wherein the edge nodes are responsible for calculating local causal relationships and the cloud nodes are responsible for aggregating global causal relationships. Specifically, each edge node (corresponding to a 50m×50m spatial partition) maintains a local causal relationship library. , contains the valid causal pairs in the partition in the past 30 minutes. The edge node calculates the local model parameters every 5 minutes (like The probability distribution parameters of The cloud nodes use the weighted average method to aggregate global parameters: , where the weight ( is the average collaborative credit of the agents in the partition), ensuring that the local knowledge of high-credit areas accounts for a higher proportion in the global model. It will be synchronized back to each edge node to update the local causal mental map. The communication volume of the whole process times (achieved through model parameter compression), with a delay of less than 1s.

[0055] In the method of this step, the intention reasoning includes: generating a counterfactual trajectory of the agent based on counterfactual reasoning, where the counterfactual trajectory is the possible predicted trajectory of the agent assuming that a specific cause event has not occurred; combining the counterfactual trajectory with the actual trajectory of the agent to calculate the target probability distribution, where the target probability distribution is used to represent the potential goals of the agent and the corresponding confidence levels.

[0056] Generating counterfactual trajectories When the event is "uncaused As a premise, the modified A* algorithm is used to solve:

[0057]

[0058] in, For intelligent agents The counterfactual trajectory (simulating "when the uncaused event E occurs, should travel”, used for causal intervention analysis), For the mathematical operation "search for the candidate trajectory P that minimizes the sum (the core is to find the "optimal counterfactual trajectory")", is the position of the candidate trajectory P at time step t (traverse all possible trajectories and verify whether they meet the constraints), For intelligent agents The actual trajectory of the position at time t (historical observation data, as a benchmark for "close to reality"), is the Euclidean distance (calculating the position deviation between the candidate trajectory and the actual trajectory to ensure that the counterfactual trajectory “does not deviate from the historical trend”), is the preset counterfactual loss function (when With "no The larger the deviation from the expected path, the higher the CF value). The preset counterfactual weights (to ensure that the trajectory is close to the historical trend and conforms to the “no constraints).

[0059] Combining counterfactual trajectories with actual trajectories When calculating the target probability distribution, the initial target probability is corrected by comparing the deviations between the two at key decision points (such as forks in the road). For example, if the counterfactual trajectory shows Originally heading towards , but the actual trajectory is Steering ,but Confidence improvement ( is the correction factor, is the maximum possible deviation used to normalize d, so that ), and finally form a probability distribution containing 5-8 candidate targets, and satisfy ,in, For intelligent agents right Goal The confidence improvement (due to E causing trajectory deviation, the target probability needs to be corrected, is the overall deviation between the actual trajectory and the counterfactual trajectory (the sum of deviations at key decision points such as forks in the road, reflecting the influence of E), For intelligent agents right by is the probability of being the target (after correction, it must satisfy the legal distribution of "total confidence is 1").

[0060] Through the above process, the causal mind map can dynamically integrate causal relationships, goal beliefs, behavioral characteristics and collaboration history, provide a quantitative basis for intention reasoning between intelligent agents, and significantly improve the accuracy of intention prediction in high-density dynamic scenarios.

[0061] S2: Assigning roles to the agent based on a quantum role superposition state, where the quantum role superposition state includes probability amplitudes of multiple roles.

[0062] The probability amplitude of each role in the quantum role superposition state is determined by calculating the probability amplitude of each role based on the marginal contribution value, role fitness and causal effect of the intelligent agent. The marginal contribution value is calculated by Shapley value, and the role fitness is used to characterize the degree of matching between the behavioral parameters of the intelligent agent and the role.

[0063] Specifically, the quantum role superposition state is the parallel representation of multiple potential roles by the intelligent agent during the decision-making process. Its mathematical form is:

[0064]

[0065] in, is the quantum state vector (right vector), describing the quantum state of the system, For a preset character collection (usually containing Leader, Follower, Coordinator, Obstacle Avoider, Standby core roles, which can be expanded according to the scenario), For the role The ground state, is the probability amplitude of the corresponding role, satisfying the normalization condition (Make sure the total probability is 1). Probability amplitude The size directly reflects the role The calculation of the possibility of being selected requires the integration of three factors: marginal contribution value, role suitability and causal effect, to form a multi-dimensional decision-making basis.

[0066] Calculation of Marginal Contribution Value (Shapley Value): Agent The marginal contribution value is calculated by Shapley value Quantification, at its core, is evaluation The average value increment brought by all possible agent subset collaborations is:

[0067]

[0068] in, For intelligent agents Shapley value (marginal contribution quantitative indicator), is the set of agents in the current task, for Not included A subset of For subset The number of agents, for The total number of agents in ), Characterization After joining, the value of the subset increases, For subset The value of completing the task (defined as the inverse of the total length of the paths planned by all agents in the subset, i.e. , the shorter the path, the higher the value). Shapley value The larger the The stronger the irreplaceability in collaboration, the higher the probability weight of it assuming a key role (such as a navigator).

[0069] Quantification of role fit: role fit To measure the agent Behavior parameter set With the role The degree of demand matching is calculated using a weighted scoring method:

[0070]

[0071] in, Behavioral parameter dimensions (such as speed stability, steering flexibility, collaborative response speed, etc.), For the role For the first The weight of each parameter (such as the weight of "navigator" on speed stability) , weight of steering flexibility ), for No. Behavioral parameter values, For the role For the first The optimal value of the parameters, is the normalization function (mapping parameter values ​​to The closer to the optimal value, the higher the score). For example, if the character As the "navigator", its speed stability is the best value (variance ), Speed ​​stability ,but , indicating a high degree of fit.

[0072] Integration of causal effects: causal effects Quantitative Role With the agent The causal relationship strength of the current task goal is derived from the effective causal relationship in the causal mind map:

[0073]

[0074] in, For Assume a role A set of related causal event pairs (e.g., the causal pair of "assuming the role of navigator" and "shortening the task completion time"), is the causal effect value of the causal pair (defined as in step S1) ). The bigger, the more character The stronger the causal relationship with the current task goal, the more its probability amplitude needs to be enhanced.

[0075] Final calculation of probability amplitude: Combining the above three factors, the role The probability amplitude The calculation formula is:

[0076]

[0077] In order to meet the normalization conditions, the probability amplitudes of all roles need to be normalized:

[0078]

[0079] Through this formula, roles with high marginal contribution, good adaptability and strong causal relationship will obtain higher probability amplitude, so that the quantum role superposition state can accurately reflect the collaborative potential and task requirements of the intelligent body, and provide a reliable initial state for subsequent role collapse (when the conflict risk exceeds the threshold).

[0080] In addition, the parallelism of quantum role superposition is reflected in the In terms of synchronous representation, the intelligent agent does not need to evaluate the roles one by one, but explores all possibilities simultaneously through superposition states. Its computing efficiency increases linearly with the number of roles (thanks to the parallel processing capabilities of the quantum simulation acceleration unit), solving the efficiency bottleneck of traditional serial role trial and error in high-density scenarios.

[0081] S3: Combining the result of the intention reasoning with the result of the role assignment to generate a collision-free global path.

[0082] The method of this step specifically includes: performing a path search based on the A* algorithm, wherein the heuristic function of the A* algorithm combines the path length and the integral value of the conflict risk field, and the integral value of the conflict risk field is the integral of the conflict risk value within a preset time interval.

[0083] The collision risk field is used to quantify the collision risk between agents. Its calculation needs to comprehensively consider the spatial distance and intention uncertainty. The formula is:

[0084]

[0085] in, For intelligent agents and At a point in time and space The intensity of the conflict risk field, is the spatial coordinate, is the timestamp; For coordinates With the agent The predicted path In time The Euclidean distance of is the preset weight coefficient (spatial distance has a greater impact on immediate risk); is the pre-acquired spatial attenuation coefficient (the value is 1.5 times the robot radius, for example, when the radius is 30 cm for right The target distribution entropy ( , the higher the entropy value, the The greater the target uncertainty; is the smoothing factor, avoiding The denominator is 0.

[0086] The integral value of the conflict risk field is the accumulation of risk values ​​within a preset time interval, which is used to evaluate the conflict risk of the entire path. The calculation formula is:

[0087] Integral field of conflict risk field =

[0088] in, For the agent to reach the current node time, Estimated destination Time (by node arrive The straight-line distance is divided by the average speed, that is, for historical average speed); Nodes on the path In time The larger the integral value is, the closer the path is to the preset time interval. The higher the cumulative risk of a collision.

[0089] Improved heuristic function of A* algorithm The path length and conflict risk integral are combined into the formula:

[0090]

[0091] in, The value of the agent The heuristic function value at node n, original is the traditional A* Euclidean distance heuristic function original is the preset risk weight coefficient (calibrated through 1000 sets of simulation experiments, so that the proportion of risk factors in the total cost is approximately , balancing path efficiency and safety); represents the sum of the risk scores of all other agents to ensure global collision-free.

[0092] The total cost function of path search is:

[0093]

[0094] in, From the starting point To Node The actual path length; RoleCost Quantify the nodes for role adaptation cost With the agent Current role The matching degree of the role (for example, the cost of the "navigator" role is lower in the main channel node, and the cost of the "obstacle avoider" role is lower in the edge area node), is calculated as RoleCost For nodes functional attributes of the area); is the role weight coefficient, which ensures the impact of role assignment results on path preference.

[0095] During the path generation process, it is necessary to dynamically adjust the intention reasoning results in real time. :when Target belief set When updating (such as from 0.3 to 0.8), and recalculate its predicted path , and update the conflict risk field simultaneously and heuristic function If the risk score of a candidate path exceeds the preset threshold (like ,correspond Collision probability under confidence level ), then prune the path directly to avoid invalid search.

[0096] Through the above method, the generated path can not only ensure that the intelligent agent reaches the target efficiently, but also avoid potential conflicts in advance based on intention reasoning and role division, and achieve global collision-free collaborative movement.

[0097] The method further includes: when the conflict risk value between the intelligent agents exceeds a preset risk threshold, measuring the quantum role superposition state to cause the quantum role superposition state to collapse into a determined role.

[0098] In the method of this step, the conflict risk value is obtained by calculating the conflict risk field, which comprehensively considers the spatial distance between intelligent agents and the target distribution entropy. The target distribution entropy is used to characterize the uncertainty of the potential target of the intelligent agent.

[0099] Specifically, the calculation of the conflict risk value is based on the conflict risk field, and the maximum risk value of the agent within a preset time window (such as 5 seconds) in the future is selected as the judgment basis. The formula is:

[0100]

[0101] in, is the current time, is the risk prediction window (based on the average speed of the agent setting, covering a 5m range of motion). For intelligent agents In time The predicted location of for and The conflict risk field value at this time and space point (calculated in the same way as step S3) .when To preset the risk threshold, calibrate through 100,000 sets of collision cases to ensure When the potential collision energy is triggered), the measurement collapse of the quantum role superposition state is triggered.

[0102] Target distribution entropy It plays a key role in the calculation of conflict risk value, and its formula is:

[0103]

[0104] in is the number of candidate targets, For intelligent agents right by is the confidence of the target (from the target belief set in step S1). The entropy range is ,when When (indicates The target is highly uncertain), the target uncertainty term in the conflict risk field Will increase, More likely to exceed threshold , trigger collapse in advance to deal with high uncertainty scenarios.

[0105] Quantum role superposition The essence of the measurement process is to calculate the square of the probability amplitude of each character , select the character with the highest probability as the collapse result:

[0106]

[0107] in To ensure the coordination between roles, the collapse process needs to refer to the role status of adjacent intelligent agents, especially through quantum entanglement cooperation state. Implemented role association:

[0108]

[0109] in is a set of complementary role pairs (e.g., {(leader, follower), (coordinator, executor)}), is the joint probability amplitude. Collapse into a character hour, will preferentially collapse into Complementary roles ,Right now and , avoiding conflicts at the role allocation level (e.g. two agents will not collapse into "navigators" at the same time).

[0110] Character after collapse It will be incorporated into subsequent path planning as a constraint condition, by adjusting the role adaptation cost Role (See step S3) Dynamically update path preferences: For example, an agent collapsing into the "avoider" role will prioritize edge nodes (with lower RoleCost values) during path search, while a "navigator" will prioritize nodes along the main path. This role-path linkage ensures that post-collapse decisions directly contribute to the generation of collision-free paths, forming a closed loop of "collision detection - role determination - path adjustment."

[0111] Through this approach, the collapse of quantum role superpositions retains the efficiency advantages of parallel exploration while enabling rapid convergence to a defined role in high-risk scenarios, balancing planning flexibility and security. Furthermore, by combining multi-factor assessment of the conflict risk field with the collaborative constraints of quantum entanglement, the role collapse results are more consistent with the goal of global collision-free operation, significantly reducing the probability of collisions in high-density scenarios.

[0112] The method further includes: constructing a quantum entangled cooperative state of adjacent intelligent agents, wherein the quantum entangled cooperative state includes a joint probability amplitude of a complementary role pair, and the complementary role pair is a role combination that can achieve collaborative work.

[0113] The method further includes: constructing a quantum entangled cooperative state of adjacent intelligent agents, wherein the quantum entangled cooperative state includes a joint probability amplitude of a complementary role pair, and the complementary role pair is a role combination that can achieve collaborative work.

[0114] The determination of the complementary role pair is based on the compatibility of the role functions and the collaboration history data, specifically: from the preset role set Screening meets Character combination , forming a set of complementary role pairs .in Score the role compatibility using the formula: Functional compatibility (e.g., the functional complementarity score of "leader" and "follower" is 0.9, and that of "dual leader" is 0.2), is the historical collaboration success rate (calculated based on the proportion of conflict-free collaborations in the past 100 collaborations), is the function weight, Compatibility threshold (ensuring that the selected roles are effective for collaboration) ).

[0115] The basis for determining adjacent agents is the intersection of spatial distance and communication range: when an agent and Euclidean distance ( , based on the effective transmission distance setting of the communication module), and collaborative credit scoring (Collaborative credit score from step S1, ensuring the reliability of historical collaboration), it is determined to be an adjacent intelligent agent, triggering the construction of the quantum entangled collaborative state.

[0116] The mathematical form of quantum entangled cooperative state is multi-particle superposition state:

[0117]

[0118] in is the joint probability amplitude, satisfying the normalization condition The calculation of the joint probability amplitude combines role fitness and collaboration credit:

[0119]

[0120] for For the role The fitness of (from step S2), for and The collaborative credit score (from step S1) is used to make the complementary role pairs with high adaptability and good collaborative credit obtain a higher joint probability amplitude.

[0121] The dynamic update mechanism of quantum entangled cooperative state is: (Based on the agent's movement speed and the frequency of environmental changes), recalculate the distance between adjacent agents Compatibility score ,like or , then the entangled state is released; otherwise, the joint probability amplitude is updated according to the latest role fitness and collaborative credit , ensuring that the entangled state always reflects the current collaboration needs.

[0122] In the process of role collapse, the quantum entangled cooperative state realizes the role coordination of adjacent intelligent agents through the constraint of joint probability amplitude: The quantum role superposition state collapses to hour, The collapse result Need to meet , and the collapse probability is Positive correlation, that is:

[0123]

[0124] in For Complementary role sets. This mechanism ensures that the roles of adjacent agents always form an effective collaborative combination, fundamentally avoiding path collisions caused by role conflicts.

[0125] Based on the above, the method in this step constructs a causal mind map containing causal relationships, a target belief set, a behavioral parameter set, and a collaborative credit score. This allows accurate distinction between causal and random associations in agent behavior, and combines counterfactual reasoning to modify the target probability distribution, thereby reducing the risk of path conflicts caused by misjudgment of intention. Multiple roles are represented in parallel through quantum role superposition states, and marginal contribution values, role fitness, and causal effects are integrated to calculate probability amplitudes. This ensures that role assignments are both consistent with the agent's collaborative potential and adapted to task requirements, avoiding the efficiency loss of traditional serial trial and error. The collision risk is quantified by integrating spatial distance and target uncertainty through a conflict risk field. The A* algorithm heuristic function is optimized based on risk integrals within a preset time interval, enabling path search to balance efficiency while preemptively avoiding cumulative risks. When the conflict risk exceeds a threshold, the roles are locked and determined through quantum role superposition state collapse. Combined with quantum entangled collaborative states to constrain the complementary roles of adjacent agents, collaborative compatibility is ensured at the role level, fundamentally reducing path crossings caused by role conflicts. The above technical approach forms a closed loop from intention understanding, role assignment, risk assessment, to conflict resolution, ultimately achieving collision-free and efficient global path planning for agents in a dynamic environment.

[0126] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to the embodiments of this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required for this application.

[0127] In summary, this application has at least the following beneficial effects:

[0128] 1. Through causal mind mapping and counterfactual reasoning, we can accurately distinguish between causal and random associations in the behavior of intelligent agents, improve the accuracy of intention prediction, and fundamentally reduce path conflicts caused by misjudgment of intentions.

[0129] 2. Leveraging the parallelism of quantum role superposition states and the complementary constraints of entangled collaborative states, role allocation and coordination can be efficiently completed in high-density scenarios, breaking through the efficiency bottleneck of traditional serial role trial and error and enhancing the flexibility of multi-agent collaboration.

[0130] 3. The improved A* algorithm integrates the spatiotemporal integral of the conflict risk field, so that the generated path can avoid the cumulative collision risk within the preset time interval in advance while ensuring arrival efficiency, achieving a balance between safety and efficiency.

[0131] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the aforementioned disclosed concepts. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A collision-free global path planning method for a mobile robot, characterized in that: include: Constructing a causal mind map to realize intention reasoning between agents. The causal mind map includes the causal relationship between agents, the target belief set, the behavior parameter set, and the collaborative credit score; Assigning roles to agents based on a quantum role superposition state, wherein the quantum role superposition state includes probability amplitudes for multiple roles; Combining the intention reasoning result with the role assignment result to generate a collision-free global path; The construction of the causal mind map includes: Extracting causal event pairs from the agent's historical trajectory data, wherein the causal event pairs include a cause event and a result event; Performing statistical significance verification on the causal event pairs to screen out valid causal relationships that meet a preset significance threshold; updating the causal mental map based on the effective causal relationship; The probability amplitude of each role in the quantum role superposition state is determined by the following method: The probability amplitude of each role is calculated based on the marginal contribution value, role fitness and causal effect of the agent. The marginal contribution value is calculated by Shapley value. The role fitness is used to characterize the degree of matching between the behavior parameters of the agent and the role.

2. The collision-free global path planning method for a mobile robot according to claim 1, characterized in that: Extracting causal event pairs from the historical trajectory data of the intelligent agent includes: The time synchronization mechanism is used to ensure that the timestamp of the cause event is earlier than the timestamp of the result event, and the synchronization accuracy of the time synchronization mechanism is not lower than a preset synchronization time threshold.

3. The collision-free global path planning method for a mobile robot according to claim 1, characterized in that: The intent reasoning includes: Generate a counterfactual trajectory of the agent based on counterfactual reasoning, where the counterfactual trajectory is the predicted trajectory of the agent if a specific causal event does not occur; The target probability distribution is calculated by combining the counterfactual trajectory with the actual trajectory of the agent. The target probability distribution is used to represent the potential target of the agent and the corresponding confidence level.

4. The collision-free global path planning method for a mobile robot according to claim 1, characterized in that: Also includes: When the conflict risk value between the agents exceeds a preset risk threshold, the quantum role superposition state is measured to cause the quantum role superposition state to collapse into a determined role.

5. The collision-free global path planning method for a mobile robot according to claim 4, characterized in that: The conflict risk value is calculated through a conflict risk field, which comprehensively considers the spatial distance between agents and the target distribution entropy. The target distribution entropy is used to characterize the uncertainty of the potential targets of the agents.

6. The collision-free global path planning method for a mobile robot according to claim 1, characterized in that: Also includes: A quantum entangled cooperative state of adjacent intelligent agents is constructed, wherein the quantum entangled cooperative state includes the joint probability amplitude of complementary role pairs, and the complementary role pairs are role combinations that can achieve collaborative work.

7. The collision-free global path planning method for a mobile robot according to claim 1, characterized in that: The constructing of the causal mental map further includes: The causal mind map is dynamically updated through a federated learning framework, which includes edge nodes and cloud nodes. The edge nodes are responsible for calculating local causal relationships, and the cloud nodes are responsible for aggregating global causal relationships.

8. The collision-free global path planning method for a mobile robot according to claim 1, characterized in that: Generating a collision-free global path comprises: Path search is performed based on the A* algorithm, where the heuristic function of the A* algorithm combines the path length and the integral value of the conflict risk field, where the integral value of the conflict risk field is the integral of the conflict risk value within a preset time interval.

Images