WO2021070239A1

WO2021070239A1 - Autonomous distributed control system

Info

Publication number: WO2021070239A1
Application number: PCT/JP2019/039594
Authority: WO
Inventors: 允裕山隅
Original assignee: 三菱電機株式会社
Priority date: 2019-10-08
Filing date: 2019-10-08
Publication date: 2021-04-15
Also published as: JP6742561B1; JPWO2021070239A1

Abstract

The purpose of the present invention is to provide a technology with which global convergence stability can be satisfied even when a coverage region changes dynamically in an autonomous distributed control system. A first mobile body is provided with: a coverage boundary acquiring unit that performs mapping for generating a virtual mobile body on the opposite side from the first mobile body with respect to the coverage boundary; a coverage control computing unit that assumes that the first mobile body, a plurality of second mobile bodies, and the virtual mobile body are coupled by a network and generates a command value which indicates a control of the first mobile body; and a mobile body control unit that controls the movement of the first mobile body on the basis of the command value.

Description

Autonomous distributed control system

The present invention relates to an autonomous decentralized control system in which a plurality of mobile bodies cover a predetermined covering region by autonomous and decentralized control.

In recent years, a control technology has been proposed in which a plurality of moving objects (hereinafter, each of which is referred to as a "moving body") form a formation to achieve a certain purpose, and the control technology solves the problem with sociality. It is expected to solve the problems that have been solved.

For example, it is expected that a drone that can move freely in space will regularly search a wide range on behalf of a person. Further, for example, it is being realized that a transport vehicle capable of autonomous driving transports goods in a factory or a distribution warehouse instead of a person. Furthermore, it is expected that autonomous vehicles capable of highly autonomous driving without human intervention will perform platooning with a reduced inter-vehicle distance in order to alleviate traffic congestion.

However, in the above technology, it is required to act densely with respect to a plurality of moving bodies and to synchronize their movements with each other. Therefore, if the control of the moving body group is unstable, a big problem may occur. There is sex. Therefore, the conventional control system in which each moving body uses its own information is replaced with a new control system in which a plurality of moving bodies use the information of a plurality of moving bodies as a whole so that a plurality of moving bodies function as a moving body group. A control system has been proposed.

A control system in which multiple mobiles function as a group of mobiles is one in which information on the entire group of mobiles is centrally processed by a central management device, etc., and one in which local information observed by a certain mobile is processed nearby. It is divided into those that are handed over to the moving body and processed in a decentralized manner. Here, the former control system is a method with excellent control efficiency from the viewpoint of control convergence time and energy optimization, but communication congestion due to information concentration and delay of convergence time due to space expansion, etc. Not suitable for large-scale systems due to physical constraints. On the other hand, in the latter control system (hereinafter referred to as "autonomous distributed control system"), each moving body acts by autonomous and decentralized control. In this autonomous distributed control system, since each moving body processes only local information, there are some problems in convergence and stability. However, it is suitable for a large-scale system because a group that processes information on a small scale cooperates with each other to realize the function of the entire control system.

In the autonomous distributed control system, multiple moving objects in the moving body group are arranged at desired positions (agreement), and multiple moving objects in the moving body group are efficiently arranged in space (covering). Do one of the above. It should be noted that, in general, an autonomous distributed control system cannot handle consensus and cover in the same scheme.

The autonomous distributed control system that handles consensus tends to be effective in relatively small-scale systems. On the other hand, autonomous distributed control systems that handle coverings are expected to be applied to relatively large-scale systems. For example, autonomous distributed control systems dealing with coverings can replace a wide range of cost, resolution, and time-constrained satellite or aircraft observations or surveillance with drones or similar inexpensive mobiles. Therefore, it is expected to build a service that is inexpensive, has a high update frequency, and is highly convenient. In addition, an autonomous distributed control system that handles coverings is an efficiency improvement by rectification of mobile objects, or in a delivery system in transportation, electric power infrastructure and shopping malls, or a system composed of 100 or more mobile objects not limited thereto. It is expected that the efficiency will be similar to those.

However, since the autonomous distributed control system controls the entire mobile group with local information, global convergence stability is an issue. In particular, since the autonomous distributed control system that handles the covering is expected to be a large-scale system, it is indispensable for reliability and safety to improve the convergence stability of the control law. In addition, machine learning, which has become widespread in recent years, or similar heuristic control methods, are not suitable for improving convergence stability in autonomous distributed control systems.

For the above problems, some design methods for improving global convergence stability have been proposed in autonomous distributed control systems.

For example, in Non-Patent Document 1, in a state where a network connecting information is formed between a plurality of mobile bodies, the global convergence stability is determined from the mathematical structure thereof, and the Voronoi region. It is proposed to define and obtain the command value based on the center of gravity of the Voronoi region. For example, Non-Patent Document 2 proposes that setting the stop point of the gradient function as a control command value is a necessary and sufficient condition for a control command value that satisfies global convergence stability. As the gradient function, it is proposed to use a function represented by a linear sum of functions whose elements are a set of mobile bodies including any two mobile bodies that form a network and exchange information with each other. .. For example, Patent Document 1 proposes a monitoring system including a flight device and a center device for monitoring the ground from the sky.

Japanese Unexamined Patent Publication No. 2019-20992

However, in the conventional covering control that defines the Voronoi region, all the moving objects to be controlled must share the information about the covering region. Specifically, each mobile must share information about the covering boundary that closes the covering region with each other in order to calculate the Voronoi region. If the information on the covering boundary of each moving body is different, the Voronoi region calculated for each moving body will be different, so that the autonomous distributed control cannot satisfy the global convergence stability. is there.

Due to the above constraints of having to share information about the covering area, it is difficult for conventional covering control to dynamically change a part of the covering boundary, and the covering area must be treated statically. There was a problem that it did not become. As a result, in conventional coating control, a part of the coating region is deformed, two coating regions are combined into one, one coating region is divided into two or more, or a similar coating region is dynamically dynamic. There was a problem that it was difficult to apply it to the real environment because it was difficult to handle such changes.

Therefore, the present invention has been made in view of the above problems, and is a technique capable of satisfying global convergence stability even if the covering region changes dynamically in an autonomous distributed control system. The purpose is to provide.

In the autonomous decentralized control system according to the present invention, a first mobile body and a plurality of second mobile bodies connected by a network connecting information cover a predetermined covering region by autonomous and decentralized control. In the dispersion control system, the first moving body includes a covering boundary acquisition unit that performs a mapping to generate a virtual moving body on the opposite side of the covering boundary that closes the covering region to the first moving body, and the first moving body. It is assumed that the mobile body, the plurality of second mobile bodies connected to the first mobile body by the network, and the virtual mobile body generated by the covering boundary acquisition unit are connected by the network. Then, a covering control calculation unit that generates a command value indicating control of the first moving body and a moving body that controls the movement of the first moving body based on the command value generated by the covering control calculation unit. It is equipped with a control unit.

According to the present invention, a mapping is performed to generate a virtual moving body on the side opposite to the first moving body with respect to the covering boundary confining the covering region, and the first moving body, a plurality of second moving bodies, and the virtual moving body are Is connected by a network, a command value indicating control of the first mobile body is generated, and the movement of the first mobile body is controlled based on the command value. According to such a configuration, in the autonomous distributed control system, it is possible to satisfy the global convergence stability even if the covering region changes dynamically.

The object, features, aspects and advantages of the present invention will be made clearer by the following detailed description and accompanying drawings.

It is a figure which shows typically the structure of the autonomous distributed control system which concerns on Embodiment 1. FIG. It is a figure which shows typically the operation of the autonomous distributed control system which concerns on Embodiment 1. FIG. It is a block diagram which shows the functional structure of the 1st moving body which concerns on Embodiment 1. FIG. It is a figure which shows typically the operation of the autonomous distributed control system which concerns on Embodiment 1. FIG. It is a figure which shows typically the structure of the autonomous distributed control system which concerns on Embodiment 2. FIG. It is a block diagram which shows the functional structure of the 1st moving body which concerns on Embodiment 2. It is a figure which shows typically the operation of the autonomous distributed control system which concerns on Embodiment 2. FIG. It is a figure which shows typically the operation of the autonomous distributed control system which concerns on Embodiment 2. FIG. It is a figure which shows typically the operation of the autonomous distributed control system which concerns on Embodiment 3. It is a figure which shows typically the operation of the autonomous distributed control system which concerns on Embodiment 3. It is a block diagram which shows the functional structure of the 1st moving body which concerns on Embodiment 4.

<Embodiment 1>
FIG. 1 is a diagram schematically showing a configuration of an autonomous distributed control system according to a first embodiment of the present invention. First, an outline of the autonomous distributed control system according to the first embodiment and terms used in the explanation will be described.

The autonomous distributed control system according to the first embodiment includes a mobile group, and the mobile group includes a first mobile body 1, a plurality of second mobile bodies 2, and a plurality of third mobile bodies 3. Includes. Each of the first mobile body 1, the plurality of second mobile bodies 2, and the plurality of third mobile bodies 3, that is, each mobile body, moves autonomously when instructed to perform formation actions in an autonomously decentralized manner. It is a mobile body that forms a group of multiple functional machines. Each moving body is, for example, a drone, an autonomous vehicle, and the like, but is not limited to this.

The first mobile body 1 and the plurality of second mobile bodies 2 are connected by a network 4 that connects information. In this case, the first mobile body 1 and the plurality of second mobile bodies 2 can be referred to as "nodes" of the network 4. Further, the fact that mobiles have a connection of information by connecting networks can be called "adjacent", and the state of having a connection of information (adjacent state) can be called a "branch" of network 4. it can.

Here, the first mobile body 1 is any one mobile body in the mobile body group.

The second moving body 2 is a moving body having an information connection with the first moving body 1 in the moving body group. The second mobile body 2 may be a mobile body having an information connection with the first mobile body 1 without communication with another second mobile body 2, or may be a mobile body with another second mobile body 2. It may be a mobile body having a connection of information with the first mobile body 1 via communication.

The third mobile body 3 is a mobile body having no information connection with the first mobile body 1 in the mobile body group.

The first mobile body 1 and the plurality of second mobile bodies 2 joined by the network 4 cover a predetermined covering region 5 by autonomous and decentralized control. The covering region 5 is a region enclosed by a covering boundary 6 such as a surface or a line, and specifically, is an area occupied by a space enclosed by a curved surface or an area occupied by a plane enclosed by a curve. is there. The number of covering boundaries 6 recognized by the first moving body 1 may be one or may be plural. Unless otherwise specified, the mobile group is intended to cover the same covering region 5.

When the covering area 5 is visible, the first mobile body 1 includes a controller that cooperates with a communication device, a sensor, and an arithmetic unit (not shown). As a result, the first mobile body 1 can acquire the state quantity of the second mobile body 2 from the network 4 and cover the covering region 5, that is, the covering of the covering region 5 can be achieved. At this time, the controller of the first mobile body 1 calculates the center of gravity of the Voronoi region from the observation results of the second mobile body 2 around the first mobile body 1 and covers the covering region 5. A control command value, which is a command value indicating the control of 1, is generated. That is, the first moving body 1 calculates the boronoy region based on all the covering boundaries 6 that close the covering region 5, the position of the first moving body 1, and the position of the second moving body 2, and the first A control command value based on the center of gravity of the Boronoi region including the moving body 1 is generated.

The second mobile body 2 or the third mobile body 3 is also configured in the same manner as the first mobile body 1, and generates a control command value based on the center of gravity of the Boronoi region including the second mobile body 2 or the third mobile body 3. .. By moving each moving body of the moving body group toward the center of gravity of the Voronoi region indicated by each control command value, it is possible to perform a coating in which global convergence stability is satisfied.

In the convergence stability by the coating control, it is a necessary requirement that each of the first moving body 1, the second moving body 2, and the third moving body 3 shares the same covering region 5 and thus the covering boundary 6. Therefore, when the covering region 5 changes, each moving body needs to share the change with the network 4. However, since high performance is required for the means for forming a network, it is difficult to construct a system suitable for a large-scale mobile group having a large number of mobiles.

In view of this, in addition to the above configuration, the first mobile body 1 includes a detector that recognizes the covering boundary 6 and an arithmetic unit that uses the covering boundary 6 as a state quantity. Then, as shown in FIG. 2, the first moving body 1 performs a mapping to generate a virtual moving body 7 on the opposite side of the first moving body 1 with respect to the covering boundary 6 that closes the covering region 5. The first moving body 1 performs the above mapping based on the geometric shape of the covering boundary 6 and the state amount of the first moving body 1 acquired by the first moving body 1, thereby performing the state quantity of the virtual moving body 7. To generate.

At this time, the first moving body 1 and the virtual moving body 7 are treated as adjacent to each other. That is, it is assumed that the first mobile body 1, the plurality of second mobile bodies 2 connected to the first mobile body 1 by the network 4, and the virtual mobile body 7 are connected by the network 4. Under this assumption, the first mobile body 1 generates a control command value indicating control of the first mobile body 1 by, for example, calculating the center of gravity of the Voronoi region.

According to the above configuration, the first moving body 1 locally acquires the information of the covering boundary 6, treats the information of the covering boundary 6 as the virtual moving body 7, and generates a control command value. As a result, even if the covering area 5 dynamically changes, the moving body such as the first moving body 1 in the autonomous distributed control system can change the information in the covering area 5 from the network 4 to the state quantity of the virtual moving body 7. Since it can be acquired as a change, global convergence stability can be satisfied.

From the above, even if the covering region 5 changes dynamically, the moving body group can realize a covering that satisfies the global convergence stability. At this time, since each mobile body generates a control command value only with local information, the means for forming the network 4 can be constructed by a simple mechanism. Therefore, the above system can be realized by a relatively simple and relatively small device.

Next, the details of the autonomous distributed control system according to the first embodiment will be described. FIG. 3 is a block diagram showing a functional configuration of the first mobile body 1 according to the first embodiment. The functional configurations of the second mobile body 2 and the third mobile body 3 may be the same as the functional configurations of the first mobile body 1.

Before explaining the functional configuration of the first mobile body 1 shown in FIG. 3, the hardware configuration of the first mobile body 1 will be described. The first moving body 1 calculates a path between the current position and the target position so that the first moving body 1 moves autonomously, and controls the drive source of the first moving body 1 along the path. It includes a controller, any of a communication device and a detector forming a network 4 with the second mobile body 2, an arithmetic device, and a detector for detecting the covering boundary 6.

The controller that controls the movement of the first mobile body 1 autonomously includes means for observing the state quantity of the first mobile body 1. The means for observing the state quantity of the first moving body 1 is, for example, a speedometer for detecting the speed vector of the first moving body 1. The means for observing the state quantity of the first moving body 1 is not limited to the detection of the velocity vector by the velocity meter, for example, the position estimation by imaging, the positioning by the illuminator (radar, laser, ultrasonic wave), the accelerometer or the angle. Any means may be used for measuring the movement of a moving object, such as integration by an accelerometer or positioning of an absolute position similar to GPS (Global Positioning System). The first mobile body 1 may include a storage device for storing the observed state quantity of the first mobile body 1.

The state quantity of the first mobile body 1 observed by the above means is information for the first mobile body 1 to move autonomously. The state quantity of the first moving body 1 is, for example, a velocity vector of the first moving body 1 in a predetermined space. The state quantity of the first moving body 1 is not limited to the velocity vector, and is, for example, a position vector, an acceleration vector, an angular velocity vector, an attitude angle, or other state quantities related to motion similar thereto in a predetermined space. You may.

The first mobile body 1 has a functional configuration shown in FIG. 3 in addition to the above hardware functions. The functional configuration of the first mobile body 1 shown in FIG. 3 will be described. The first mobile body 1 in FIG. 3 includes an information acquisition unit 81, a cover boundary detection unit 82 which is a cover boundary acquisition unit, a network calculation unit 83, a graph calculation unit 84, a cover control calculation unit 85, and a mobile body. It includes a control unit 86.

<Information acquisition unit 81>
The information acquisition unit 81 acquires the state quantities of the plurality of second mobile bodies 2. The information acquisition unit 81 is composed of, for example, one of a communication device and a detector forming the second mobile body 2 and the network 4, and an arithmetic unit.

The information acquisition unit 81 includes a transmission / reception unit 81a including a transmitter / receiver that transmits the state amount of the first mobile body 1 to the second mobile body 2 and receives the state amount of the second mobile body 2 from the second mobile body 2. ..

The transmission / reception unit 81a is, for example, a communication means using radio waves. The transmission / reception unit 81a is not limited to the communication means using radio waves, and may be, for example, a communication means for imparting information to light, sound, or a carrier wave similar thereto. Further, the transmission / reception unit 81a may be a communication means that converts information into a light amount distribution by an image output device, a lighting device, or a similar device, and restores the information from the light amount distribution by an imaging device. The transmission / reception unit 81a may be any kind of means as long as it is a means for transmitting and receiving information.

In the transmission / reception unit 81a, when the state amount cannot be acquired from one of the second mobile bodies 2 of the plurality of second mobile bodies 2 for a certain period of time, or when the mobile body among the plurality of second mobile bodies 2 is autonomously distributed and controlled. If it is determined that the function of the system is not satisfied, the one second mobile body 2 may be treated as a third mobile body 3 instead of the second mobile body 2. That is, the first mobile body 1 and the one second mobile body may not be connected by a network.

On the other hand, the transmission / reception unit 81a was the third mobile body 3 at the initial time, but after that, when the state quantity can be acquired from the third mobile body 3, the third mobile body 3 is used. It may be treated as the second moving body 2.

The information acquisition unit 81 includes a measurement unit 81b for observing the state quantity of the second mobile body 2.

The measuring unit 81b is a distance direction measuring means including, for example, a detector such as a stereo camera and an arithmetic unit for observing the relative position of the second moving body 2 from the detection result of the detector. The measuring unit 81b is not limited to the distance directional measuring means using a stereo camera. For example, the measuring unit 81b emits an electromagnetic wave, an ultrasonic wave, or a carrier wave similar thereto to the second moving body 2 or its surroundings, and measures the reflected wave. It may be a measuring means. The measuring unit 81b may be any kind of means as long as it is a means for observing the state quantity.

The information acquisition unit 81 of FIG. 3 includes both a transmission / reception unit 81a capable of acquiring the state quantity of the second mobile body 2 and a measurement unit 81b capable of acquiring the state quantity of the second mobile body 2. However, the configuration of the information acquisition unit 81 is not limited to this. For example, if the information acquisition unit 81 can acquire the state quantity of the second moving body 2 required for generating the control command value from the transmission / reception unit 81a, the information acquisition unit 81 may not be provided with the measurement unit 81b, or may be provided from the measurement unit 81b. If the state quantity of the second moving body 2 required for generating the control command value can be acquired, the transmission / reception unit 81a may not be provided.

In the first embodiment, the state quantities of the plurality of second mobile bodies 2 combine the positions and velocities of the plurality of second mobile bodies 2 with respect to the covering region 5 and the plurality of second mobile bodies 2. It may include the information of the network 4 that is in operation. Then, the information acquisition unit 81 detects the relative position and relative speed of the second moving body 2 with respect to the first moving body 1, and includes it in the state quantity of the second moving body 2 based on the relative position and the relative speed. The position and speed of the second mobile body 2 are estimated, and the position and speed of the second mobile body 2 included in the state quantity of the second mobile body 2 are received from the second mobile body 2 via the network 4. You may do at least one of the things you do.

<Covering boundary detection unit 82>
The covering boundary detection unit 82 detects the covering boundary 6 that closes the covering region 5, and performs a mapping that generates a virtual moving body 7 on the opposite side of the covering region 5 from the first moving body 1. The covering boundary detection unit 82 maps the state of the virtual moving body 7 based on, for example, the geometric shape of the covering boundary 6 and the state quantity of the first moving body 1 acquired by the first moving body 1. Generate a quantity. The covering boundary detection unit 82 is included in, for example, a device for autonomously moving the first mobile body 1.

The covering boundary 6 may be, for example, an obstacle similar to a wall or the like. The covering boundary detection unit 82 is, for example, a means for measuring the distance and orientation with respect to the same surface, and detects the measured geometric shape of the obstacle as the geometric shape of the covering boundary 6. The covering boundary detection unit 82 is not limited to the means for measuring the distance direction of the obstacle, and may, for example, estimate the geometric shape of the covering boundary 6 by state estimation by distance measurement.

Further, the covering boundary 6 may be, for example, a sign similar to a white line or the like. The covering boundary detection unit 82 extracts a sign included in the environmental information obtained by, for example, imaging, and detects the geometric shape of the sign as the geometric shape of the covering boundary. The covering boundary detection unit 82 is not limited to the means for extracting the label by imaging, and may estimate the geometric shape of the covering boundary 6 by state estimation based on the label posted in the environment.

The covering boundary detection unit 82 generates, for example, a virtual moving body 7 having line symmetry with the first moving body 1 by using a plurality of line segment elements that divide the geometric shape of the covering boundary 6 as axes of symmetry. .. The line segment element is derived, for example, by performing an operation on the covering boundary 6 using a predetermined logical formula. The covering boundary detection unit 82 may determine whether the first moving body 1 is located inside or outside the covering region 5 from the state quantity of the first moving body 1. It should be noted that this determination does not have to be performed by the covering boundary detection unit 82. For example, a device that manages a group of moving bodies, which is installed separately from the first moving body 1, detects the first moving body 1 and determines whether it is located inside or outside the covering area 5, and determines whether the result is inside or outside. It may be transmitted to the transmission / reception unit 81a or the measurement unit 81b of the first mobile body 1.

When the state quantity of the virtual moving body 7 is generated so as to be line-symmetric with the state quantity of the first moving body 1 with the line segment element of the covering boundary 6 as the axis of symmetry, the first moving body 1 and the virtual moving body 7 The perpendicular bisector with and coincides with the line segment element.

<Network calculation unit 83>
The network calculation unit 83 includes the state amount of the first mobile body 1 acquired by the first mobile body 1, the state amount of the plurality of second mobile bodies 2 acquired by the information acquisition unit 81, and the covering boundary detection unit 82. Receives the state quantity of the virtual mobile body 7 generated in. The network calculation unit 83 is included in, for example, a device for autonomously moving the first mobile body 1. The network calculation unit 83 may include a storage device for storing the received state quantity.

The network calculation unit 83 includes the first mobile body 1 and the plurality of first moving bodies 1 based on the state quantities of the first mobile body 1, the state quantities of the plurality of second mobile bodies 2, and the state quantities of the virtual mobile body 7. 2 The network that connects the mobile body 2 and the virtual mobile body 7 is obtained. That is, the network calculation unit 83 includes the virtual mobile body 7 in the network of the first mobile body 1 assuming that the virtual mobile body 7 has an information connection with the first mobile body 1 and the plurality of second mobile bodies 2 and is adjacent to the virtual mobile body 7.

In the network calculation unit 83, for example, based on the position information included in the state amount of the first mobile body 1 and the position information included in the state amount of one non-virtual mobile body, the one mobile body is the first. 1 Estimate whether or not the state quantity of the moving body 1 is acquired. Then, when the network calculation unit 83 estimates that the one mobile body has acquired the state quantity of the first mobile body 1, the network calculation unit 83 has the information connection between the mobile body and the first mobile body 1. 2 It is determined that it is a moving body 2.

Further, the network calculation unit 83 is, for example, one of the networks of the mobile group including the first mobile 1, the second mobile 2, and the third mobile 3 from the identifier included in the state quantity of the second mobile 2. The part is estimated to estimate the structure of the network in the set of local mobiles included in the mobile group. The structure of the network is represented by a set of a plurality of moving bodies (nodes) having information connections and adjacent states (branches) in which information is connected between the moving bodies. Similarly, the network calculation unit 83 further includes the virtual mobile 7 in the network structure of the first mobile 1.

Further, the network calculation unit 83 stores, for example, the network information of the first mobile body 1 in a specific buffer and includes it in the state quantity of the first mobile body 1. At this time, the network calculation unit 83 may include the state quantity of the virtual mobile body 7 in the network information of the first mobile body 1.

Further, for example, when the transmission / reception unit 81a acquires the information of the network of the second mobile body 2, the network calculation unit 83 performs an operation of superimposing the network of the first mobile body 1 and the network of the second mobile body 2. .. By this calculation, the network calculation unit 83 can obtain a network including a set of several mobile bodies having information connections in the mobile body group, that is, a set of mobile bodies in a locally adjacent state.

<Graph calculation unit 84>
The graph calculation unit 84 generates a control graph for generating a control command value from the network obtained by the network calculation unit 83. A graph such as a control graph is a set element of a plurality of nodes and a plurality of branches representing a network, and is a figure in which a node (vertex) and a branch (side) are combined. The graph calculation unit 84 is included in, for example, a controller for autonomously moving the first mobile body 1. The graph calculation unit 84 classifies the attribute information for each graph based on the graph theory and holds it as graph information.

The graph calculation unit 84 obtains, for example, a set of combinations of nodes and branches in which two arbitrary moving bodies have information connections with each other and are adjacent to each other as a complete graph from the network obtained by the network calculation unit 83. The graph calculation unit 84 obtains a plurality of complete graphs as a graph set of complete graphs by comprehensively combining nodes and branches. The graph calculation unit 84 classifies a plurality of complete graphs according to the number of nodes of the complete graph (hereinafter, also referred to as “orders”), and generates a control command value for covering control from the plurality of classified complete graphs. Select the control graph for.

An example in which the graph calculation unit 84 generates a control graph by performing Delaunay triangulation division using the operation of the Delaunay graph based on the network obtained by the network calculation unit 83 will be described below.

First, the graph calculation unit 84 includes the first mobile body 1 from the first mobile body 1, the plurality of second mobile bodies 2, and the virtual mobile body 7 based on the network obtained by the network calculation unit 83. , Find a plurality of triangular complete graphs with a digit of 3. That is, the graph calculation unit 84 obtains a set of complete graphs including the first mobile body 1 and having three mobile bodies having information connections as three nodes.

Then, the graph calculation unit 84 selects the control graph from the obtained plurality of triangular complete graphs. For example, the graph calculation unit 84 determines whether or not the nodes of another complete graph are included inside the circumscribed circle of the triangular complete graph, and a plurality of complete graphs in which the nodes of the other complete graph are included in the circumscribed circle. Remove from the complete graph of and select the remaining complete graph as the control graph. When a plurality of control graphs are selected by this, the plurality of control graphs become a triangular graph in which the branches do not intersect each other.

FIG. 4 is a diagram showing a control graph 8 obtained by the above calculation. FIG. 4 shows, as an example of the control graph 8, a control graph 8 including a first moving body 1, a second moving body 23, and a virtual moving body 7. The control graph 8 obtained in the above example is equivalent to the above-mentioned drone graph with respect to the first moving body 1, the plurality of second moving bodies 2, and the virtual moving body 7. When all the moving bodies included in the plurality of second moving bodies 2 are closer to the first moving body 1 than the third moving body 3, the control graph obtained in the above example shows the first moving body 1, Not only the plurality of second mobile bodies 2 and the virtual mobile body 7 but also the third mobile body 3 is equivalent to the above-mentioned drone graph.

<Cover control calculation unit 85>
The covering control calculation unit 85 generates a control command value based on the control graph generated by the graph calculation unit 84. As a result, the covering control calculation unit 85 controls the first mobile body 1 on the assumption that the first mobile body 1, the plurality of second mobile bodies 2, and the virtual mobile body 7 are connected by a network. It is possible to generate a control command value indicating. The cover control calculation unit 85 is included in, for example, a device for autonomously moving the first mobile body 1.

First, from the control graph 8 in which the covering control calculation unit 85 includes the first moving body 1 and the plurality of second moving bodies 2 but does not include the virtual moving body 7, the control command value indicating the control of the first moving body 1 is shown. Will be described in the case of generating. In this case, the covering control calculation unit 85 obtains a control command value using, for example, a gradient function including the state quantity of the first mobile body 1 and the state quantity of the plurality of second mobile bodies 2.

In the gradient function, for example, as shown in the following equation (1), the first moving body 1 and the plurality of second moving bodies 1 have a closed region defined based on the first moving body 1 and the plurality of second moving bodies 2. A function of the sum of the squares of the Euclidean norm with an arbitrary point located closer to the first mobile 1 than the two mobile 2 is used.

The gradient function is a function capable of determining whether a point on a closed region based on the first mobile body 1 and the plurality of second mobile bodies 2 is closer to the first mobile body 1 or the plurality of second mobile bodies 2. Is. By moving toward the stop point which is the extreme value point of the gradient function, the first moving body 1 can cover any of the plurality of second moving bodies 2 without colliding with each other.

Hereinafter, among the closed regions based on the first mobile body 1 and the plurality of second mobile bodies 2, the region based on the first mobile body 1 and the two second mobile bodies 2 constituting one control graph 8 is referred to as ". Partially closed area ". The covering control calculation unit 85 determines, for example, a polygonal region having a triangular outer center of the control graph 8 and the first moving body 1 as vertices as a partially closed region for each control graph 8.

In the case of this example, the covering control calculation unit 85 starts from the node of the control graph 8 to the first moving body 1, the midpoint between the first moving body 1 and the second moving body 21, and the first moving body 1. A partially closed region surrounded by the midpoint between the second moving body 22 and the outer center of the triangular control graph 8 is determined for each control graph 8. Further, in the case of this example, the point included in the partially closed region is the first point in the triangular control graph 8 in which the number of digits having the first moving body 1 and the two second moving bodies 2 as nodes is three. It is equivalent to the point included in the area surrounded by the two sides including the moving body 1 and the two vertical bisectors of the two sides. Then, the cover control calculation unit 85 obtains the stop point of the gradient function from the gradient function showing the union of the squares of the Euclidean norms of arbitrary points included in the partially closed region of the plurality of control graphs 8.

As described above, the cover control calculation unit 85 has the gradient from the gradient function which is the sum of the functions whose elements are the set of the control graph 8 including the plurality of second mobile bodies 2 which are connected and adjacent to each other. When finding the stop point of a function, it is possible to realize a globally convergent and stable cover.

Here, when the gradient function is a function that is the union of the squares of the Euclidean norms of arbitrary points in a closed region consisting of a plurality of partially closed regions, the stop point of the gradient function is equivalent to the center of gravity of the closed region. is there. Accordingly, coating control calculation unit 85, the center of gravity g ⁱ _k of the partial closed area obtained for each control graph 8, the area S ⁱ _k of the partial closed area, by applying the following equation (2) , The center of gravity cent (C _i (x)) of the entire closed area, that is, the stop point of the gradient function can be obtained.

The covering control calculation unit 85 obtains a control command value by applying the following equation (3) to the center of gravity of the entire closed region, that is, the stop point of the gradient function. Note that u _i of the following equation (3), a control input of a first control of the moving body 1, a control for moving the first moving body 1 in the direction of the position of the centroid of the formula (2) Corresponds to the command value. Further, the mass of the following equation (3) corresponds to the mass when the closed region is assumed to be a rigid body. The first moving body 1 evenly covers the closed region according to the control command value represented by the following equation (3).

The covering boundary of the first moving body 1 is equivalent to the boundary of the closed region of the gradient function, and when the control graph 8 is triangular, the covering region of the first moving body 1 is equivalent to the boronoy region. The covering boundary of the first moving body 1 is equivalent to the boundary of the boronoy region.

The gradient function may be, for example, a function obtained by multiplying the Euclidean norm by a weight. Weights Specifically, the coating control calculation unit 85, the center of gravity g ⁱ _k of the partial closed area obtained for each control graph 8, the area S ⁱ _k of the partial closed area, which is assigned to the partial closed area By _{applying Ni} and to the following equation (4), the stop point of the gradient function may be obtained. The control command value at this time is expressed in the same manner as in the above equation (3).

As described above, the control command value indicating the control of the first moving body 1 from the control graph 8 in which the covering control calculation unit 85 includes the first moving body 1 and the plurality of second moving bodies 2 but does not include the virtual moving body 7. The case of generating is described. The covering control calculation unit 85 generates a control command value indicating control of the first moving body 1 from the control graph 8 including the first moving body 1, the plurality of second moving bodies 2, and the virtual moving body 7. In this case, the virtual moving body 7 may be treated in the same manner as any one of the plurality of second moving bodies 2. When the virtual moving body 7 is included in the control graph 8, the boundary of the closed region described above coincides with the vertical bisector between the first moving body 1 and the virtual moving body 7.

<Mobile control unit 86>
The moving body control unit 86 controls the movement of the first moving body 1 based on the control command value generated by the covering control calculation unit 85. The mobile body control unit 86 is included in, for example, a device for autonomously moving the first mobile body 1. The mobile body control unit 86 is, for example, a feedback control system for the drive source of the first mobile body 1. The mobile control unit 86 may be of any type as long as it is a control system that controls the motion of the first mobile body 1 with the control command value of the cover control calculation unit 85 as a target.

<Summary of Embodiment 1>
According to the autonomous distributed control system according to the first embodiment, even if the second mobile body 2 dynamically changes with respect to the first mobile body 1, the condition that the global convergence stability is guaranteed is guaranteed. , The virtual moving body 7 generated by the covering boundary detection unit 82 and the second moving body 2 are treated as equivalent. Therefore, since the first moving body 1 can treat the dynamic change of the covering boundary 6 as the change of the state quantity of the virtual moving body 7 which is equivalent to the second moving body 2, the global convergence stability is achieved. Can satisfy the sex. That is, the mobile autonomous distributed control system according to the first embodiment can satisfy the global convergence stability even if the covering region 5 changes dynamically.

Further, in the mobile autonomous distributed control system according to the first embodiment, the means for forming the network 4 shown in FIG. 4 can be constructed by a simple mechanism. Therefore, it is possible to realize an autonomous distributed control system in which the above effects can be obtained with a relatively simple device.

Further, since the mobile autonomous distributed control system according to the first embodiment includes a graph calculation unit 84 that generates a control graph from the network, it is global as described in Non-Patent Document 2. The necessary and sufficient condition of the control command value satisfying the convergence stability can be satisfied.

The gradient function of the cover control calculation unit 85 is not limited to the function represented by the above equation (2) as long as it is the sum of the functions based on the control graph. The gradient function of the cover control calculation unit 85 is, for example, a first moving body 1 and an arbitrary point located closer to the first moving body 1 than the plurality of second moving bodies 2 with respect to the closed region included in the control graph. It may be the union of the Euclidean norms of. Further, when the gradient function used in the covering control calculation unit 85 is a function obtained by multiplying the Euclidean norm by a weight, the distribution of the moving body group can be arbitrarily designed, so that the moving body group is locally concentrated at a certain position. In addition to the even distribution, the coating can be applied with a distribution suitable for various situations.

Next, the social significance of the autonomous distributed control system according to the first embodiment will be described. Covering with a group of mobiles is used, for example, in space sensing applications. The demand for sensing and the improvement in coverage are increasing due to the use of monitoring people and objects in the city and indoors against various terrorist attacks.

On the other hand, autonomously controlled mobile objects are becoming widespread due to the rise in autonomous driving technology for robots and automatic driving technology for automobiles. Traditionally, mobiles have been equipped with independent autonomous controls for each purpose, but as the area of application expands, the demand for collaborative work or similar advanced work is increasing.

From the above, covering with a mobile group is an important technology in which social demand and technical demand match. However, the coating of a conventional mobile group satisfies the global convergence stability without dynamically changing the covering region, or the covering against a dynamic change of the covering region although it does not satisfy the convergence stability. Either the method of locally performing the control of giving the command for correction by performing the operation independent of the operation of the control command value was selected.

However, if the covering area is treated statically, it becomes a low-performance system in which the robustness of the covering of the mobile group is impaired. On the other hand, in the conventional calculation independent of the calculation of the control command value of the covering, a big problem may occur if the global convergence stability is not satisfied. For this reason, it has been difficult to put the conventional mobile group covering into practical use widely because it satisfies only either robustness against dynamic change of the covering region and global convergence stability. On the other hand, the autonomous distributed control system according to the first embodiment can satisfy both robustness against dynamic changes in the covering boundary and global convergence stability, and thus contributes widely to society. Can be done.

<Embodiment 2>
FIG. 5 is a diagram schematically showing a configuration of an autonomous distributed control system according to a second embodiment of the present invention. Hereinafter, among the components according to the second embodiment, the components that are the same as or similar to the above-mentioned components are designated by the same or similar reference numerals, and different components will be mainly described.

The configuration of the autonomous distributed control system according to the second embodiment is the same as the configuration in which the covering boundary control unit 10 is added to the configuration of the autonomous distributed control system (FIG. 1) according to the first embodiment. The covering boundary control unit 10 virtually changes the geometric shape of the covering boundary 6 while maintaining global convergence stability, and transmits the change information to the first mobile body 1. In FIG. 5, the covering boundary control unit 10 is provided outside the first moving body 1, but the present invention is not limited to this.

The covering boundary detection unit 82 according to the second embodiment updates the virtual moving body 7 by performing the mapping described in the first embodiment with respect to the covering boundary 6 changed by the covering boundary control unit 10. The update referred to here includes at least one of deleting the virtual moving body 7 and newly generating the virtual moving body 7. The network calculation unit 83, the graph calculation unit 84, the covering control calculation unit 85, and the moving body control unit 86 perform the operations described in the first embodiment with respect to the newly generated virtual moving body 7.

Next, the details of the autonomous distributed control system according to the second embodiment will be described. FIG. 6 is a block diagram showing a functional configuration of the first mobile body 1 according to the second embodiment. As shown in FIG. 6, the covering boundary control unit 10 of FIG. 5 transmits change information of the covering boundary 6 to the covering boundary detecting unit 82 of the first mobile body 1.

Before explaining the functional configuration of the first mobile body 1, the covering boundary control unit 10 briefly described above will be described in detail.

The covering boundary control unit 10 is provided in, for example, an arithmetic unit that manages the covering region 5 of the mobile group. The covering boundary control unit 10 uses, for example, a calculation unit for virtually changing the covering area 5 and a moving body located in the vicinity of the covering boundary 6 of the changed covering area 5 as the first moving body 1. The cover boundary detection unit 82 of the first moving body 1 has a communication unit that transmits change information of the cover boundary 6. At least a part of the calculation unit and the communication unit of the covering boundary control unit 10 may be provided inside the first mobile body 1.

The communication unit of the covering boundary control unit 10 detects information related to the change operation for the covering boundary 6 by using a sign recognition device that recognizes a traffic sign or the like as the covering boundary 6 by imaging or the like, and based on the information. The change information of the covering boundary 6 may be acquired. Further, the communication unit of the covering boundary control unit 10 detects information related to the change operation for the covering boundary 6 by, for example, determining the internal processing using a predetermined logical formula, and the covering boundary is based on the information. The change information of 6 may be acquired.

The calculation unit of the covering boundary control unit 10 translates a part of the covering boundary 6 in the normal direction, for example, in order to translate the first moving body 1 in the normal direction of a part of the covering boundary 6. The moved virtual covering element is given to the covering boundary detection unit 82 of the first moving body 1.

FIG. 7 is a diagram showing a principle of giving an operation of translating the first moving body 1 in the normal direction of a part of the covering boundary 6 by the covering boundary control unit 10. For example, the covering boundary detection unit 82 performs a virtual translation of a line segment element that is a part of the covering boundary 6 from the virtual covering element given by the covering boundary control unit 10, that is, the change information of the covering boundary 6. Ask. By such an operation of the covering boundary detecting unit 82, the covering boundary control unit 10 substantially moves the line segment element which is a part of the geometric shape of the covering boundary 6 to the inside or the outside of the covering region 5. , The geometric shape of the covering boundary 6 can be changed.

Hereinafter, the line segment element that is a part of the covering boundary 6 before the virtual movement is referred to as the line segment element 16a before the virtual movement, and the line segment element that is a part of the covering boundary 6 after the virtual movement is referred to. , It is described as a line segment element 16b after virtual movement. Note that FIG. 7 shows a virtual moving body 7a generated based on the line segment element 16a before the virtual moving and the state quantity of the first moving body 1.

When the covering boundary control unit 10 acquires the change information of the covering boundary 6, a new virtual moving body 7b is generated at the position of the inner or outer dividing point between the first moving body 1 and the virtual moving body 7a. , Gives an instruction to the covering boundary detection unit 82 of the first moving body 1.

Upon receiving an instruction from the covering boundary control unit 10, the covering boundary detecting unit 82 removes the virtual moving body 7a and performs mapping based on the line segment element 16b after the virtual moving and the state quantity of the first moving body 1. , Generates a new state quantity of the virtual moving body 7b. At this time, the covering boundary detection unit 82 virtually moves so that the vertical bisectors of the first moving body 1 and the virtual moving body 7b coincide with the covering boundary 6 (line segment element 16b) after the virtual movement. The position of the body 7b (the position of the inner or outer division point) is calculated.

The covering boundary control unit 10 performs the same operation as the above operation for all the moving bodies that detect the covering boundary 6. The first mobile body 1 may hold the input of the first mobile body 1 to the covering boundary detection unit 82 as the state quantity of the first mobile body 1 in the buffer, or may transmit the input to the second mobile body 2. Good.

From the above, all the moving bodies that detect the covering boundary 6 recognize the covering boundary 6 as the covering boundary 6 after the virtual movement and perform coating. Here, since the moving body that does not detect the covering boundary 6 does not receive the command of the covering boundary control unit 10 or is not affected by the command, the global convergence stability is satisfied.

The configuration in which the covering boundary control unit 10 substantially moves the line segment element, which is a part of the geometric shape of the covering boundary 6, to the inside or the outside of the covering region 5 has been described above. The covering boundary control unit 10 performs an operation of increasing or decreasing the covering boundary 6 of the covering region 5 so as to avoid the failure when a failure that the moving body needs to avoid occurs in the vicinity of the covering boundary 6, for example. May be given. This case will be described below.

The covering boundary control unit 10 transmits, for example, information on a line segment element to be added to the covering boundary 6 or a line segment element to be deleted from the covering boundary 6 to the covering boundary detecting unit 82. That is, the covering boundary control unit 10 changes the geometric shape of the covering boundary 6 by adding or deleting a plurality of line segment elements that divide the geometric shape of the covering boundary 6. The covering boundary detection unit 82 updates the virtual moving body 7 by mapping the covering boundary 6 changed by the covering boundary control unit 10 based on the information transmitted from the covering boundary control unit 10. As a result, the first mobile body 1 recognizes the new covering boundary 6.

FIG. 8 is a diagram showing a principle of giving an operation of increasing the covering boundary 6 of the covering region 5. The first moving body 1a corresponds to the second moving body 2 when viewed from the first moving body 1b, and the first moving body 1b corresponds to the second moving body 2 when viewed from the first moving body 1a. The first moving body 1a generates a virtual moving body 7c with respect to the covering boundary 6, and the first moving body 1b generates a virtual moving body 7d with respect to the covering boundary 6.

The covering boundary control unit 10 newly forms a triangle surrounding the obstacle with a part of the covering boundary 6 as an edge in order to avoid a certain obstacle. In the example of FIG. 8, as this triangle, a triangle composed of a two-dot chain line and

covering boundaries

6a and 6b added as line segment elements is shown. The first moving body 1a is close to the covering boundary 6a, and the first moving body 1b is close to the covering boundary 6b.

The covering boundary control unit 10 transmits information on the covering boundary 6a, which is a line segment element close to the first moving body 1a, to the covering boundary detecting unit 82 of the first moving body 1a. The covering boundary detection unit 82 of the first moving body 1a newly generates a virtual moving body 7e from the covering boundary 6a. Since the first moving body 1a stays inside the area surrounded by the covering boundary 6 excluding the alternate long and short dash line which is the reference of the virtual moving body 7c and the covering boundary 6a which is the reference of the virtual moving body 7e, the first moving body 1a moves. The body 1a covers a new covering region 5 having a covering boundary 6 excluding the alternate long and short dash line and a covering boundary 6a as a boundary.

Similarly, the covering boundary control unit 10 transmits information on the covering boundary 6b, which is a line segment element close to the first moving body 1b, to the covering boundary detecting unit 82 of the first moving body 1b. The covering boundary detection unit 82 of the first moving body 1a newly generates a virtual moving body 7e from the covering boundary 6a. Since the first moving body 1b stays inside the area surrounded by the covering boundary 6 excluding the alternate long and short dash line which is the reference of the virtual moving body 7d and the covering boundary 6b which is the reference of the virtual moving body 7e, the first moving body 1b moves. The body 1b covers a new covering region 5 having a covering boundary 6 excluding the alternate long and short dash line and a covering boundary 6b as a boundary.

Since the first moving body 1a is surrounded by the perpendicular bisector with the first moving body 1b, the covering boundary 6, and the covering boundary 6a, it does not invade the inside of the triangle surrounding the obstacle. Similarly, since the first mobile body 1b is surrounded by the perpendicular bisector with the first mobile body 1a, the covering boundary 6 and the covering boundary 6b, it does not invade the inside of the triangle surrounding the obstacle. ..

The calculation unit of the covering boundary control unit 10 assumes, for example, a virtual moving body in order to increase the covering boundary 6, and has a first moving body 1b and a first moving body 1b that are close to both the first moving body 1a and the covering boundary 6. A vertical bisector may be derived for the moving body 1a. Then, the calculation unit of the covering boundary control unit 10 may use the line segments cut by the two vertical bisectors of the covering boundary 6 as the covering boundary 6a and the covering boundary 6b. Further, the operation of increasing the covering boundary 6 by the covering boundary control unit 10 is not limited to the above.

<Summary of Embodiment 2>
Next, the social significance of the autonomous distributed control system according to the second embodiment will be described. According to the autonomous distributed control system according to the second embodiment, the covering region shared by the mobile group can be changed by locally giving information from the outside or the like. Traditionally, the cover area of a mobile group is static information, and the autonomous distributed control system cannot change the cover area during cover control.

It is assumed that the operation of changing the covering area during the control of the covering is performed as a non-stationary command such as a command for giving an instruction to prohibit intrusion when a problem occurs in a part of the covering area. As an operation for this purpose, either an operation of sharing a non-stationary command as a whole or an operation of locally sharing a non-stationary command in a part where a problem has occurred can be considered.

In an autonomous distributed control system that shares non-stationary commands as a whole, it takes a relatively long time for the commands to be reflected in the covering because it is necessary to re-share the new covering area as a whole. Further, in the autonomous distributed control system, it is necessary to stop the control of all the moving bodies until the information is shared. Further, a new function of determining whether or not the information is shared is required. When these are satisfied, the autonomous distributed control system is satisfied with global convergence stability for unsteady commands.

On the other hand, in an autonomous distributed control system that locally shares non-stationary commands, the covering area is corrected only for moving objects in the vicinity of the covering boundary that require a change in the covering area. The command can be reflected on the coating in a shorter time than that. Further, in the autonomous distributed control system, the control of the moving body may be stopped locally, so that the covering can be continued as a whole. However, since the mobile group is divided into two or more groups having different covering regions and discontinuity occurs at the interface, the autonomous distributed control system has global convergence stability with respect to unsteady commands. Is not satisfied.

The conventional autonomous distributed control system was one of the two autonomous distributed control systems having a trade-off relationship with each other as described above. For this reason, in conventional autonomous distributed control systems, it is necessary to sacrifice either the time until the command is reflected in the cover or the global convergence stability, depending on the problem to be applied.

On the other hand, in the autonomous distributed control system according to the second embodiment, since the unsteady command is locally shared, the time until the command is reflected in the covering can be shortened, and the virtual movement can be performed. It is possible to prevent the global convergence stability from being impaired by the formation of the body.

In such an autonomous distributed control system according to the second embodiment, since the covering area can be dynamically handled by the covering boundary control unit 10, the observation area changes from moment to moment, for example, for monitoring forest fires. It can be applied to monitor the phenomenon of fire.

Further, in the autonomous distributed control system according to the second embodiment, for example, an autonomous driving vehicle that detects an obstacle generated in a traveling lane transmits obstacle information to surrounding autonomous vehicles, thereby causing an obstacle. It can be applied to systems that share the coverage area modified by. In this case, it is possible to realize safe control in which the robustness is not impaired in the platooning of the autonomous driving vehicle. As described above, according to the second embodiment, it is possible to realize an autonomous distributed control system in which robustness is not impaired by changing the covering region with time, and it is safe in autonomous distributed control of various mobile groups. And reliability can be improved.

<Embodiment 3>
9 and 10 are diagrams schematically showing the operation of the autonomous distributed control system according to the third embodiment of the present invention. The block diagram showing the functional configuration of the first mobile body 1 according to the third embodiment is the same as the block diagram (FIG. 6) showing the functional configuration of the first mobile body 1 according to the second embodiment. Hereinafter, among the components according to the third embodiment, the components that are the same as or similar to the above-mentioned components are designated by the same or similar reference numerals, and different components will be mainly described.

In the third embodiment, the covering boundary control unit 10 changes the geometric shape of the covering boundary 6 by merging a plurality of adjacent covering regions into one covering region. That is, in the autonomous distributed control system according to the third embodiment, it is possible to perform an operation of coupling the mobile group corresponding to the confluence of the covering regions.

Alternatively, in the third embodiment, the covering boundary control unit 10 changes the geometric shape of the covering boundary 6 by branching one covering region into a plurality of adjacent covering regions. That is, in the autonomous distributed control system according to the third embodiment, it is possible to perform an operation of dividing the mobile group corresponding to the branching of the covering region.

FIG. 9 shows a first mobile group surrounded by the covering region 5c and a second mobile group surrounded by the covering region 5d. The first mobile group surrounded by the covering region 5c includes a first mobile 1c and a plurality of second mobiles 2c connected by a network 4c, and has a system similar to the autonomous distributed control system according to the second embodiment. It is configured. The second mobile group surrounded by the covering region 5d includes a first mobile 1d and a plurality of second mobiles 2d connected by a network 4d, and has a system similar to the autonomous distributed control system according to the second embodiment. It is configured. The covering boundary control unit 10 is commonly used in the first mobile group and the second mobile group.

FIG. 10 shows a third mobile group surrounded by a covering region 5e. The third mobile group surrounded by the covering region 5e includes the first mobile group and the second mobile group similar to those in FIG. 9, and some of them are connected by the network 4e to form the second embodiment. It constitutes a system similar to the autonomous distributed control system related to.

Here, the information acquisition unit 81 (FIG. 6) of the first moving

bodies

1c and 1d provides information including a state quantity from the moving body when a new moving body appears in the vicinity of a network that does not include the moving body. By receiving the mobile, it is possible to include the mobile in the network. As an example, in FIG. 9, the first mobile body 1c included in the first mobile body group and the first mobile body 1d included in the second mobile body group form a network via their respective information acquisition units 81. Imagine a case where it can be formed. In this case, the covering boundary control unit 10 operates the covering boundary detection unit 82 (FIG. 6) and the network calculation unit 83 (FIG. 6) of the first

mobile bodies

1c and 1d, respectively, to operate the first mobile unit 1c and FIG. The mobile group and the second mobile group may be combined into one giant third mobile group in FIG. Hereinafter, an example of binding of the mobile group will be described.

In the covering boundary control unit 10, the network calculation unit 83 of the first mobile body 1c or the network calculation unit 83 of the first mobile body 1d determines the first mobile body 1c based on the state quantities of the first

mobile bodies

1c and 1d, respectively. It is determined whether and the first mobile body 1d have a connection of information and can be adjacent to each other. When it is determined that the covering boundary control unit 10 can be adjacent to the covering boundary detecting unit 82 of the first moving body 1c, the covering boundary control unit 10 causes the virtual moving body 7f (FIG. 9) generated on the second moving body group side with respect to the covering boundary detecting unit 82 of the first moving body 1c. Give a command to delete. Further, the covering boundary control unit 10 gives a command to the covering boundary detecting unit 82 of the first moving body 1d to delete the virtual moving body 7g (FIG. 9) generated on the first moving body group side. Then, the covering boundary control unit 10 connects the first mobile body 1c and the first mobile body 1d to the network calculation unit 83 of the first mobile body 1c and the network calculation unit 83 of the first mobile body 1d. Command to form 4e (FIG. 10).

By the above procedure, the covering boundary control unit 10 connects the first mobile body group and the second mobile body group to the third mobile body group. The determination of whether or not to combine the first mobile group and the second mobile group may be performed by the first mobile 1c or the first mobile 1d instead of the covering boundary control unit 10. The third mobile group formed by combining the first mobile group and the second mobile group covers the covering region 5e, which is the union of the covering region 5c and the covering region 5d.

The example of binding of mobile groups has been explained above. Next, it is assumed that the closed region including all the first mobile group and the closed region including all the second mobile group do not intersect with each other. In this case, the covering boundary control unit 10 operates the covering boundary detection unit 82 and the network calculation unit 83 of the first

mobile bodies

1c and 1d to obtain one third mobile body group in FIG. 9 may be divided into a first mobile group and a second mobile group. Hereinafter, an example of division of the mobile group will be described.

The covering boundary control unit 10 determines whether the closed region including all the first mobile group and the closed region including all the second mobile group intersect each other. When it is determined that the covering boundary control unit 10 does not intersect, the first moving body issues a command to delete the network 4e (FIG. 10) connecting the first moving body 1c and the first moving body 1d. A command to add a new covering boundary 6c (FIG. 9) to the network calculation unit 83 of 1c and generate a new virtual moving body 7f using the covering boundary 6c is issued to detect the covering boundary of the first moving body 1c. Give to part 82. Further, the covering boundary control unit 10 gives a command to the network calculation unit 83 of the first mobile body 1d to delete the network 4e (FIG. 10) connecting the first mobile body 1c and the first mobile body 1d. , A new covering boundary 6d (FIG. 9) is added, and a command to generate a new virtual moving body 7g using the covering boundary 6d is given to the covering boundary detection unit 82 of the first moving body 1d.

According to the above procedure, the covering boundary control unit 10 divides the third mobile body group into a first mobile body group and a second mobile body group. The first moving body 1c of the first moving body group transmits the covering boundary 6c as a state quantity to the information acquisition unit 81 of the second moving body 2c, and the covering boundary of the second moving body 2c of the first moving body group. The detection unit 82 may generate the virtual mobile 7f from the state quantity of the covering boundary 6c obtained by the information acquisition unit 81. This also applies to the second mobile group.

<Summary of Embodiment 3>
Next, the social significance of the autonomous distributed control system according to the third embodiment will be described.

In the conventional autonomous distributed control system, it was necessary to stop the autonomous distributed control both when combining the mobile group and when dividing the mobile group. On the other hand, in the autonomous distributed control system according to the third embodiment, it is possible to perform an operation related to combining or separating the covering regions of the regions where the covering regions of different mobile groups overlap. At this time, in the autonomous distributed control system according to the third embodiment, since it is not necessary to transmit the change of the covering area to the entire moving body to be controlled, the autonomous distributed control is performed regardless of the scale of the moving body group or the covering area. It is possible to satisfy the global convergence stability while continuing.

According to Non-Patent Document 1, when covering a space where merging or branching is performed, the covering region is accompanied by a complicated shape such as a recess, so that the global convergence stability may not be satisfied. As a solution, there is a method of combining two mobile groups into one in the case of merging, and a method of dividing one moving body group into two in the case of branching.

However, in the conventional autonomous distributed control system, the covering area cannot be updated during the control of the binding or separation of the mobile group, so that the actual operation is difficult. For this reason, it is not realistic to perform the coating described in Non-Patent Document 1 with an autonomous distributed control system, and it is difficult to deploy the moving body in a complicated shape. On the other hand, in the autonomous distributed control system according to the third embodiment, the mobile body can be deployed in a complicated shape in order to realize the coupling or separation of the mobile body group. For example, it can be expected to be used for quick rescue by deploying a large number of drones in a complicated environment caused by damage to a building due to an earthquake or tsunami.

<Embodiment 4>
FIG. 11 is a block diagram showing a functional configuration of the first mobile body 1 in the autonomous distributed control system according to the fourth embodiment of the present invention. Hereinafter, among the components according to the fourth embodiment, the same or similar components as those described above will be designated by the same or similar reference numerals, and different components will be mainly described.

The cover control calculation unit 85 according to the fourth embodiment is different from the method of the cover control calculation unit 85 described in the first embodiment for a part of the moving body group as long as the global convergence stability is satisfied. It is configured to calculate.

The graph calculation unit 84 according to the fourth embodiment has a plurality of mobile bodies having a plurality of mobile bodies having information connections as a plurality of nodes based on the network obtained by the network calculation unit 83. Find the complete graph. Then, the graph calculation unit 84 obtains a graph having the minimum configuration necessary for the covering control calculation unit 85 to obtain the control command value by a method different from that of the first embodiment from the plurality of complete graphs.

The graph calculation unit 84 may obtain, for example, a plurality of complete graphs having four or more nodes, and select a control graph from the plurality of complete graphs. Further, a complete graph other than the Delaunay triangle may be used as a control graph.

The covering control calculation unit 85 may generate a control command value from the gradient function, for example, using the sum of arbitrary functions having a complete graph included in the control graph, which is the minimum configuration graph, as an argument as a gradient function.

Further, the cover control calculation unit 85 performs an calculation based on a part of the cover boundary 6 detected by the cover boundary detection unit 82, for example, when the virtual moving body 7 is included in the node of the control graph which is the minimum configuration graph. The obtained result may be replaced with the stationary value of the gradient function and thus the control command value. It should be noted that Non-Patent Document 2 describes that global convergence safety is ensured by performing an operation of replacing a gradient function having one of the complete graphs as an element or its stationary value with another complete graph. It is clear from.

Further, for example, when the virtual moving body 7 is included in the node of the control graph which is the minimum configuration graph, the covering control calculation unit 85 performs Voronoi division based on the control graph to obtain the Voronoi region, and obtains the Voronoi region. The control command value may be obtained based on the center of gravity and the area of the region.

The cover control calculation unit 85 may add, for example, a new gradient function having a complete graph as an element to the gradient function provided in advance.

<Summary of Embodiment 4>
According to the fourth embodiment configured as described above, if the covering boundary 6 composed of curves and fine discontinuous lines is approximated to a linear element, the number of virtual moving bodies 7 increases, and for processing. When the load of calculation becomes heavy, a method different from the method described in the first embodiment is performed. With such a configuration, the calculation load can be reduced.

Next, the social significance of the autonomous distributed control system according to the fourth embodiment will be described. The uses and scope of application of an autonomous distributed control system in which a plurality of mobile bodies make autonomous decisions based on local information are not limited to covering, and may be agreed.

For example, with the social spread of autonomous vehicles, which are mobiles equipped with autonomous control, it is expected that efficient and high-speed transportation of goods using platooning on highways will be realized. The platooning is realized by an autonomous distributed control system in which a large number of moving bodies form a network with each other. An autonomous distributed control system that performs platooning agrees to form a formation for a specific shape if the distance between vehicles can be defined in advance, and if the size of the platoon can be defined in advance, multiple movements. A coating is applied to maintain the coverage area so that the body does not collide. That is, an autonomous distributed control system that performs platooning handles both consensus and cover problems, depending on the given situation.

However, in the conventional autonomous distributed control system, the handling of the consensus problem and the handling of the cover problem are mathematically different. Therefore, the conventional autonomous distributed control system cannot handle both the consensus problem and the cover problem.

On the other hand, in the autonomous distributed control system according to the fourth embodiment, the gradient function can be arbitrarily designed by a complete graph. Considering that Non-Patent Document 2 describes that the consensus problem can be handled by the gradient function based on the complete graph, the autonomous distributed control system according to the fourth embodiment has both the consensus problem and the covering problem. Can be handled. In the above, a system for platooning has been described as an example of an autonomous distributed control system that handles both the consensus problem and the cover problem, but the present invention is not limited to this.

In the present invention, each embodiment can be freely combined, and each embodiment can be appropriately modified or omitted within the scope of the invention.

Although the present invention has been described in detail, the above description is exemplary in all embodiments and the present invention is not limited thereto. It is understood that innumerable variations not illustrated can be assumed without departing from the scope of the present invention.

1 1st mobile, 2 2nd mobile, 4 network, 5 covering area, 6 covering boundary, 7 virtual moving body, 8 control graph, 10 covering boundary control unit, 81 information acquisition unit, 82 coating boundary detection unit, 83 Network calculation unit, 84 graph calculation unit, 85 covering control calculation unit, 86 mobile control unit.

Claims

An autonomous distributed control system in which a first mobile body and a plurality of second mobile bodies connected by a network connecting information cover a predetermined covering region by autonomous and decentralized control.
The first moving body is
A covering boundary acquisition unit that performs a mapping to generate a virtual moving body on the opposite side of the first moving body with respect to the covering boundary that closes the covering region.
The first mobile body, the plurality of second mobile bodies connected to the first mobile body by the network, and the virtual mobile body generated by the covering boundary acquisition unit are connected by the network. A covering control calculation unit that generates a command value indicating control of the first mobile body, assuming that the first moving body is controlled.
An autonomous distributed control system including a moving body control unit that controls the movement of the first moving body based on the command value generated by the covering control calculation unit.
The autonomous distributed control system according to claim 1.
The covering boundary acquisition portion is
The state quantity of the virtual moving body is generated by performing the mapping based on the geometric shape of the covering boundary and the state quantity of the first moving body acquired by the first moving body.
The first moving body is
An information acquisition unit that acquires the state quantities of the plurality of second mobile bodies, and
Based on the state amount of the first moving body, the state amount of the plurality of second moving bodies, and the state amount of the virtual moving body, the first moving body, the plurality of second moving bodies, and the like. A network calculation unit that obtains the network that connects the virtual mobile body, and
Further provided with a graph calculation unit for generating a control graph for generating the command value from the network obtained by the network calculation unit.
The covering control calculation unit
An autonomous distributed control system that generates the command value based on the control graph generated by the graph calculation unit.
The autonomous distributed control system according to claim 2.
The state quantity of the plurality of second mobile bodies includes information on the positions and velocities of the plurality of second mobile bodies with reference to the covering region, and the information of the network connecting the plurality of second mobile bodies. Including
The information acquisition unit of the first mobile body
The relative position and relative speed of the second moving body with respect to the first moving body are detected, and the second moving body included in the state quantity of the second moving body based on the relative position and the relative speed. Estimating the position and the velocity, and
An autonomous distributed control system that receives at least one of the position and the speed of the second mobile included in the state quantity of the second mobile from the second mobile via the network. ..
The autonomous distributed control system according to claim 2 or 3.
The covering boundary acquisition portion is
A plurality of line segment elements that divide the geometric shape of the covering boundary are used as axes of symmetry, and the virtual moving body that is line-symmetric with the first moving body is generated by the mapping.
The graph calculation unit
Based on the network obtained by the network calculation unit, any two mobile bodies including the first mobile body from the first mobile body, the plurality of second mobile bodies, and the virtual mobile body. Obtains a plurality of triangular complete graphs having three moving bodies having information connections as three nodes, and performs Dronay triangle division for selecting the control graph from the plurality of triangular complete graphs.
The covering control calculation unit
An autonomous distributed control system that obtains the command value based on the position of the center of gravity of a polygonal region having a triangular outer center and the first moving body as vertices in the control graph.
The autonomous distributed control system according to any one of claims 1 to 4.
A covering boundary control unit for changing the geometric shape of the covering boundary is provided.
The covering boundary acquisition portion is
An autonomous distributed control system that updates the virtual moving body by performing the mapping on the covering boundary changed by the covering boundary control unit.
The autonomous distributed control system according to claim 5.
The covering boundary control unit is
An autonomous distributed control system that modifies the geometry of a coating boundary by moving a portion of the geometry of the coating boundary to the inside or outside of the coating region.
The autonomous distributed control system according to claim 5 or 6.
The covering boundary control unit is
An autonomous distributed control system that changes the geometry of a covering boundary by adding or removing a plurality of line segment elements that divide the geometry of the covering boundary.
The autonomous distributed control system according to any one of claims 5 to 7.
The covering boundary control unit is
An autonomous distributed control system that changes the geometric shape of the coating boundary by merging a plurality of adjacent coating regions into one coating region.
The autonomous distributed control system according to any one of claims 5 to 8.
The covering boundary control unit is
An autonomous distributed control system that changes the geometric shape of the covering boundary by branching one covering region into a plurality of adjacent covering regions.
The autonomous distributed control system according to claim 2 or 3.
The graph calculation unit
Based on the network obtained by the network calculation unit, any two mobile bodies including the first mobile body from the first mobile body, the plurality of second mobile bodies, and the virtual mobile body. Obtains a plurality of triangular complete graphs having three moving bodies having information connections as three nodes, and performs Dronay triangle division for selecting the control graph from the plurality of triangular complete graphs.
When the virtual moving body is included in the node of the control graph, the covering control calculation unit obtains the Voronoi region based on the control graph, and obtains the command value based on the center of gravity and the area of the Voronoi region. Autonomous distributed control system.
The autonomous distributed control system according to claim 2 or 3.
The covering control calculation unit obtains the command value by a different method, and obtains the command value.
The graph calculation unit
Based on the network obtained by the network calculation unit, a plurality of complete graphs in which any two mobile bodies have a plurality of mobile bodies having information connections as a plurality of nodes are obtained, and from the plurality of complete graphs, a plurality of complete graphs are obtained. An autonomous distributed control system in which the covering control calculation unit obtains a graph having the minimum configuration required to obtain the command value by a different method.