Intelligent agent control system and intelligent agent control method

Abstract

Provided are an agent control system and the like that enable an agent to perform a prescribed task while performing monitoring. An agent control system is provided with: a monitoring state evaluation unit (305) for calculating, on the basis of monitoring information transmitted from an agent moving within a prescribed area, a monitoring evaluation index indicating the quality of the monitoring state for each location within the prescribed area; a large-area route generation unit (306) that generates a route plan for the agent on the basis of the monitoring evaluation index and traffic management information including the movement destination of the agent and the type of task; and a route plan transmission unit (309) that transmits route plan data to the agent.

Description

Technical Field

[0001] The present invention relates to an agent control system, etc. Background Art

[0002] Automatic transportation systems for enabling intelligent agents (transportation vehicles such as cars) to transport goods or people from a certain location within a specified area to a destination are used in a variety of fields. In automatic transportation systems, in areas where controlled objects (e.g., autonomous vehicles) and uncontrolled objects (e.g., manually driven vehicles and people) are mixed, such as in urban areas, it is expected that intelligent agents perform path planning that reduces the risk of collision with uncontrolled objects.

[0003] Regarding such a technology, for example, Patent Document 1 states that "In a mobile body allocation unit, a potential risk object to be monitored is determined in such a manner that the exploration degree and the following degree each satisfy a predetermined level, and a plurality of mobile bodies are allocated to the potential risk object determined as a potential risk object to be monitored."

[0004] Prior art literature

[0005] Patent Literature

[0006] Patent Document 1: Japanese Patent Application Publication No. 2019-16306 Summary of the invention

[0007] However, the technology described in Patent Document 1 only considers monitoring within a specified area, and does not specifically consider the case of performing other tasks (e.g., transporting goods or people). In the technology described in Patent Document 1, if a specified task other than monitoring is also performed, a dedicated intelligent agent for the task is required, and there is a possibility that many intelligent agents are required.

[0008] Therefore, an object of the present invention is to provide an agent control system or the like that enables an agent to perform a predetermined task while performing monitoring.

[0009] In order to solve such a problem, the intelligent agent control system of the present invention comprises: a monitoring status evaluation unit, which calculates a monitoring evaluation index based on the monitoring information sent from the intelligent agent moving in the specified area, and the monitoring evaluation index represents the quality of the monitoring status of each place in the above-mentioned specified area; a path planning generation unit, which generates the path planning of the above-mentioned intelligent agent based on the business management information including the moving destination and task category of the above-mentioned intelligent agent, and the above-mentioned monitoring evaluation index; and a path planning communication unit, which sends the data of the above-mentioned path planning to the above-mentioned intelligent agent.

[0010] Effects of the Invention

[0011] According to the present invention, it is possible to provide an agent control system or the like that enables an agent to perform a prescribed task while performing monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a functional block diagram of the agent control system according to the first embodiment.

[0013] Figure 2 This is a functional block diagram showing the system configuration of an agent in the agent control system according to the first embodiment.

[0014] Figure 3 This is an explanatory diagram regarding the size and orientation of the agent in the agent control system according to the first embodiment.

[0015] Figure 4 This is a functional block diagram showing the system configuration of a base station in the agent control system according to the first embodiment.

[0016] Figure 5A This is a diagram showing an example of map information in the agent control system according to the first embodiment.

[0017] Figure 5B The first embodiment of the intelligent control system used in Figure 5A A partially enlarged view of area K1 in the map information.

[0018] Fig. 6A is the coefficient α used to calculate the monitoring evaluation index in the agent control system of the first embodiment. oj 's illustration.

[0019] Figure 6B is the coefficient α used to calculate the monitoring evaluation index in the agent control system of the first embodiment. mj 's illustration.

[0020] Fig. 7A It is an explanatory diagram showing the path of an agent in a comparative example.

[0021] Figure 7B It is an explanatory diagram showing a large-area path of an agent in the agent control system according to the first embodiment.

[0022] Figure 8 is a monitoring evaluation index C in the agent control system of the first embodiment. mj With weight W mj A diagram illustrating the relationship between .

[0023] Fig. 9A This is an explanatory diagram showing the traffic conditions of a section within a predetermined area at time t=k in the agent control system according to the first embodiment.

[0024] Fig. 9B This is an explanatory diagram showing the traffic conditions of a section within a predetermined area at time t=k+1 in the agent control system according to the first embodiment.

[0025] Fig.10 This is an explanatory diagram showing conditions for avoiding contact with an obstacle in the agent control system according to the first embodiment.

[0026] Fig.11 This is a flowchart showing the operation flow of the agent control unit included in the agent control system according to the first embodiment.

[0027] Fig.12 This is a functional block diagram of the agent control system according to the second embodiment.

[0028] Fig.13 This is a functional block diagram showing the system configuration of a base station in the agent control system according to the second embodiment.

[0029] Fig.14 It is an explanatory diagram showing an operation example of the monitoring rate calculation unit in the agent control system according to the second embodiment.

[0030] Fig.15 This is an explanatory diagram of a theme park to which the agent control system according to the third embodiment is applied.

[0031] Fig.16A It is an explanatory diagram showing the path of an agent in a comparative example.

[0032] Fig. 16B It is an explanatory diagram showing a large-area path of an agent in the agent control system according to the third embodiment.

[0033] Fig.17 It is an explanatory diagram showing a travel path of an agent in the agent control system according to the third embodiment. DETAILED DESCRIPTION

[0034] 《First Implementation Method》

[0035] Figure 1 This is a functional block diagram of the agent control system W1 according to the first embodiment.

[0036] The intelligent agent control system W1 is a system that calculates the path planning (time series data such as coordinates, posture, speed, etc.) of each intelligent agent 200-1 to 200-n (controllable movable bodies such as robots or vehicles) within a specified area (for example, a specific area or a theme park) and moves the intelligent agents 200-1 to 200-n based on the path planning.

[0037] In the first embodiment, as an example, a system in which an autonomous driving vehicle delivers people or materials within a specified area is described. In addition, the application target of the first embodiment is not limited to autonomous driving vehicles within a specified area, and it can also be applied to transportation vehicles within areas such as ports or theme parks as described later. In addition, the intelligent bodies 200-1 to 200-n are mobile bodies that perform autonomous driving based on instructions from the base station 100 (server). In addition, in the case where a specified intelligent body assumes the role of the base station 100, the base station 100 can also be omitted.

[0038] like Figure 1 As shown, the agent control system W1 is configured to include a base station 100 that calculates the paths of the agents 200 - 1 to 200 - n in a predetermined area, and the agents 200 - 1 to 200 - n that move along the paths calculated by the base station 100 .

[0039] <Base Station>

[0040] The base station 100 calculates the driving routes of the agents 200-1 to 200-n. The base station 100 includes a RAM 101 (Random Access Memory) as a volatile storage element, a ROM 102 (Read Only Memory) as a non-volatile storage element, and a CPU 103 (Central Processing Unit) including a processor. In addition to the above-mentioned structure, the base station 100 also includes a bus 104, an input / output interface 105, and a communication device 106. Furthermore, the CPU 103 reads out the program stored in the ROM 102 and expands it in the RAM 101, thereby executing various processes.

[0041] The bus 104 is a signal line for interconnecting the RAM 101, the ROM 102, the CPU 103, and the input / output interface 105. The input / output interface 105 is used to transmit data via the communication device 106. The communication device 106 performs wireless communication with the agents 200-1 to 200-n in a predetermined manner. The communication device 106 is connected to the bus 104 via the input / output interface 105.

[0042] also, Figure 1 , the number of base stations 100 is not limited thereto. For example, multiple base stations may function as one server. In addition, one or more of the agents 200-1 to 200-n may function as the base station 100.

[0043] <Agent 200>

[0044] Agents 200-1 to 200-n monitor a predetermined area and follow the route based on route information transmitted by wireless communication from base station 100. In the following, except for the case where agents 200-1 to 200-n are represented individually, agents 200-1 to 200-n are collectively referred to as agents 200. Figure 1 As shown, the agent 200 includes RAM 201, ROM 202, CPU 203, bus 204, and input / output interface 205. In addition to the above configuration, the agent 200 also includes a communication device 206, an external recognition sensor 207, a position measurement device 208, a posture measurement device 209, and a control device 210.

[0045] exist Figure 1 In the example of FIG. 2 , the RAM 201, the ROM 202, and the CPU 203 are connected to the input / output interface 205 via the bus 204. In addition, the communication device 206, the outside world recognition sensor 207, the position measurement device 208, the posture measurement device 209, and the control device 210 are connected to the input / output interface 205. The agent 200 having such a structure not only moves based on the path information transmitted from the base station 100 by wireless communication, but also monitors the driving environment using the outside world recognition sensor 207 and transmits the result to the base station 100.

[0046] The communication device 206 is a terminal capable of wireless communication such as Bluetooth, Wi-Fi, or a mobile phone line. The external recognition sensor 207 is a sensor for measuring the surrounding environment of the intelligent body 200. As such an external recognition sensor 207, for example, LiDAR (Light Detection And Ranging) and a camera are used.

[0047] The position measuring device 208 is a device for measuring the position of the intelligent body 200 on the map. In the processing of the position measuring device 208, for example, GNSS (Global Navigation Satellite System) is used. In addition, instead of the position measuring device 208, the position and orientation of the intelligent body 200 on the map can be calculated based on the SLAM (Simultaneous Localization and Mapping) technology using LiDAR and a camera. The posture measuring device 209 is a device for measuring the orientation and posture of the intelligent body 200. As such a posture measuring device 209, for example, an IMU (Inertia Measurement Unit) and an encoder are used. The control device 210 is a device that converts the speed command and orientation command of the intelligent body 200 into the actuator output of the intelligent body 200. As such a control device 210, a control microcomputer or the like is used.

[0048] <Description of Agent 200>

[0049] Figure 2 It is a functional block diagram showing the system structure of the intelligent agent 200.

[0050] like Figure 2 As shown, the CPU 203 of the agent 200 includes a state detection unit 211, a follow-up control unit 212, and a risk monitoring unit 213 as functional structures. The state detection unit 211 calculates the state (position, orientation) of the agent 200 based on the output value of the position measurement device 208 in addition to the output value of the posture measurement device 209.

[0051] Figure 3 This is an explanatory diagram related to the size and orientation of the intelligent body 200.

[0052] exist Figure 3 In the example, the agent 200 is configured as a vehicle and has front wheels 221, 221 and rear wheels 222, 222. In addition, the position and orientation of the agent 200 are assumed to be represented by a state p(t) = [x(t), y(t), θ(t)] (t is a time). Figure 3 The orientation θ(t) shown shows the orientation of the agent 200 with respect to a predetermined direction. is the angle representing the turning direction of the agent 200 based on the orientation θ(t). Figure 3 The length L shown is the distance between the front wheel 221 and the rear wheel 222 in the front-rear direction. The length L is used for following the control unit 212 and the like (see Figure 2 ) processing.

[0053] Figure 2 The following control unit 212 shown in FIG. Figure 1 ) The target path r(t) = [xr(t), yr(t), θr(t)] of the agent 200 obtained via the communication device 206 and the state p(t) = [x(t), y(t), θ(t)] of the agent 200 obtained from the state detection unit 211 are feedback controlled to reduce the difference between the target path r(t) and the state p(t). In addition, the following control unit 212 outputs control values ​​such as the steering amount and acceleration / deceleration (acceleration amount, braking amount) to the control device 210.

[0054] The control device 210 performs control based on the control value calculated by the following control unit 212 so that the tires of the agent 200 have a predetermined rotation speed and the steering has a predetermined rotation angle.

[0055] The risk monitoring unit 213 detects other vehicles (vehicles not belonging to the agent 200) and pedestrians around the agent 200 based on the sensor information obtained from the external recognition sensor 207, and sends the detection results together with the position information of the agent 200 to the base station 100 (see Figure 1 ).

[0056] <Description of Base Station 100>

[0057] Figure 4 It is a functional block diagram showing the system configuration of the base station 100 .

[0058] like Figure 4 As shown, the base station 100 includes an agent control unit 300 in addition to an agent information management unit 301, a monitoring information management unit 302, a map information management unit 303, and a service management unit 304. The agent control unit 300 has a function of calculating the agent 200 (see Figure 1 )’s moving path. Figure 4 As shown, the agent control unit 300 is connected to the agent information management unit 301, the monitoring information management unit 302, the map information management unit 303, and the service management unit 304. The processing of the agent control unit 300 is performed by the CPU 103 of the base station 100 (see Figure 1 ) is executed in

[0059] <Intelligent Body Information Management Unit 301>

[0060] The agent information management unit 301 manages the individual information (referred to as agent individual information) of the agent 200 traveling in the specified area. In such agent individual information, in addition to the size of the agent 200 and the external recognition sensor 207 (refer to Figure 2 ) in addition to the performance value, it also includes the cruising distance of the agent 200. In addition, the agent individual information managed by the agent information management unit 301 is not limited to this, and can also include, for example, the number of passengers that the agent 200 can take and the weight of cargo that can be loaded.

[0061] As the agent individual information, for example, information collected in advance and stored in the base station 100 (see Figure 1 ) in the ROM 102. Alternatively, the base station 100 may periodically acquire and update the agent individual information managed by a server building (not shown) outside the base station 100. The agent individual information is output from the agent information management unit 301 to the monitoring state evaluation unit 305 and is also output to the path correction unit 308.

[0062] <Monitoring Information Management Unit 302>

[0063] The monitoring information management unit 302 monitors the agents 200 in the specified area (see Figure 1 ) in addition to the position of the intelligent body 200, the position information of the vehicles and pedestrians other than the intelligent body 200 is also collected in a timely manner. Such monitoring information is obtained by, for example, transmitting the position measuring device 208 (see Figure 1 ) and the measured value of the external recognition sensor 207 (refer to Figure 1 ) is sent to the base station 100 via the communication device 206 for collection. The monitoring information collected in this way is output from the monitoring information management unit 302 to the monitoring state evaluation unit 305. In addition, the method of collecting monitoring information is not limited to this. For example, fixed sensors (not shown) may be installed at intersections in urban areas, and the information of pedestrians and vehicles around obtained from the fixed sensors may be sent to the base station 100. As such fixed sensors, for example, cameras and LiDAR are used.

[0064] <Map Information Management Unit 303>

[0065] The map information management unit 303 determines the predetermined area 500 (see Figure 5A The map information is output from the map information management unit 303 to the large area path generation unit 306.

[0066] Figure 5A This is an explanatory diagram showing an example of map information.

[0067] For example, Figure 5A The map information of the predetermined area 500 (the area indicated by dots) shown in FIG. 1 is stored in advance in the map information management unit 303 (see FIG. 1 ). Figure 4 ).

[0068] Figure 5B yes Figure 5A A partially enlarged view of area K1 in the map information.

[0069] Map information is provided by, for example, a graphic G(V j , E j ) indicates that the graph G(V j , E j ) is provided with a branch (side / branch road) E divided at a specified distance in a manner that passes through the center of the lane j (j = 0, 1, 2, ... N), and on each side E j The starting point and the end point are set with nodes V j In the first embodiment, the first E at a point on the map and on a road corresponding to the map is j The side is set to interval E j , the V jSet the node to node V j .

[0070] Regarding the map information, for example, information on a map of the predetermined area 500 may be stored in advance in the base station 100 (see Figure 1 ) of the ROM 102. Alternatively, a map that is periodically updated by a specified map management server (not shown) outside the base station 100 may be received by wireless communication.

[0071] <Business Management Department 304>

[0072] Figure 4 The business management unit 304 shown in FIG. 1 manages the business management unit 304 for each agent 200 (see Figure 1 ) The destination information of the agent 200 is based on the node V contained in the map information. j In addition, the destination information can be either location information based on a specified coordinate system or longitude and latitude information. In addition, the main task (task) is a task related to services such as transportation of people or goods based on the intelligent agent 200. As such main tasks, for example, transportation, emptying, shunting, and charging (refueling) can be listed, but are not limited to these.

[0073] For example, a server (not shown) that accepts customer requests for shunting, etc. may distribute destination information and a main task to the agent 200, and periodically transmit this information to the base station 100 using wireless communication.

[0074] <Intelligent body control unit 300>

[0075] like Figure 4 As shown in FIG. 1 , the agent control unit 300 includes a monitoring state evaluation unit 305, a large area path generation unit 306 (path planning generation unit), a speed limit calculation unit 307, a path correction unit 308, and a path planning communication unit 309. Furthermore, the agent control unit 300 calculates the path of each agent 200 based on the agent individual information, monitoring information, map information, and business information (destination, main task), and transmits the calculation result to each agent 200.

[0076] <Monitoring Status Evaluation Unit 305>

[0077] The monitoring status evaluation unit 305 generates a risk map showing monitoring evaluation indicators for each location based on the agent individual information obtained from the agent information management unit 301, the monitoring information obtained from the monitoring information management unit 302, and the destination and main task information of the agent 200 obtained from the business management unit 304. In addition, the monitoring evaluation indicator is a risk map showing the monitoring evaluation indicators within the specified area 500 (refer to Figure 5A ) is a numerical value that indicates the quality of the monitoring status of each location.

[0078] The risk map is a map in which the values ​​of the monitoring evaluation indexes are associated with each location within the predetermined area 500. The risk map does not need to include edges. j and node V j Graphics (see Figure 5B ), or it may be, for example, to define the area 500 (see Figure 5A ) is divided into grid-like grid data.

[0079] For example, the higher the value of the monitoring evaluation index, the higher the risk of the place (such as the risk of contact with other vehicles or people), and the higher the possibility that monitoring is needed. On the other hand, the lower the value of the monitoring evaluation index, the more thorough the monitoring of the place, and the higher the possibility that the risk is low. In addition, within the specified time ΔT, from the interval E j The more pedestrians and vehicles pass by, the higher the value of the monitoring evaluation index. j The more vehicles (that is, agents 200) that pass by with monitoring tasks, the lower the value of the monitoring evaluation index. The predetermined time ΔT is, for example, a cycle for repeatedly calculating the monitoring evaluation index.

[0080] In the first embodiment, the monitoring state evaluation unit 305 performs a monitoring operation on each section E associated with the map information. j The monitoring status of the edge E at time (T+ΔT) is evaluated. j Monitoring and evaluation index C mj (T+ΔT) is, for example, as shown in the following formula (1), by substituting a predetermined coefficient α oj , coefficient α mj , coefficient α envj , coefficient α esp The product of the product and the monitoring evaluation index C at time T mj (T) is calculated by multiplying. In addition, the coefficient α oj It is based on the time ΔT from interval E j The value is determined by the number of vehicles and pedestrians passing by. mj It is based on the time ΔT from interval E j The value is determined by the number of agents 200 that are responsible for monitoring tasks. envj Is the reflection interval E j The road environment and accident rate of the road. esp is the coefficient of reaction time (forgetfulness coefficient). mj (T) and coefficient α mj The subscript "m" indicates monitoring.

[0081]

Formula 1

[0082] c mj (T+ΔT)=α oj α mj α envj α esp ·c mj (T)···(1)

[0083] <Explanation of monitoring evaluation indicators>

[0084] Fig. 6A is the coefficient α used in the calculation of the monitoring evaluation index oj 's illustration.

[0085] also, Fig. 6A The horizontal axis is the time from interval E to j The number of pedestrians and vehicles passing by, the vertical axis is the coefficient α oj The value of Fig. 6A As shown, the coefficient α oj For example, with an intercept of 1 and a slope of a o is a positive linear function. That is, from the interval E j The more pedestrians and vehicles pass by, the more the coefficient α oj Set a larger value.

[0086] Figure 6B is the coefficient α used in the calculation of the monitoring evaluation index mj 's illustration.

[0087] also, Figure 6B The horizontal axis is the time from interval E to j The vertical axis is the number of monitoring vehicles (that is, the agents 200 that are responsible for monitoring) that pass through, and the vertical axis is the coefficient α mj The value of Figure 6B As shown, the coefficient α mj For example, with an intercept of 1 and a slope of m is a negative linear function. That is, from the interval E j The more monitoring vehicles pass through, the more the coefficient α mj Set a smaller value. In addition, the coefficient α oj (Refer to Fig. 6A ) and coefficient α mj (Refer to Figure 6B ) does not need to be a linear function in particular. For example, these coefficients can also be changed in the form of an exponential function. In addition, table data created in advance based on statistical data related to traffic volume and the number of accidents can also be used.

[0088] The coefficient α in the above formula (1) isenvj It is set by considering the accident rate and the quality of the estimate in the preliminary investigation. For example, for places such as intersections where there are many accidents in the preliminary investigation, the coefficient α envj is set to a value greater than 1, and is set so that the monitoring evaluation index increases as time passes. In addition, for roads that are estimated to be good, and sections with high safety such as guardrails at the boundaries between sidewalks and lanes, the coefficient α envj is set low (e.g., less than 1), and the monitoring evaluation index is set so that it decreases as time passes. By adding a coefficient α that depends on the location to the monitoring evaluation index envj , which can not only reflect the pedestrian traffic and traffic volume, but also reflect the necessity of monitoring corresponding to the on-site conditions.

[0089] In addition, the coefficient α that takes into account the passage of time esp A positive constant is used to set the monitoring evaluation index to increase as time passes. By determining the monitoring evaluation index in this way, a risk map can be created in which the monitoring evaluation index becomes high in sections with few monitored vehicles, in addition to sections with a large number of pedestrians and traffic, and sections with a large number of accidents. In addition, the value of the monitoring evaluation index in each section in the risk map changes all the time.

[0090] <Large Area Path Generating Unit 306>

[0091] Figure 4 The large-area path generation unit 306 (path plan generation unit) shown in the figure generates a large-area path (path plan) for the agent 200 based on the business management information including the moving destination and the type of the task of the agent 200 and the monitoring evaluation index. Specifically, the large-area path generation unit 306 generates a large-area path (path plan) for each agent 200 (see FIG. 1 ) according to a graph-based search method such as the Dijkstra algorithm. Figure 1 ) in a regional path. In addition, the “regional path” represents a rough path for the agent 200 to reach a predetermined destination.

[0092] Fig. 7A It is an explanatory diagram showing the path of an agent in a comparative example.

[0093] also, Fig. 7A The reference numerals CAV1 and CAV2 shown in FIG. 2 are shown as the agent 200 (see FIG. Figure 1 ) is an example of a vehicle. Hereinafter, the term "agent" is also used for these reference numerals CAV1 and CAV2. Fig. 7A, a comparative example of path planning that only considers path length is shown. That is, the shortest path Gp1 is generated when the agent CAV1 moves to the destination Tg1 and then moves to the destination Tg2 after the person gets on the vehicle. In addition, the shortest path Gp2 is generated when another agent CAV2 moves to the destination Tg3.

[0094] <Description of Operation of Large Area Path Generating Unit 306>

[0095] Figure 7B It is an explanatory diagram showing a large-area path of an agent in the agent control system according to the first embodiment.

[0096] For example, it is assumed that the agents CAV1 and CAV2 existing in the predetermined area 500 perform patrol (that is, monitoring) and carry out tasks such as transporting, emptying, and charging. In addition, the above-mentioned "patrol" specifically refers to the control of the slave agent control unit 300 (refer to Figure 4 ) sends the path planning, the agents CAV1 and CAV2 move within the specified area 500 and use the external recognition sensor 207 (refer to Figure 1 ) detects people and other vehicles around, and sends the detection results as monitoring information to the monitoring information management unit 302 (refer to Figure 4 ). After the agent CAV1 reaches the destination Tg1 after completing the shunting task, it picks up passengers and performs the transportation task to the destination Tg2. At this time, let Figure 7B The monitoring evaluation index of the section E1 shown is relatively high. The agent CAV1 carrying passengers and moving performs a transportation task, and if the shortest path is selected, it will select a path Gp1 that passes through the section E1.

[0097] In contrast, in the first embodiment, the agent CAV1 detours to reduce risk and selects the path Gp3 leading to the destination Tg2. As a result, the reliability of the agent CAV1 in performing the transportation task is improved. In addition, the agent CAV1 also monitors its own driving path while performing the transportation task.

[0098] in addition, Figure 7B The agent CAV2 shown is assumed to be traveling to the destination Tg3 without any passengers for charging. In addition, the monitoring evaluation index of the intermediate section E2 is assumed to be relatively high. In this case, in the first embodiment, the monitoring status of the sections with high monitoring evaluation indexes within the specified area 500 is improved by making the agent CAV2 take on the monitoring task. Specifically, the agent CAV2 selects Fig. 7A The shortest path Gp2 shown in FIG. 1 is different from the path Gp4 that passes through the interval E2. Furthermore, the agent CAV2 uses the external recognition sensor 207 (refer to FIG. 1 ). Figure 2) performs perimeter recognition of section E2, and after improving the monitoring evaluation index of section E2 based on the above formula (1), leads to destination Tg3.

[0099] Thus, in the first embodiment, the agents CAV1 and CAV2 perform both monitoring and main tasks (transportation, shunting, emptying, charging, etc.). Figure 5A )A dedicated monitoring intelligent agent (not shown) and fixed sensors (not shown) are set up, so in addition to improving the overall efficiency of the system, costs can also be reduced.

[0100] <Monitoring task ratio β i >

[0101] In order to realize the above-mentioned path planning, in the first embodiment, the large area path generation unit 306 (refer to Figure 4 ) In addition to the type of the main task, it is also based on the external recognition sensor 207 (refer to Figure 2 ) capabilities (sensing range, resolution, environmental tolerance, etc.), and assign monitoring tasks to the agent 200. In addition, the ratio of the importance of monitoring to the main task (task) is called the monitoring task ratio β i (i is the identification number of agent 200).

[0102] For example, the monitoring task ratio β i The higher the value, the agent 200 will select a path through which the monitoring evaluation index of the surrounding environment is relatively high. i The lower the monitoring evaluation index is, the higher the safety is. Figure 5A ) of each interval E j (Refer to Figure 5B ) is calculated. On the other hand, the monitoring task ratio β i Individually set in a one-to-one correspondence with a plurality of agents 200.

[0103] <Monitoring task ratio β i How to set >

[0104] Monitoring task ratio β i Based on the type of main task of the agent 200 and the external recognition sensor 207 (see Figure 2 ) is determined by the sensor capabilities (sensing range, resolution, environmental tolerance, etc.). In addition, the monitoring task ratio β iIt can also be determined based on factors other than these. For example, the range of the agent 200 can be considered, and environmental factors such as weather and time can also be added. For example, it can be set that the longer the range of the agent 200 is, the higher the monitoring task ratio β is. i In addition, it can also be set that in the case of bad weather such as rainy or snowy days, the monitoring task ratio β is higher than that in the case of sunny and cloudy days. i In addition, it can also be set that the monitoring task ratio β is higher during the night time period than during other time periods. i Get taller.

[0105] Data identifying the main tasks assigned to the agent 200 is received from the business management unit 304 (see Figure 4 ) is output to the large-area path generation unit 306. The main tasks of the agent 200 include, for example, transportation, emptying, shunting, charging, and storage. In the case where there is no special passenger or cargo (emptying, shunting, charging, and storage), the large-area path generation unit 306 increases the monitoring task ratio β i , increasing the ratio of monitoring tasks that the agent 200 is responsible for. That is, compared with the case where the main task (task) is the transportation of people or goods, the large area path generation unit 306 (path planning generation unit) increases the monitoring task ratio β when the main task is to return to empty, shunting or charging. i Set to a high value.

[0106] On the other hand, for the agent 200 assigned the transport task as the main task, the large-area path generation unit 306 reduces the monitoring task ratio β i , reducing the ratio of monitoring tasks that the agent 200 is responsible for. In addition, there are also cases where the main task of the agent 200 is switched midway, such as switching from "shunting" to "transportation". In addition to the switching time of the main task, the monitoring task ratio β can also be adjusted according to the cruising range of the agent 200, weather, or time period. i There are also changes.

[0107] Monitoring task ratio β i is set to a value predetermined for each main task, for example. Figure 2 ) is different, the higher the performance (sensing range, resolution, durability, etc.) of the external recognition sensor 207, the higher the monitoring task ratio β of the large area path generation unit 306 (path planning generation unit) i Specifically, it is also possible to consider the outside world recognition sensor 207 (refer to Figure 2 ) monitoring task ratio β iThe added value is preset for each agent 200, and the monitoring task ratio β of the agent 200 is calculated by adding the added value to the monitoring task ratio specified based on the main task category. i .

[0108] <Interval E in path calculation j Weight calculation method of

[0109] The interval E used in the calculation of the path of the agent 200 j The weight w j For example, using the interval E j The path length is related to the weight w dj Sum and interval E j The weight w of the monitoring evaluation index mj , calculated based on the following formula (2).

[0110]

Formula 2

[0111] w j =w dj +w mj (c mj ,β i )···(2)

[0112] The large area path generation unit 306 (refer to Figure 4 ) For example, the weight w when the agent 200 moves to a specified destination j The regional path of the agent 200 is set in such a way that the sum of is minimized. The weight w related to the path length contained in formula (2) dj By measuring the section E in advance from the map information j In addition, interval E j The longer the distance, the greater the weight w dj In addition, the weight w related to the size of the monitoring evaluation index is mj For example, based on the monitoring task ratio β i And monitoring evaluation index C mj And calculated.

[0113] Figure 8 It represents the monitoring evaluation index C mj With weight w mj A diagram illustrating the relationship between .

[0114] also, Figure 8 The horizontal axis is the interval E j Monitoring evaluation index C mj .in addition, Figure 8 The vertical axis is related to the monitoring evaluation index C mj The relevant weight w mjFor example, set the monitoring task ratio β i is the specified value β th In the above case, monitoring evaluation index C mj The larger the weight w is, the mj The smaller (refer to Figure 8 That is, the large area path generation unit 306 (path planning generation unit) monitors the task ratio β i The higher the value, the intelligent agent 200 will be driven to monitor the evaluation index C. mj The path planning is generated in a relatively high-speed way. i The agent 200 with a relatively high value gives priority to monitoring the path with poor monitoring status. th The weight w is set in the large area path generation unit 306 mj When it becomes a decision to use Figure 8 The monitoring task ratio β of the judgment standard is which of the straight lines M1 and M2 shown in FIG. i The threshold is pre-set.

[0115] In addition, it is set to monitor the task ratio β i Less than the specified value β th In the case of mj The larger the weight w is, the mj The larger (refer to Figure 8 That is, the large area path generation unit 306 (refer to Figure 4 ) makes the monitoring evaluation index C mj The weight of a large and risky path is w mj This makes it possible to prevent the agent 200 that is transporting people or goods from traveling on a path with poor monitoring conditions, for example.

[0116] In addition, the weight w mj The calculation method is not limited to Figure 8 For example, it can also be that the monitoring evaluation index C mj and weight w mj The slope of the straight line becomes the monitoring task ratio β i Another option is to use the weight w mj Relative to monitoring evaluation index C mj changes in the form of an exponential function. In addition, the weight w j The calculation formula of is not limited to formula (2), for example, it can also be, in order to increase the monitoring task ratio β i The large area path generation unit 306 uses the following formula (3). In addition, the monitoring task ratio β contained in formula (3) i The range is 0≤β i ≤1.

[0117]

Formula 3

[0118] w j =(1-β i )·w dj +β i ·w mj (c mj ,β i )(0≤βi≤1)···(3)

[0119] The large area path generation unit 306 (refer to Figure 4 ) Using the interval E calculated in this order j The weight w j ,Through graph-based search methods such as Dijkstra’s algorithm, the regional paths of each agent 200 are calculated.

[0120] <Speed ​​Limit Calculation Unit 307>

[0121] Figure 4 The speed limit calculation unit 307 shown in FIG. Figure 4 ) The monitoring evaluation index C calculated mj and the large area path generation unit 306 (refer to Figure 4 ) The monitoring task ratio β calculated i , and calculate the speed limit of the agent 200. In addition, the calculation method of the speed limit will be described later.

[0122] <Path Correction Unit 308>

[0123] Figure 4 The path correction unit 308 shown in FIG. 2 corrects the path of the agent 200. That is, the path correction unit 308 refers to the path generated by the large area path generation unit 306 (refer to Figure 4 ) is used to calculate the large-area path, and the speed limit calculated by the speed limit calculation unit 307 is further used as a constraint to generate a target path for the agent 200 to follow. In addition, by using a framework such as model predictive control for path correction, a formula that explicitly incorporates speed constraints can be achieved.

[0124] Fig. 9A It is an explanatory diagram showing the traffic conditions of a section within a predetermined area at time t=k.

[0125] In addition, Fig. 9A In the example, it is assumed that the intelligent body CAV4 is equipped with an external recognition sensor 207 with a large sensing range (refer to Figure 2 )(See the dotted triangle.) In addition, Fig. 9AIn the figure, the temporal position of each calculation step of agents CAV3 and CAV4 is indicated by a black triangle mark.

[0126] At time t = k (k is the calculation step), Fig. 9A As shown, the agent CAV3 performs the delivery task, and the other agent CAV4 performs the return task. In addition, it is assumed that both agents CAV3 and CAV4 are in the process of moving through the intersection CR13 to the node V5. In addition, it is assumed that the interval E in the intersection CR13 j The monitoring evaluation index shares the same value in agents CAV3 and CAV4.

[0127] <Description of Operation of Path Correction Unit 308>

[0128] If set to Fig. 9A In the state shown, the monitoring evaluation index of the intersection CR13 is relatively large, which means that the risk for the passage of the intelligent body CAV3, which is in the process of transporting people or goods, is high. Therefore, in the first embodiment, by making the intelligent body CAV3 bear the speed restriction, the arrival time at the intersection CR13 is delayed, and the passing speed at the intersection CR13 is reduced to reduce the risk. On the other hand, when the intelligent body CAV4, which is equipped with an external recognition sensor 207 with high sensing capability and is in the air, passes through the intersection CR13, compared with the intelligent body CAV3, there is no worry about damage to the goods, etc. and the risk is low, so the speed limit calculation unit 307 (refer to Figure 4 ) No special speed restriction is required.

[0129] Fig. 9B It is an explanatory diagram showing the traffic conditions of a section within a predetermined area at time t=k+1.

[0130] For example, in Fig. 9A At time t=k, it is assumed that the environment recognition sensor 207 of the agent CAV4 senses the condition of the intersection CR13 and the monitoring evaluation index C of the intersection CR13 is m13 In the case of improvement, the risk of intersection CR13 is reduced. m13 When the value of is improved, the risk of the agent CAV3 passing through the intersection CR13 is reduced, so the speed limit calculation unit 307 (refer to Figure 4 )Remove the speed constraint imposed on agent CAV3.

[0131] In addition, when there are pedestrians or other vehicles (vehicles not belonging to the agent 200) at the intersection CR13, the speed limit calculation unit 307 (see Figure 4 ) It is also possible to impose constraints on the agent CAV4 to avoid collisions with obstacles. Thus, it is possible to generate a path for the agent CAV4 to avoid pedestrians and vehicles, or to generate a path that slows down or stops right in front of pedestrians and vehicles.

[0132] <Formulation of the system using model predictive control>

[0133] If the vector containing the position and orientation of the i-th agent 200 is set to p i , the large area path generation unit 306 (refer to Figure 4 ) The nodes on the regional path Gpi of the i-th agent 200 calculated are set to V i , and set the imaginary target position to r i , then it is expressed as the following equations (4a) and (4b). In addition, vector p i and the imaginary target position r i The deviation between them is set to e i , and is defined as in the following formula (4c). In addition, k refers to a calculation step (time).

[0134]

Formula 4

[0135] p i (k) = [x i (k) y i (k) θ i (k)] T ···(4a)

[0136]

Formula 5

[0137] r i (k) = [xr i (k) yr i (k) θr i (k)] T ···(4b)

[0138]

Formula 6

[0139] e i (k): = p i (k)-r i (k)···(4c)

[0140] The motion model of the agent 200 is formulated as, for example, equation (5). i and steering angle is equivalent to the control command of the agent 200. In addition, L is the control command of the agent 200 (refer to Figure 3 )'s distance in the front-to-rear direction between the front wheel 221 and the rear wheel 222.

[0141]

Formula 7

[0142]

[0143] In addition, the speed v of the i-th agent 200 is integrated i and steering angle The control command vector u i It is defined by formula (6).

[0144]

Formula 8

[0145] u i (k)=[v i φ i ] T ···(6)

[0146] Furthermore, by discretizing equation (5) at a predetermined sampling period Δt, the following equation (7) is obtained.

[0147]

Formula 9

[0148] p i (k+1)=p i (k)+Δt·f(p i (k),u i (k))···(7)

[0149] Path correction unit 308 (see Figure 4 ) The position p of agent 200 at each time k i (k) and the imaginary target position r on the large area path Gpi i (k) The deviation between i (k) is reduced, and the instruction vector u is calculated i (k).

[0150] In addition, by using the model predictive control formulation, Q i , R i Assuming that the weight matrix is ​​Np and the prediction step is Np, the following equation (8) is obtained regarding the evaluation function J. Equation (8) is used for optimizing the path from a predetermined time k0 to a time (Np+k0-1).

[0151]

Formula 10

[0152]

[0153] In model predictive control, at each time k, the optimal control input u is calculated in such a way as to minimize the evaluation function J represented by equation (8): i (k) By using the optimal control input u iSubstitute (k) into equation (7) to calculate the position (x(k), y(k)) and orientation θ(k) of agent 200 at each time k. Then, generate time series data (data from time k0 to Np steps ago) determined by the position (x(k), y(k)) and orientation θ(k).

[0154] <Ratio to monitoring tasks β i Corresponding speed constraints>

[0155] Figure 4 The speed limit calculation unit 307 shown in FIG. 1 monitors the task ratio β. i In addition, based on interval E j Monitoring evaluation index C mj For example, the speed constraint represented by the following equation (9) is calculated. Thus, a constraint condition is given to the predetermined evaluation function J, so that the speed of the agent 200 is subjected to the predetermined constraint.

[0156]

Formula 11

[0157]

[0158] That is, in monitoring the evaluation index C mj Than the specified threshold C th High, and monitoring task ratio β i Less than the specified value β th In the case of i Suppressed to the specified speed limit v slow In addition, in other cases, the speed v of the agent 200 is i Suppressed to the specified speed upper limit v max In addition, the speed limit v slow The value is set to the upper limit value v of the specific speed max As described above, the speed limit calculation unit 307 (refer to Figure 4 ) in monitoring task ratio β i Less than the specified value β th In the case of mj The speed limit when driving in relatively high areas v slow Ratio of monitoring tasks to β i is the specified value β th The speed limit (speed upper limit v max ) is low. By using the above formula (9), the monitoring task ratio β can be set i The low agent 200 monitors the evaluation index C at a low speed mj High driving area.

[0159] Furthermore, the path correction unit 308 suppresses the traveling speed of the agent 200 to be below the speed limit based on the prescribed speed limit and monitoring information, and corrects the traveling path of the agent 200 in a manner to avoid contact between the agent 200 and surrounding objects.

[0160] In addition, the speed restriction is not limited to the above conditions. For example, in order to make the agent 200 in the transport task enter the monitoring evaluation index C mj The speed is restricted before the high section, and the speed limit calculation unit 307 (refer to Figure 4 ) Based on the monitoring evaluation index C of the next section that the agent 200 is to travel mj , impose speed constraints on each agent 200. The speed limit v slow It is desirable to have a speed that can adequately cope with sudden jumps, etc. In addition, for the speed upper limit value v max , for example, using the lower value of the maximum speed of agent 200 and the legal speed.

[0161] <Restrictions on avoiding obstacles>

[0162] Fig.10 It is an explanatory diagram showing conditions for avoiding contact with an obstacle.

[0163] Alternatively, in order to avoid contact with vehicles (including other agents 200) or pedestrians, the agent 200 may impose a constraint on the relative distance between the agent 200 and other objects, for example, on the evaluation function J. Fig.10 As shown, if the width of the i-th agent 200 is set to w i , length is set to L i , then the agent 200 can be moved with the radius ra of formula (10) i surrounded by a circle.

[0164]

Formula 12

[0165]

[0166] In addition, the distance between the obstacle Q1 that may come into contact with the i-th agent 200 and the agent 200 is expressed by the following equation (11).

[0167]

Formula 13

[0168]

[0169] Therefore, if the constraint condition of the following formula (12) is satisfied, the contact between the i-th agent 200 and the obstacle Q1 can be avoided. obj Indicates the size of obstacles Q1 such as pedestrians or other vehicles.obj The value of can be used, for example, by the external recognition sensor 207 (refer to Figure 2 ) calculated, or a pre-set value that does not depend on the type of obstacle Q1 may be used.

[0170]

Formula 14

[0171] d ij >ra i +r obj ··· (12)

[0172] Path correction unit 308 (see Figure 4 ) Based on the constraint conditions of formula (12), a control input sequence U(k)=[U1(k)···UN(k)] is calculated to minimize the evaluation function J of formula (8). In this way, each intelligent agent 200 can avoid collision with other vehicles or pedestrians and calculate an efficient movement path.

[0173] <Route planning communication unit 309>

[0174] Figure 4 The route planning transmission unit 309 shown in FIG. 1 transmits the route planning data to each agent 200. That is, the route planning transmission unit 309 transmits the route correction unit 308 (see FIG. 1 ) Figure 4 ) The calculated path is sent to each agent 200 by wireless communication. Each agent 200 performs following control in a prescribed manner based on the path sent by the path planning communication unit 309.

[0175] <Flowchart of the intelligent body control unit>

[0176] Fig.11 is a flowchart showing the operation flow of the intelligent body control unit (see also Figure 4 ).

[0177] First, in step S601 , the agent control section 300 acquires individual information (agent individual information) of the agent 200 as a control target from the agent information management section 301 .

[0178] Next, in step S602 , the agent control unit 300 obtains the destination information and task information of the agent 200 from the business management unit 304 .

[0179] In step S603 , the agent control unit 300 acquires monitoring information including the position of the agent 200 and the current positions of pedestrians and traveling vehicles in the predetermined area 500 from the monitoring state evaluation unit 305 .

[0180] In step S604, the agent control unit 300 uses the monitoring state evaluation unit 305 to calculate the monitoring evaluation index of each section in the specified area 500 (monitoring state evaluation processing). That is, the agent control unit 300 calculates the monitoring evaluation index of each section based on the position of the agent 200 obtained from the monitoring state evaluation unit 305 and the current positions of pedestrians and vehicles in the specified area 500.

[0181] In step S605, the agent control unit 300 uses the large-area path generation unit 306 to calculate the large-area path of the agent 200 and calculate the monitoring task ratio β i (Path planning generation process). That is, in addition to the monitoring evaluation index of each section calculated in step S604, the agent control unit 300 also calculates the regional path Gpi and the monitoring task ratio β of each agent 200 based on the map information obtained by the map information management unit 303 and the destination information and main task information of the agent 200 obtained by the business management unit 304. i In addition, for the agent 200 that has already calculated the large-region path, in order to shorten the calculation time, the calculation result of the previous large-region path can also be used directly.

[0182] Next, in step S606, the agent control unit 300 uses the speed limit calculation unit 307 to calculate the speed limit of the agent 200. That is, the agent control unit 300 calculates the speed limit of the agent 200 based on the risk map representing the monitoring evaluation index of each location calculated by the monitoring state evaluation unit 305 and the monitoring task ratio β calculated by the large area path generation unit 306. i , calculate the speed limit of agent 200.

[0183] In step S607, the agent control unit 300 uses the path correction unit 308 to correct the path of the agent 200. That is, the agent control unit 300 calculates the path between nodes based on the speed limit of the agent 200 calculated by the speed limit calculation unit 307 and the obstacle information around the agent 200.

[0184] In step S608, the agent control unit 300 uses the path planning communication unit 309 to transmit the information of the path calculated in step S607 to each agent 200 by wireless communication (path planning communication processing). Figure 1 ) Each intelligent agent 200 that receives the path information performs following control within the specified area 500 based on the path information.

[0185] By repeatedly performing the above-mentioned processing, each agent 200 within the predetermined area 500 is controlled by the agent control unit 300 .

[0186] <Effect>

[0187] According to the first embodiment, the intelligent agent 200 takes on the monitoring task according to the capabilities of each intelligent agent 200 and the type of the main task within the specified area 500. Therefore, there is no need to set up an intelligent agent (not shown) and a fixed sensor (not shown) that only performs monitoring, so monitoring and transportation can be performed with a smaller number of intelligent agents 200 as a whole. As a result, in addition to improving the efficiency of the entire system, the operating cost can also be reduced. In addition, the intelligent agent control unit 300 calculates the path of the intelligent agent 200 based on the monitoring evaluation index, thereby ensuring the reliability of each intelligent agent 200 and establishing efficient path planning. In this way, according to the first embodiment, an intelligent agent control system W1 can be provided that enables the intelligent agent 200 to perform specified tasks while performing monitoring.

[0188] <<Second Implementation Method>>

[0189] In the second embodiment, the agent control unit 300A (see Fig.12 ) is provided with a monitoring rate calculation unit 310 (refer to Fig.12 ), and based on fixed sensors 400, etc. (refer to Fig.12 ) is different from the first embodiment in that a predetermined monitoring rate is set corresponding to the monitoring evaluation index of the detection result ...

[0190] Fig.12 This is a functional block diagram of the agent control system W2 according to the second embodiment.

[0191] like Fig.12 As shown, the agent control system W2 is configured to include a base station 100A, agents 200-1 to 200-n, and fixed sensors 400-1 to 400-m. The fixed sensors 400-1 to 400-m are, for example, cameras, and are installed in a predetermined area 500 (see Figure 5A ) intersection, etc. And the detection results of the fixed sensors 400-1 to 400-m are transmitted to the base station 100A at all times. In addition, the fixed sensors 400-1 to 400-m are collectively referred to as fixed sensors 400.

[0192] In the second embodiment, a fixed sensor 400 and an agent 200 having a monitoring task are installed at an intersection where there is a lot of traffic of people and vehicles, and a predetermined area 500 (see Figure 5A). By setting up multiple fixed sensors 400, the overall sensing capability of the system is improved, but on the other hand, the overall communication volume of the system is also increased. Therefore, in the second embodiment, based on the monitoring evaluation index of each place, the monitoring rate of the intelligent body 200 and the fixed sensor 400 in the place with poor monitoring conditions and the place with a lot of pedestrian traffic is set high, and the monitoring rate of the place with less pedestrian traffic and less traffic is set low. In this way, monitoring in the specified area 500 can be properly performed and the communication volume can be reduced.

[0193] <Monitoring rate control when there is a fixed sensor>

[0194] Fig.13 It is a functional block diagram showing the system configuration of the base station 100A.

[0195] like Fig.13 As shown, the agent control unit 300A includes a monitoring state evaluation unit 305, a large area path generation unit 306, a speed limit calculation unit 307, a path correction unit 308, and a path planning communication unit 309, and also includes a monitoring speed calculation unit 310. Fig.12 ) in addition to the moving path, the monitoring rate of the intelligent agent 200 and the fixed sensor 400 is also calculated, and the calculation results are sent to the intelligent agent 200 and the fixed sensor 400.

[0196] <Intelligent Body Information Management Unit 301>

[0197] Fig.13 The agent information management unit 301 shown in FIG. Fig.12 ) also retains the individual information of the fixed sensor 400 (refer to Fig.12 The individual information of the fixed sensor 400 includes the installation position of the fixed sensor 400 and sensor information (sensing range, resolution, environmental tolerance, etc.).

[0198] <Monitoring Rate Calculation Unit 310>

[0199] Fig.14 is an explanatory diagram showing an operation example of the monitoring rate calculation unit (see also Fig.13 ).

[0200] also, Fig.14 The "Sensor Acquisition" in FIG. 4 shows a fixed sensor 400 (refer to Fig.12 ) and agent 200 (refer to Fig.12 ) generates monitoring information from the external recognition sensor 207. Fig.14 The "monitoring process" in FIG. 1 shows that the fixed sensor 400 and the agent 200 send a signal to the base station 100A (refer to Fig.12 ) sends monitoring information. In addition, Fig.14 The horizontal axis of the graph on the left side of the paper is time, and the vertical axis is the monitoring evaluation index C mj . Fig.14 The solid line graph in the figure shows the monitoring evaluation index C in the specified place. mj The dotted line graph shows the monitoring evaluation index C in another place. mj The progress of.

[0201] The monitoring rate calculation unit 310 calculates the communication rate (that is, monitoring rate) of the monitoring information of the agent 200 and the fixed sensor 400 based on the monitoring information obtained from the monitoring state evaluation unit 305 in addition to the individual information of the agent 200 and the individual information of the fixed sensor 400 obtained from the agent information management unit 301. The monitoring rate calculated in this way is sent to the agent 200 and the fixed sensor 400 via the path planning communication unit 309.

[0202] The monitoring rate is the number of times (that is, frequency) the agent 200 or the fixed sensor 400 provides monitoring information to the base station 100A (monitoring state evaluation unit 305 , etc.) per unit time, and is set for each location within the specified area 500 .

[0203] The monitoring rate calculation unit 310 monitors the evaluation index C mj The monitoring rate is calculated. That is, the monitoring rate calculation unit 310 calculates the monitoring evaluation index C mj With the specified threshold C th The monitoring rate is calculated by comparing the size of mj The traffic volume and pedestrian traffic increase and the monitoring evaluation index C mj Become threshold C th In the above case, the monitoring rate calculation unit 310 increases the monitoring rate of the intelligent agent 200 and the fixed sensor 400 in the area. mj The higher the value, the higher the monitoring rate calculation unit 310 makes the monitoring evaluation index C mj The higher the monitoring rate of the corresponding place.

[0204] On the other hand, when the traffic volume decreases at night, the evaluation index C is monitored. mj Less than threshold C th In the case of , the monitoring rate calculation unit 310 reduces the monitoring rate of the agent 200 and the fixed sensor 400. In addition, the method for calculating the monitoring rate is not limited to this, for example, it can also be calculated by the monitoring evaluation index C mj The monitoring rate is varied inversely proportional to the value of

[0205] <Effect>

[0206] According to the second embodiment, the monitoring rate calculation unit 310 changes the monitoring rate of the agent 200 and the fixed sensor 400 based on the monitoring information. Thus, since the communication volume is adjusted according to the monitoring status in the specified area 500, the increase in the communication volume of the entire system can be suppressed. In addition, by installing the fixed sensor 400 at a place where there are many pedestrians, such as an intersection, the sensing capability of the entire system can be improved.

[0207] 《Third Implementation Method》

[0208] In the third embodiment, the agent control system W1 (see Figure 1 ) Theme Park 700 (reference Fig.15 ) is different from the first embodiment in terms of the transportation and monitoring of people, etc. In addition, (the structure of the intelligent control system W1, etc., refer to Figure 1 , Figure 2 , Figure 4 ) is the same as the first embodiment. Therefore, the parts different from the first embodiment will be described, and the description of the overlapping parts will be omitted.

[0209] <Example of application to a moving vehicle in a theme park>

[0210] Fig.15 This is an explanatory diagram of a theme park 700 to which the agent control system according to the third embodiment is applied.

[0211] exist Fig.15 In the example, in theme park 700, in addition to the entrance area and the central area, areas A to D are arranged in a manner surrounding the central area. In theme park 700, agents CAV5, CAV6, CAV7, and CAV8 transport passengers from a certain location to a specified destination. Fig.15 The shown situation is that the intelligent agent CAV5 picks up passengers at section A and transports them to the destination Tg5 at section C.

[0212] Fig.16A It is an explanatory diagram showing the path of an agent in a comparative example.

[0213] In addition, it is assumed that a graph G (E j , V j ) map information, the graphic G(E j , V j ) includes nodes V set in front of the main attractions j And each node V j Connected interval E j. Assume that currently, the section E10, which is the passage from the central section to the section C, is crowded with pedestrians, and the monitoring evaluation index of the section E10 is relatively high (that is, the risk is high). Assume that the agent CAV5 is carrying passengers and moving along the shortest path, and calculate Fig.16A A predetermined path Gp5 passing through the section E10 as shown in the comparative example.

[0214] Fig. 16B It is an explanatory diagram showing a large-area path of an agent in the agent control system according to the third embodiment.

[0215] In the third embodiment, the agent CAV5 for transporting people is set to pass through a path where the monitoring evaluation index is low. For example, a path Gp6 is calculated to pass through the section E11 where the risk is lower than the above-mentioned section E10. In addition, when the agent CAV6 is in the process of returning to empty in the central area, a path Gp5 is calculated to allow the agent CAV6 to patrol the section E10 where the monitoring evaluation index is relatively large.

[0216] Fig.17 It is an explanatory diagram showing a travel path of an agent in the agent control system according to the third embodiment.

[0217] Based on the above calculation results, agents CAV5 and CAV6 finally Fig.17 In this way, the agent control system W1 (refer to Figure 1 ) According to the content of the main tasks of the agents CAV5 and CAV6, a path planning is generated in which each agent is responsible for the monitoring task. Thus, during the process of transporting passengers, the agent CAV5 and the like can advance along a path with low risk, thereby ensuring safety. In addition, by having the agent CAV6 monitor a high-risk area during the emptying process, the monitoring status of the theme park 700 can be improved.

[0218] 《Variation Example》

[0219] As mentioned above, although the agent control systems W1, W2, etc. of the present invention are described in each embodiment, the present invention is not limited to these descriptions, and various modifications can be made.

[0220] For example, in each embodiment, the case where the monitoring evaluation index is calculated based on the agent individual information, monitoring information, and the destination and main task information of the agent 200 is described, but it is not limited to this. That is, when calculating the monitoring evaluation index, part of the above information (for example, the agent individual information and / or the destination, the main task information) may be omitted. That is, the monitoring status evaluation unit 305 may calculate the monitoring evaluation index representing the quality of the monitoring status of each place in the specified area based on the monitoring information sent from the agent 200 moving in the specified area. In such a structure, the same effect as in each embodiment is obtained.

[0221] In the first embodiment, the agent control unit 300 (see Figure 4 ) is provided with a speed limit calculation unit 307 and a path correction unit 308, but is not limited thereto. That is, one or both of the speed limit calculation unit 307 and the path correction unit 308 may be omitted from the structure of the agent control unit 300. In addition, the same can be said for the second embodiment and the third embodiment.

[0222] In addition, in the first embodiment, if Figure 4 As shown in the figure, the case where the agent information management unit 301, the monitoring information management unit 302, the map information management unit 303, the business management unit 304 and the agent control unit 300 are arranged in one base station 100 is described, but it is not limited to this. That is, part of the above-mentioned structures can also be arranged in other servers or specified agents 200.

[0223] In the first embodiment, the following description is given of the method based on the following example: j The number of vehicles and pedestrians passing by, the number of agents 200 responsible for monitoring, and the interval E j The monitoring evaluation index C is calculated based on the road environment and accident rate. mj That is, the monitoring state evaluation unit 305 may calculate the monitoring evaluation index C based on at least one of the traffic volume of pedestrians and ordinary vehicles (other vehicles than the agent 200, and vehicles not specifically monitored), the traffic volume of monitoring vehicles including the agent 200, and the road condition. mj .

[0224] In the first embodiment, the monitoring task ratio β is set based on the type of the main task (task) of the agent 200 and the performance of the external recognition sensor 207. i That is, the large area path generation unit 306 (path planning generation unit) may also monitor the task ratio β based on the type of the main task (task) contained in the business management information. iThe setting is established corresponding to the intelligent agent 200.

[0225] In the second embodiment, the agent control system W2 (see Fig.12 ) is provided with a plurality of fixed sensors 400, but the present invention is not limited thereto. That is, the monitoring rate calculation unit 310 (see Fig.13 ) calculates the monitoring rate based on the monitoring evaluation index and sends the monitoring rate to the agent 200. In such a structure, the same effect as the second embodiment is achieved.

[0226] Alternatively, the risk map showing the monitoring evaluation index of each location described in the first embodiment may be displayed on a display (display device) of a manager or the like. Alternatively, the monitoring rate described in the second embodiment may be associated with the identification number of the fixed sensor 400 and the intelligent body 200 and displayed on a display (display device) of a manager or the like. Thus, the manager or the like can easily grasp the monitoring evaluation index and the monitoring rate in the specified area 500.

[0227] In addition, in each embodiment, the predetermined area 500 (see Figure 5A )'s automatic conveyor system and theme park 700 (reference Fig.15 ), but the applicable targets of the present invention are not limited to this. For example, it can also be applied to factories, ports or logistics warehouses where people and mobile bodies are mixed, and it can also be applied to other areas such as agriculture (farms) or tourism.

[0228] In addition, each embodiment can be combined appropriately. For example, the second embodiment and the third embodiment can be combined, and the system structure of the second embodiment (see Fig.12 , Fig.13 ) is applicable to the theme park 700 described in the third embodiment (refer to Fig.15 ).

[0229] In addition, the processing in the agent control system W1 and the like may be executed as a predetermined program of a computer. The above-mentioned program can be provided via a communication line, and can also be written in a recording medium such as a CD-ROM and distributed.

[0230] In addition, each embodiment is described in detail to explain the present invention in an easy-to-understand manner, but is not limited to having all the structures described. In addition, with respect to a part of the structure of the embodiment, other structures can be added, deleted, or replaced. In addition, the above-mentioned mechanisms and structures show the mechanisms and structures that are considered necessary for the description, but the product is not limited to having to show all the mechanisms and structures.

[0231] Description of Reference Numerals

[0232] 100, 100A base station

[0233] 200-1, 200-2, ..., 200n agents

[0234] 207 External Recognition Sensor

[0235] 301, 300A Intelligent body control unit

[0236] 301 Intelligent Information Management Department

[0237] 302 Surveillance Information Management Department

[0238] 303 Map Information Management Department

[0239] 304 Business Management Department

[0240] 305 Monitoring Status Evaluation Department

[0241] 306 Region Path Generation Department (Path Planning and Generation Department)

[0242] 307 Speed ​​limit calculation unit

[0243] 308 Path Correction Department

[0244] 309 Route Planning and Communication Department

[0245] 310 Monitoring rate calculation unit

[0246] 400, 400-1, 400-2, ..., 400-m fixed sensor

[0247] 500 Specified Area

[0248] 700 Theme Park (specified area)

[0249] CAV1, CAV2, CAV3, CAV4, CAV5, CAV6, CAV7, CAV8 agents

[0250] W1, W2 intelligent agent control system.

Claims

1. An intelligent agent control system, characterized in that: have: A monitoring status evaluation unit that calculates a monitoring evaluation index based on monitoring information sent from an agent moving within a specified area, the monitoring evaluation index indicating the quality of the monitoring status of each location within the specified area; A path planning generating unit, which generates a path planning for the agent based on the business management information including the moving destination of the agent and the type of the task, and the monitoring evaluation index; and A path planning communication unit sends the path planning data to the intelligent agent.

2. The intelligent agent control system according to claim 1, characterized in that: The path planning generation unit sets a monitoring task ratio corresponding to the intelligent agent based on the category of the task contained in the business management information, which represents the importance of monitoring relative to the ratio of the task, and generates the path planning in a manner such that the higher the monitoring task ratio, the higher the intelligent agent travels on the path with the higher monitoring evaluation index.

3. The intelligent agent control system according to claim 2, characterized in that: The route planning generation unit sets the monitoring task ratio to a higher value when the task is movement for emptying, shunting, or charging, compared to when the task is transportation of people or goods.

4. The intelligent agent control system according to claim 2, characterized in that: The higher the performance of the external environment recognition sensor mounted on the intelligent body, the higher the value that the path planning generation unit sets the monitoring task ratio.

5. The intelligent agent control system according to claim 2, characterized in that: A speed limit calculation unit is provided, which calculates a speed limit of the agent based on the monitoring evaluation index and the monitoring task ratio, The speed limit calculation unit sets a speed limit when the agent travels in an area where the monitoring evaluation index is relatively high, when the monitoring task ratio is less than a predetermined value, lower than the speed limit when the monitoring task ratio is greater than the predetermined value.

6. The intelligent agent control system according to claim 5, characterized in that: A path correction unit is provided for correcting the path of the intelligent body. The path correction unit corrects the travel path of the agent based on the speed limit and the monitoring information so as to suppress the travel speed of the agent to be below the speed limit and avoid contact between the agent and surrounding objects.

7. The intelligent agent control system according to claim 1, characterized in that: The monitoring state evaluation unit calculates the monitoring evaluation index based on at least one of the traffic volume of pedestrians and ordinary vehicles, the traffic volume of monitoring vehicles including the agent, and the road condition.

8. The intelligent agent control system according to claim 1, characterized in that: A monitoring rate calculation unit is provided, which sets a monitoring rate for each place in the specified area based on the monitoring evaluation index, and the monitoring rate is the number of times the intelligent agent provides monitoring information to the monitoring state evaluation unit per unit time, The monitoring rate calculation unit increases the monitoring rate of a location corresponding to the monitoring evaluation index as the monitoring evaluation index increases.

9. An intelligent agent control method, characterized in that: Including the following processing: A monitoring status evaluation process, based on monitoring information sent from an intelligent agent moving within a specified area, calculates a monitoring evaluation index, the monitoring evaluation index indicating the quality of the monitoring status of each place within the specified area; A path planning generation process, which generates a path planning of the agent based on the business management information including the moving destination of the agent and the type of the task, and the monitoring evaluation index; and Path planning communication processing sends the path planning data to the intelligent agent.

Images