Government affair digital human system based on agent framework and interaction method

Abstract

The invention relates to the technical field of intelligent service management, in particular to a government affair digital human system based on an intelligent agent framework and an interaction method. The system comprises a path planning module, an interaction module, a collision detection module and a response adjustment module. By sensing the position data of the static obstacle in real time, monitoring the motion state of the adjacent agents and introducing the relative speed and the path conflict probability, the future collision risk in the multi-agent scene is accurately predicted, that is, the collision risk in the motion process of the agent body and other agents is automatically evaluated in combination with the static obstacle, and the collision risk in the multi-agent scene is accurately predicted. And the patrol path and the movement speed are adjusted according to the obstacle environment, so that collision among multiple agents is avoided when the government affair service center patrols in the peak period, and the speed of the agents is increased when the collision risk is relatively low, so that the patrol efficiency is improved, the multi-dimensional risk is quantified, and the safety of the government affair service center is improved. And safety patrol and efficiency optimization of the government affair digital human in a dynamic environment are effectively realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent service management, and particularly to a government digital human system and an interaction method based on an agent framework. Background Art

[0002] With the advancement of digital transformation, the intelligent demand for government services is increasing day by day. The government digital human system based on the agent framework aims to improve the efficiency of government services and the user experience through artificial intelligence technology. How to combine human-machine collaborative decision-making, utilize deep learning and mixed reality technologies to achieve efficient government service interaction has increasingly become a topic of concern to relevant field personnel.

[0003] The Chinese patent document with the publication number CN114367976A discloses a lobby robot with detection and security functions, including: a body disposed in the service hall; the body includes: a touch screen operation main body for performing interactive operations; a storage box body detachably connected to the touch screen operation main body; a terminal server control system disposed in the body; the terminal server control system includes: a main control module for regulating the operation of the system; connected to the main control module: a visitor information collection module, a body temperature information detection module, a 3D tour management module, a remote visual interaction module; and a terminal database for storing information data of each module and system pre-configured data; It can be seen that the existing government digital human system lacks the quantification of multi-dimensional risks according to complex environmental information during the path planning process, resulting in a low degree of intelligence of the government digital human and thus a low patrol efficiency. Summary of the Invention

[0004] Therefore, the present invention provides a government digital human system and an interaction method based on an agent framework to overcome the problem in the prior art that there is a lack of quantification of multi-dimensional risks during the patrol of the digital human, timely adjustment of the path and movement speed, resulting in a low patrol efficiency.

[0005] To achieve the above object, the present invention provides a government digital human system based on an agent framework, including,

[0006] A path planning module, which is used to generate safety guidance points according to static obstacle information and construct an initial patrol path for each agent body according to the safety guidance points;

[0007] An interaction module, which is connected to the path planning module, and includes a number of agent bodies. Any one of the agent bodies is used to patrol along a set path to reach a target position, and perceive the surrounding environment to obtain the real-time position of static obstacles and the interaction state of adjacent agent bodies;

[0008] A collision detection module, which is connected to the interaction module, is used to monitor the patrol process of each agent body in real time, determine the risk of the body's movement and the risk of agent interaction according to the movement data of the agent body and the interaction state between the agent body and adjacent agents, so as to calculate the comprehensive collision risk level;

[0009] A response adjustment module, which is respectively connected to the collision detection module, the path planning module and the interaction module, is used to adjust the speed and patrol path of the agent body according to the comprehensive collision risk level of the agent body.

[0010] Further, the path planning module includes a key point extraction unit, a connectivity construction unit and a path generation unit, where,

[0011] The key point extraction unit is used to generate a number of candidate guiding points in the non-obstacle area, extract the edge contours corresponding to each static obstacle, generate a safety buffer corresponding to each static obstacle, and generate a number of candidate guiding points according to each safety buffer and the surrounding safety distance;

[0012] The connectivity construction unit is used to connect adjacent candidate guiding points to form a number of paths to be selected;

[0013] The path generation unit is used to select the optimal path from the paths to be selected according to the path length as the initial patrol path.

[0014] Further, the collision detection module includes an ontology movement risk monitoring unit, an agent interaction risk monitoring unit, a static obstacle risk monitoring unit and a comprehensive risk calculation unit, where,

[0015] The ontology movement risk monitoring unit is used to calculate the ontology movement risk according to the real-time speed, current step length and predicted turning angle of the agent body;

[0016] The agent interaction risk monitoring unit is used to calculate the agent interaction risk according to the Euclidean distance, relative speed magnitude and path conflict probability between the agent body and adjacent agents;

[0017] The static obstacle risk monitoring unit is used to calculate the static obstacle risk according to the distance to the nearest static obstacle, the circumradius of the static obstacle and the dynamic safety ratio coefficient;

[0018] The comprehensive risk calculation unit is used to calculate the comprehensive collision risk level of the agent body according to the ontology movement risk, agent interaction risk and static obstacle risk.

[0019] Further, the calculation formula for the ontology movement risk is,

[0020] ;

[0021] Among them, Rm is the risk of the body's movement;

[0022] kv is the speed risk weight, ks is the step length risk weight, and ko is the turning risk weight;

[0023] is the real-time speed of the agent's body;

[0024] is the maximum allowable speed;

[0025] is the current step length of the agent's body;

[0026] is the maximum step length of the agent's body, = 1.5 × the radius of the agent's body;

[0027] θ is the predicted turning angle;

[0028] θmax is the sharp turn threshold.

[0029] Furthermore, the calculation formula for the agent's interaction risk is,

[0030] ;

[0031] Among them, Ri is the agent's interaction risk;

[0032] σ is the type of motion state between the agent's body and the adjacent agent;

[0033] wd is the distance weight;

[0034] λ is the distance attenuation coefficient;

[0035] d is the Euclidean distance between the agent's body and the adjacent agent;

[0036] wv is the speed weight;

[0037] wc is the path conflict probability weight;

[0038] is the magnitude of the relative speed between the agent's body and the adjacent agent;

[0039] Vrel-max is the maximum allowable relative speed;

[0040] Pc is the path conflict probability between the agent's body and the adjacent agent.

[0041] Furthermore, the calculation formula for the path conflict probability is,

[0042] ;

[0043] Where Pc is the path conflict probability;

[0044] Tp is the total prediction time;

[0045] N is the number of prediction steps;

[0046] k is the current detection point;

[0047] is the collision indication function;

[0048] The safety radius is the sum of the radius of the agent, the radius of the adjacent agents, and the buffer distance;

[0049] Tk is the current prediction time corresponding to the current detection point;

[0050] is the time decay constant.

[0051] Furthermore, the calculation formula for static obstacle risk is:

[0052] ;

[0053] Among them, Rs is the static obstacle risk;

[0054] is the distance between the agent and the nearest static obstacle;

[0055] f is the dynamic safety proportional coefficient;

[0056] u is the safety margin;

[0057] ro is the radius of the circumscribed circle of the static obstacle.

[0058] Furthermore, the calculation formula for the comprehensive collision risk level is:

[0059] R = Rm × α + Ri × β + Rs × γ;

[0060] Among them, R is the comprehensive collision risk level;

[0061] Rm is the proprioceptive motion risk;

[0062] Ri is the agent interaction risk;

[0063] Rs is the static obstacle risk;

[0064] α is the weight coefficient of the body motion risk;

[0065] β is the weight coefficient of the agent interaction risk;

[0066] γ is the weight coefficient of the static obstacle risk.

[0067] Furthermore, the response adjustment module includes a risk determination unit, a speed control unit, and a path replanning unit, where

[0068] the risk determination unit is used to obtain the comprehensive collision risk level of the agent ontology in real time, and compare the first standard collision risk level and the second standard collision risk level with the comprehensive collision risk level;

[0069] the speed control unit is used to accelerate the agent ontology;

[0070] the path replanning unit is used to adjust the initial expansion distance of the static obstacle to the temporary expansion distance, so as to adjust the initial patrol path to the target patrol path.

[0071] On the other hand, the present invention also provides a government digital human interaction method based on the agent framework, which is applied to the above-mentioned government digital human system based on the agent framework, including

[0072] generating safety guiding points according to the static obstacle information, and constructing the initial patrol paths of each agent ontology according to the safety guiding points;

[0073] any one of the agents patrols along the initial patrol path, and senses the surrounding environment during the patrol, including the real-time position of the static obstacle and the interaction state of the adjacent agents, so as to reach the target position;

[0074] monitoring the patrol process of each agent in real time, obtaining any one of the agents as the agent ontology, and calculating the comprehensive collision risk level according to the motion data of the agent ontology and the interaction state between the agent and the adjacent agents;

[0075] adjusting the speed and patrol path of the agent ontology according to the comprehensive collision risk level of the agent ontology.

[0076] Compared with the prior art, the beneficial effects of the present invention are as follows: during the multi-agent patrol process, by real-time sensing the position data of the static obstacles and monitoring the motion states of the adjacent agents, introducing the relative speed and the path conflict probability, the future collision risk in the multi-agent scenario can be accurately predicted, that is, combining the static obstacles to automatically evaluate the collision risk during the movement of the agent ontology and other agents, and adjusting the patrol path and the motion speed according to the obstacle environment, so as to avoid collisions between multiple agents during the peak period of the government service center, and at the same time increase the speed of the agents when the collision risk is low, so as to improve the patrol efficiency, quantify the multi-dimensional risks, and effectively realize the safe patrol and efficiency optimization of the government digital human in the dynamic environment.

[0077] Further, a safety buffer zone is formed by expanding along the edge contour of the obstacle, and a certain distance is further extended outside the safety buffer zone to form a feasible area for guiding points, thereby improving path smoothness and obstacle avoidance redundancy. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 FIG. is a schematic structural diagram of a government digital human system based on an agent framework according to an embodiment of the present invention;

[0079] Figure 2 FIG. is a schematic structural diagram of a collision detection module according to an embodiment of the present invention;

[0080] Figure 3 FIG. is a schematic structural diagram of a response adjustment module according to an embodiment of the present invention;

[0081] Figure 4 FIG. is a schematic flowchart of a government digital human interaction method based on an agent framework according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0082] In order to make the objectives and advantages of the present invention clearer and more understandable, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0083] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0084] It should be noted that in the description of the present invention, the terms indicating directions or positional relationships such as "upper", "lower", "left", "right", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings. This is only for convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention.

[0085] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0086] Please refer to Figure 1As shown in the figure, it is a schematic structural diagram of the government affairs digital human system based on the agent framework in the embodiment of the present invention. The present invention provides a government affairs digital human system based on the agent framework.

[0087] A path planning module, which is used to generate safety guiding points according to static obstacle information and construct an initial patrol path for each agent ontology based on the safety guiding points;

[0088] An interaction module, which is connected to the path planning module and includes a number of agents. Any one of the agents patrols along the initial patrol path and can perceive the surrounding environment during the patrol, including the real-time position of static obstacles and the interaction state of adjacent agents, so as to reach the target position;

[0089] A collision detection module, which is connected to the interaction module and is used to monitor the patrol process of each agent in real time, obtain any agent as the agent ontology, and calculate the comprehensive collision risk level according to the motion data of the agent ontology and the interaction state between the agent and adjacent agents;

[0090] A response adjustment module, which is respectively connected to the collision detection module, the path planning module and the interaction module, and is used to adjust the speed and patrol path of the agent ontology according to the comprehensive collision risk level of the agent ontology.

[0091] In this embodiment, the government affairs digital human is a smart unmanned vehicle. The static obstacle information of each static obstacle in the current patrol scene is obtained through GPS or map matching. The current patrol scene is to patrol in the government affairs hall. The static obstacles include static facilities such as service desks and queuing railings. During the multi-agent patrol process, by real-time sensing the position data of static obstacles and monitoring the motion states of adjacent agents, the relative speed and path conflict probability are introduced to accurately predict the future collision risk in the multi-agent scenario, that is, the collision risk between the agent ontology and other agents during the movement is automatically evaluated in combination with static obstacles, and the patrol path and motion speed are adjusted according to the obstacle environment, so as to avoid collisions between multiple agents during the peak period of the government affairs service center patrol. At the same time, when the collision risk is relatively low, the speed of the agent is increased to improve the patrol efficiency, quantify the multi-dimensional risks, and effectively realize the safe patrol and efficiency optimization of the government affairs digital human in the dynamic environment.

[0092] Specifically, the path planning module includes a key point extraction unit, a connectivity construction unit and a path generation unit, where

[0093] The key point extraction unit is used to generate a number of candidate guiding points in the non-obstacle area, extract the edge contours corresponding to each static obstacle, generate a safety buffer zone corresponding to each static obstacle, and generate a number of candidate guiding points according to each safety buffer zone and the surrounding safety distance;

[0094] The connection construction unit is used to connect adjacent candidate guiding points to form a number of paths to be selected;

[0095] The path generation unit is used to select the optimal path from the paths to be selected according to the path length as the initial patrol path.

[0096] In this embodiment, the non-obstacle area is the passable area within the patrol area. The safety buffer zone represents the safety distance range to avoid collisions. The surrounding safety distance is the distance range outside the safety distance range. The safety buffer zone corresponding to each static obstacle is an extended range of 10 cm to 30 cm along the corresponding edge contour, and the set value is dynamically adjusted according to the type of static obstacle. For example, 30 cm is taken for the service desk and 10 cm is taken for the railing. The surrounding safety distance is set to 5 cm to 10 cm. The set value of the surrounding safety distance is related to the number of agents and the pedestrian flow density in the patrol area. The higher the pedestrian flow and the higher the number of agents, the larger the set value of the surrounding safety distance is set to ensure a larger buffer between the agent and pedestrians and obstacles. The lower the pedestrian flow and the smaller the number of agents, the smaller the set value of the surrounding safety distance is set to shorten the path length and improve the patrol efficiency; The path length is the total length of the path from the starting point to the ending point, represented by the Euclidean distance, and the shorter its value is, the better. That is, the path with the minimum path length in the paths to be selected is selected as the initial patrol path.

[0097] By expanding the safety buffer zone along the outer edge contour of the obstacle and further expanding a certain distance outside the safety buffer zone to form a feasible area for guiding points, the path smoothness and obstacle avoidance redundancy are improved. The double isolation of the safety buffer zone of the static obstacle and the dynamic surrounding safety distance reduces the risk of accidental touch.

[0098] Refer to Figure 2 as shown, which is a schematic structural diagram of the collision detection module in an embodiment of the present invention;

[0099] Specifically, the collision detection module includes an ontology motion risk monitoring unit, an agent interaction risk monitoring unit, a static obstacle risk monitoring unit, and a comprehensive risk calculation unit. Among them,

[0100] The ontology motion risk monitoring unit is used to calculate the ontology motion risk according to the real-time speed, current step length, and predicted turning angle of the agent ontology;

[0101] The intelligent agent interaction risk monitoring unit is used to calculate the intelligent agent interaction risk according to the Euclidean distance, relative speed magnitude, and path conflict probability between the intelligent agent body and adjacent intelligent agents;

[0102] The static obstacle risk monitoring unit is used to calculate the static obstacle risk according to the distance to the nearest static obstacle, the circumradius of the static obstacle, and the dynamic safety ratio coefficient;

[0103] The comprehensive risk calculation unit is used to calculate the comprehensive collision risk level of the intelligent agent body according to the body motion risk, intelligent agent interaction risk, and static obstacle risk.

[0104] Specifically, the calculation formula for the body motion risk is,

[0105] ;

[0106] where, Rm is the body motion risk;

[0107] kv is the speed risk weight, ks is the step length risk weight, and ko is the turning risk weight;

[0108] is the real-time speed of the intelligent agent body;

[0109] is the maximum allowable speed;

[0110] is the current step length of the intelligent agent body;

[0111] is the maximum step length of the intelligent agent body, = 1.5 × the radius of the intelligent agent body;

[0112] θ is the predicted turning angle, and the predicted turning angle is the steering angle of the intelligent agent body;

[0113] θmax is the sharp turning threshold, and θmax = 90°.

[0114] Specifically, the calculation formula for the intelligent agent interaction risk is,

[0115] ;

[0116] Ri is the intelligent agent interaction risk;

[0117] σ is the motion state type between the intelligent agent body and adjacent intelligent agents. When the intelligent agent body and adjacent intelligent agents are moving towards each other, σ = 1.2; when the intelligent agent body and adjacent intelligent agents are moving in the same direction, σ = 0.8; when the intelligent agent body and adjacent intelligent agents are in other states, σ = 1.0;

[0118] wd is the distance weight, wd = 0.4;

[0119] λ is the distance decay coefficient, λ = 1.5;

[0120] d is the Euclidean distance between the agent's body and the adjacent agent;

[0121] wv is the speed weight, wv = 0.4;

[0122] wc is the conflict probability weight, wc = 0.2;

[0123] is the magnitude of the relative speed between the agent's body and the adjacent agent;

[0124] Vrel-max is the maximum allowable relative speed;

[0125] Pc is the path conflict probability between the agent's body and the adjacent agent, and Pc is between 0 and 1.

[0126] Specifically, the calculation formula for the path conflict probability is

[0127] ;

[0128] where Pc is the path conflict probability;

[0129] Tp is the total prediction duration, Tp = 5s;

[0130] N is the number of prediction steps, N = Tp / ΔT, ΔT = 0.2s, N = 25, indicating that the trajectory for the next 5 seconds is discretely predicted 25 times (predicting the position every 0.2 seconds), generating a total of 25 prediction points;

[0131] k is the current detection point;

[0132] is the collision indication function. When the actual distance < safety radius, the collision indication function is 1, otherwise, the collision indication function is 0. The actual distance is , is the predicted position of the agent's body A at time t, is the predicted position of the adjacent agent of the agent's body A at time t;

[0133] The safety radius is the sum of the radius of the agent's body, the radius of the adjacent agent, and the buffer distance. The buffer distance = 0.2m;

[0134] Tk is the current prediction time corresponding to the current detection point, Tk = k×ΔT. When k = 1, T1 = 0.2s. When k = 25, T 25 = 5.0s;

[0135] τ is the time decay constant, τ = 2.0.

[0136] In this embodiment, the path conflict probability is predicted through the trajectory in the next 5 seconds to avoid potential collisions in advance, such as intelligent agents moving towards each other.

[0137] Specifically, the calculation formula for the static obstacle risk is

[0138] ;

[0139] Rs is the static obstacle risk;

[0140] is the distance between the agent's own body and the nearest static obstacle;

[0141] f is the dynamic safety ratio coefficient, f = 1.2;

[0142] u is the safety margin, u = 0.3m;

[0143] ro is the radius of the circumscribed circle of the static obstacle.

[0144] Specifically, the calculation formula for the comprehensive collision risk level is

[0145] R = Rm×α + Ri×β + Rs×γ;

[0146] R is the comprehensive collision risk level;

[0147] Rm is the risk of the body's movement;

[0148] Ri is the risk of agent interaction;

[0149] Rs is the static obstacle risk;

[0150] α is the weight coefficient of the risk of the body's movement, α = 0.5;

[0151] β is the weight coefficient of the risk of agent interaction, β = 0.3;

[0152] γ is the weight coefficient of the static obstacle risk, γ = 0.2.

[0153] Refer to Figure 3 shown, which is the structural schematic diagram of the response adjustment module of the embodiment of the present invention;

[0154] Specifically, the response adjustment module includes a risk determination unit, a speed control unit, and a path replanning unit, where

[0155] The risk determination unit is used to obtain the comprehensive collision risk level of the agent's own body in real time, and compare the first standard collision risk level and the second standard collision risk level with the comprehensive collision risk level;

[0156] The speed control unit is used to control the speed of the agent body so that the agent body accelerates;

[0157] The path replanning unit is used to adjust the initial expansion distance of the static obstacle to the temporary expansion distance to adjust the initial patrol path to the target patrol path.

[0158] In this embodiment, the first standard collision risk level is set to 0.3, and the second standard collision risk level is set to 0.7. The first standard collision risk level and the second standard collision risk level are compared with the comprehensive collision risk level:

[0159] If the comprehensive collision risk level is less than or equal to the first standard collision risk level, it is determined that the current state of the agent body is a low risk;

[0160] If the comprehensive collision risk level is between the first standard collision risk level and the second standard collision risk level, it is determined that the current state of the agent body is a medium risk;

[0161] If the comprehensive collision risk level is greater than or equal to the second standard collision risk level, it is determined that the current state of the agent body is a high risk;

[0162] The speed control unit increases the speed when the current state of the agent body is a low risk to shorten the task completion time, and automatically decelerates when the current state of the agent body is a high risk.

[0163] When the current state of the agent body is a medium risk, the path replanning unit increases the dynamic safety ratio coefficient to insert temporary guiding points, thereby realizing local fine-tuning of the path;

[0164] When the current state of the agent body is a high risk, the priority is negotiated through the communication protocol, so that the agent with a higher task urgency passes first, triggering global path replanning. That is, the high-risk agent A broadcasts a negotiation request containing the following information to the surrounding agents through a wireless communication network, such as 5G / WiFi: its own ID, current location, target location, current comprehensive collision risk level, recommended avoidance area (safety buffer calculated according to the predicted path), task urgency, and the task urgency is a preset value and can be updated in real time to meet the actual operating environment. For example, when there is a sudden fire alarm, the priority of the patrol vehicle is raised to the highest; and the neighboring agents respond. The neighboring agent B that receives the broadcast performs the following operations:

[0165] Calculate its own task urgency and compare it with the urgency of the high-risk agent A;

[0166] If its own urgency is lower, for example, the urgency of the neighboring agent B = 5 < the urgency of the high-risk agent A = 9, then actively avoid:

[0167] Send a "consent to avoid" response to the high-risk agent, along with its own position and the planned avoidance path;

[0168] Call its own response adjustment module to reduce speed and re-plan the local path, such as bypassing static obstacles.

[0169] If its own urgency is higher, broadcast a "refusal to avoid" response, asking the high-risk agent A to wait.

[0170] The global path is re-planned by adjusting the initial expansion distance of the static obstacle to the temporary expansion distance, and the temporary expansion distance is 1.2 times the initial expansion distance, generating a bypass path to adjust the initial patrol path to the target patrol path.

[0171] In this embodiment, group congestion is avoided through risk data sharing. For example, when the calculated comprehensive collision risk level is 0.8, the priority is negotiated through the communication protocol to allow the agent with a higher task urgency to pass first, and the initial expansion distance of the service desk is adjusted to the temporary expansion distance of 36 cm.

[0172] Refer to Figure 4 As shown, it is a schematic flowchart of the government affairs digital human interaction method based on the agent framework in the embodiment of the present invention;

[0173] The present invention also provides a government affairs digital human interaction method based on the agent framework, which is applied to the above-mentioned government affairs digital human system based on the agent framework,

[0174] Step S1, generate safety guidance points according to the static obstacle information, and construct the initial patrol path of each agent ontology according to the safety guidance points;

[0175] Step S2, any one of the agent ontologies patrols along the set path to reach the target position, and perceives the surrounding environment to obtain the real-time position of the static obstacle and the interaction state of the adjacent agent;

[0176] Step S3, monitor the patrol process of each agent ontology in real time, determine the ontology movement risk and the agent interaction risk according to the movement data of the agent ontology and the interaction state of the agent ontology with the adjacent agent, so as to calculate the comprehensive collision risk level;

[0177] Step S4, adjust the speed and patrol path of the agent ontology according to the comprehensive collision risk level of the agent ontology.

[0178] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, those skilled in the art can easily understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the protection scope of the present invention.

[0179] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A government digital human system based on an agent framework, characterized in that, including, a path planning module, which is used to generate safety guiding points according to static obstacle information and construct an initial patrol path for each agent ontology according to the safety guiding points; an interaction module, which is connected to the path planning module and includes several agent ontologies. Any one of the agent ontologies is used to patrol along a set path to reach a target position, and to sense the surrounding environment to obtain the real-time position of static obstacles and the interaction status of adjacent agents; a collision detection module, which is connected to the interaction module and is used to monitor the patrol process of each agent ontology in real time, determine the ontology movement risk and agent interaction risk according to the movement data of the agent ontology and the interaction status of the agent ontology with adjacent agents, so as to calculate the comprehensive collision risk level; a response adjustment module, which is respectively connected to the collision detection module, the path planning module and the interaction module, and is used to adjust the speed and patrol path of the agent ontology according to the comprehensive collision risk level of the agent ontology.

2. The government affairs digital human system based on the agent framework according to claim 1, characterized in that The path planning module includes a key point extraction unit, a connectivity construction unit and a path generation unit, where the key point extraction unit is used to generate several candidate guiding points in the non-obstacle area, extract the edge contours corresponding to each static obstacle, generate a safety buffer zone corresponding to each static obstacle, and generate several candidate guiding points according to each safety buffer zone and the surrounding safety distance; the connectivity construction unit is used to connect adjacent candidate guiding points to form several paths to be selected; the path generation unit is used to select the optimal path from the paths to be selected according to the path length as the initial patrol path.

3. The government affairs digital human system based on the agent framework according to claim 2, characterized in that, The collision detection module includes an ontology movement risk monitoring unit, an agent interaction risk monitoring unit, a static obstacle risk monitoring unit and a comprehensive risk calculation unit, where the ontology movement risk monitoring unit is used to calculate the ontology movement risk according to the real-time speed, current step length and predicted turning angle of the agent ontology; the agent interaction risk monitoring unit is used to calculate the agent interaction risk according to the Euclidean distance, relative speed magnitude and path conflict probability between the agent ontology and adjacent agents; the static obstacle risk monitoring unit is used to calculate the static obstacle risk according to the distance to the nearest static obstacle, the circumradius of the static obstacle and the dynamic safety ratio coefficient; the comprehensive risk calculation unit is used to calculate the comprehensive collision risk level of the agent ontology according to the ontology movement risk, agent interaction risk and static obstacle risk.

4. The government affairs digital human system based on the agent framework according to claim 3, characterized in that, The calculation formula of the ontology movement risk is ； where Rm is the ontology movement risk; kv is the speed risk weight, ks is the step length risk weight, and ko is the turning risk weight; is the real-time speed of the agent body; is the maximum allowable speed; is the current step size of the agent ontology; is the maximum step length of the agent body, = 1.5 × the radius of the agent body; θ is the predicted turning angle; θmax is the sharp turning threshold.

5. The government affairs digital human system based on the agent framework according to claim 3, characterized in that The calculation formula of the agent interaction risk is ； where Ri is the agent interaction risk; σ is the movement state type of the agent ontology and adjacent agents; wd is the distance weight; λ is the distance attenuation coefficient; d is the Euclidean distance between the agent ontology and adjacent agents; wv is the speed weight; wc is the conflict probability weight; is the magnitude of the relative velocity between the agent body and the adjacent agent; Vrel-max is the maximum allowable relative speed; Pc is the path conflict probability between the agent's body and adjacent agents.

6. The government affairs digital human system based on the agent framework according to claim 5, characterized in that, The calculation formula for the path conflict probability is ； where Pc is the path conflict probability; Tp is the total prediction duration; N is the number of prediction steps; k is the current detection point; is a collision indication function; The safety radius is the sum of the radius of the agent's body, the radius of adjacent agents, and the buffer distance; Tk is the current prediction moment corresponding to the current detection point; is the time decay constant.

7. The government affairs digital human system based on the agent framework according to claim 3, wherein, The calculation formula for the static obstacle risk is ； where Rs is the static obstacle risk; is the distance between the agent's body and the nearest static obstacle; f is the dynamic safety ratio coefficient; u is the safety margin; ro is the radius of the circumscribed circle of the static obstacle.

8. The government affairs digital human system based on the agent framework according to claim 7, characterized in that, The calculation formula for the comprehensive collision risk level is R = Rm×α + Ri×β + Rs×γ; where R is the comprehensive collision risk level; Rm is the risk of the agent's body movement; Ri is the risk of agent interaction; Rs is the static obstacle risk; α is the weight coefficient of the risk of the agent's body movement; β is the weight coefficient of the risk of agent interaction; γ is the weight coefficient of the static obstacle risk.

9. The government affairs digital human system based on the agent framework according to claim 1, characterized in that, The response adjustment module includes a risk determination unit, a speed control unit, and a path replanning unit, where the risk determination unit is used to obtain the comprehensive collision risk level of the agent's body in real time, and compare the first standard collision risk level and the second standard collision risk level with the comprehensive collision risk level; the speed control unit is used to accelerate the agent's body; the path replanning unit is used to adjust the initial expansion distance of the static obstacle to the temporary expansion distance to adjust the initial patrol path to the target patrol path.

10. A government affairs digital human interaction method based on an agent framework, which is applied to the government affairs digital human system based on an agent framework according to any one of claims 1-9, and is characterized in that generate safety guiding points according to static obstacle information, and construct the initial patrol paths of each agent's body according to the safety guiding points; any one of the agent's bodies patrols along a set path to reach the target position, and senses the surrounding environment to obtain the real-time position of static obstacles and the interaction state of adjacent agents; monitor the patrol process of each agent's body in real time, determine the risk of the agent's body movement and the risk of agent interaction according to the movement data of the agent's body and the interaction state of the agent's body and adjacent agents, so as to calculate the comprehensive collision risk level; adjust the speed and patrol path of the agent's body according to the comprehensive collision risk level of the agent's body.

Images