Reference movement plan formulation device, reference movement plan formulation system, and reference movement plan formulation method

Abstract

The purpose of the present invention is to provide a technique with which it is possible to classify movement result data in an appropriate manner. This reference movement planning device comprises: a classification-criteria-defining unit for defining one or more classification criteria; a result-data-classifying unit for, on the basis of the classification criteria, classifying movement result data into respective result clusters in which the movement result data is reflected; a representative-route-creating unit for creating a representative route for each result cluster on the basis of the classified movement result data; a standard-route-creating unit for creating a standard route on the basis of feature points of the representative routes; and a standard-route-transmitting unit for transmitting the standard route to a user or the like.

Description

[Technical Field]

[0001] The present disclosure relates to a standard movement plan development device, a standard movement plan development system, and a standard movement plan development method. [Background technology]

[0002] Air traffic control sets standard flight plan routes, and route information showing these routes is published in air traffic control publications. In actual operations, routes that deviate from the standard routes may be used due to congestion or flight restrictions, so the air traffic control system sets reference routes as commonly used flight routes in addition to the standard routes.

[0003] Conventionally, air traffic control systems have been proposed that create movement plans for individual aircraft based on a reference route so as to satisfy the processing capacity of the destination. For example, Patent Document 1 proposes a technique for setting a reference route that reflects flight performance, in which a profile that defines a reference movement plan for each operation classification is generated from flight performance. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2017-212012 Summary of the Invention [Problem to be solved by the invention]

[0005] In conventional technology, a profile, which is a standard movement plan indicating the route and speed to the destination, is defined and stored in advance for each classification criterion of aircraft operation, such as type of aircraft, time (e.g., year, month, day of the week, time zone), and weather (e.g., wind direction, wind speed, visibility, weather). Then, for an aircraft that is the subject of a plan during or before operation, a profile with the same classification criterion as the operation of the aircraft is selected, and an individual movement plan indicating the individual route and speed to the destination is formulated using the selected profile.

[0006] However, if the operational classification is inappropriate, the accuracy of the profile will decrease. For example, if the classification is excessive, the profile will be meaninglessly subdivided, and there will be a problem that a sufficient number of samples will not be obtained to generate an appropriate profile. Conversely, if the classification is insufficient, that is, if profiles that should have been divided are combined, the results of different flight situations will be mixed, and a profile containing incorrect samples will be generated.

[0007] Therefore, the present disclosure has been made in consideration of the above-mentioned problems, and aims to provide a technology that can appropriately classify travel history data. [Means for solving the problem]

[0008] A reference movement plan formulation device according to the present disclosure includes an actual data receiving unit that acquires movement actual data of a moving object, a classification criterion definition unit that defines one or more classification criteria that are standards for classifying the movement actual data, an actual data classification unit that classifies the movement actual data for each actual cluster to which the movement actual data is reflected based on the classification criterion, a representative route creation unit that creates, for each actual cluster, a representative route that is a route that represents the actual cluster based on the movement actual data classified into the actual cluster, a reference route creation unit that creates a reference route based on feature points of the representative route, and a reference route transmission unit that transmits the reference route. an achievement cluster evaluation unit that changes the classification criteria based on the classification result of the travel achievement data; Equipped with. [Effects of the Invention]

[0009] According to the present disclosure, the travel history data is classified for each record cluster to which the travel history data is reflected based on the classification criteria, so that the travel history data can be appropriately classified.

[0010] The objects, features, aspects and advantages of the present disclosure will become more apparent from the following detailed description and the accompanying drawings. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a diagram illustrating an example of the configuration of a mobile object control support system according to a first embodiment. [Figure 2] 1 is a block diagram showing an example of the configuration of a standard movement plan formulation device according to a first embodiment. [Figure 3] FIG. 2 is a diagram illustrating an example of track data. [Figure 4] FIG. 10 is a diagram illustrating an example of situation data. [Figure 5] FIG. 10 is a diagram illustrating an example of classification criteria. [Figure 6] FIG. 10 is a diagram illustrating an example of the distance between track data. [Figure 7] FIG. 10 is a diagram illustrating an example of an area between track data. [Figure 8] 4 is a flowchart showing the operation of the standard movement plan formulation device according to the first embodiment. [Figure 9] 4 is a flowchart showing the operation of the standard movement plan formulation device according to the first embodiment. [Figure 10] FIG. 3 is a diagram for explaining the operation of the standard movement plan formulation device according to the first embodiment. [Figure 11] FIG. 3 is a diagram for explaining the operation of the standard movement plan formulation device according to the first embodiment. [Figure 12] FIG. 3 is a diagram for explaining the operation of the standard movement plan formulation device according to the first embodiment. [Figure 13] FIG. 3 is a diagram for explaining the operation of the standard movement plan formulation device according to the first embodiment. [Figure 14] 4 is a flowchart showing the operation of the standard movement plan formulation device according to the first embodiment. [Figure 15] FIG. 3 is a diagram for explaining the operation of the standard movement plan formulation device according to the first embodiment. [Figure 16] 1 is a diagram illustrating an example of a monitoring device according to a first embodiment. [Figure 17] 4 is a flowchart showing the operation of the standard movement plan formulation device according to the first embodiment. [Figure 18]FIG. 10 is a block diagram showing an example of the configuration of a standard movement plan formulation device according to a second embodiment. [Figure 19] FIG. 10 is a diagram for explaining the operation of the standard movement plan formulation device according to the second embodiment. [Figure 20] FIG. 10 is a diagram for explaining the operation of the standard movement plan formulation device according to the second embodiment. [Figure 21] FIG. 11 is a block diagram showing an example of the configuration of a standard movement plan formulation device according to a third embodiment. [Figure 22] FIG. 10 is a diagram for explaining the operation of the standard movement plan formulation device according to the third embodiment. [Figure 23] FIG. 10 is a diagram for explaining the operation of the standard movement plan formulation device according to the third embodiment. [Figure 24] FIG. 10 is a diagram for explaining the operation of the standard movement plan formulation device according to the third embodiment. [Figure 25] FIG. 10 is a diagram for explaining the operation of the standard movement plan formulation device according to the third embodiment. [Figure 26] FIG. 10 is a diagram for explaining the operation of the standard movement plan formulation device according to the third embodiment. [Figure 27] FIG. 10 is a block diagram showing a hardware configuration of a standard movement plan formulation device according to a modified example. [Figure 28] FIG. 10 is a block diagram showing a hardware configuration of a standard movement plan formulation device according to a modified example. DETAILED DESCRIPTION OF THE INVENTION

[0012] <First Embodiment> In the following description, the moving object is an aircraft monitored and guided by air traffic control, but is not limited to this. The moving object may be a moving object of other types of transportation, such as a ship in maritime traffic, a vehicle in road traffic, or a train in rail traffic.

[0013] FIG. 1 is a diagram showing an example of the configuration of a mobile object control support system including a standard movement plan formulation device 11 according to the first embodiment and a mobile object control system 13. As shown in FIG.

[0014] The reference movement plan formulation device 11 formulates a reference movement plan, which is a standard movement plan for a mobile object. The mobile object control system 13 presents the reference movement plan formulated by the reference movement plan formulation device 11 and the operation status of the mobile object. A mobile object controller checks the operation status of the mobile object using the mobile object control system 13 and controls the mobile object based on the reference movement plan. Here, the reference movement plan formulation device 11 and the mobile object control system 13 are shown as separate servers connected via the network 12, but the present invention is not limited to this. For example, the reference movement plan formulation device 11 and the mobile object control system 13 may be realized on the same server via a database or memory.

[0015] Fig. 2 is a block diagram showing an example of the configuration of the standard movement plan formulation device 11 according to Embodiment 1. The standard movement plan formulation device 11 in Fig. 2 includes an actual data receiving unit 21, a classification criteria defining unit 22, an actual data classifying unit 23, a representative route creating unit 24, a reference route creating unit 25, and a reference route transmitting unit 26.

[0016] The performance data receiving unit 21 in Fig. 2 acquires performance data, which is the movement history data of a mobile object. The performance data includes track data that indicates the movement history of the mobile object. Fig. 3 is a diagram showing an example of the track data. The track data is time-series data obtained by a sensor such as a radar of the mobile object control system 13. The track data in Fig. 3 includes the date 31 and time 32 when the mobile object was observed, a mobile object ID 33 such as a flight number for identifying the mobile object, latitude 34 and longitude 35 that indicate the two-dimensional position of the mobile object, speed 36 and altitude 37 of the mobile object, and model 38 that indicates the type of mobile object.

[0017] Note that the travel history data may include not only flight path data but also situation data that indicates the travel environment at the time of operation. Fig. 4 is a diagram showing an example of situation data. The situation data is, for example, observation data such as weather that affects the aircraft's flight path and the runway used for takeoff and landing. The situation data in Fig. 4 includes date 41 and time 42 when the aircraft was observed, latitude 43 and longitude 44 that indicate the observation point, and weather 45, wind direction 46, wind speed 47, and visibility 48 at the observation point. The date 41, time 42, latitude 43, and longitude 44 in Fig. 4 correspond to the date 31, time 32, latitude 34, and longitude 35 in Fig. 3.

[0018] 2 defines one or more classification criteria that are used to classify the track data of the movement history data. As will be described later, the classification criteria defined by the classification criteria definition unit 22 are used by the history data classification unit 23.

[0019] Fig. 5 is a diagram showing an example of classification criteria. The classification criteria in Fig. 5 include a classification number 51, a classification criteria name 52, the number of classifications 53, a classification content 54, and a valid flag 55. That is, in the first embodiment, the classification criteria include at least one of the similarity between routes traveled by aircraft, an indicator of the aircraft's travel history, the type, the date and time, and an indicator of the aircraft's surrounding conditions.

[0020] The classification content 54 defines the classification method and classification category of the classification criteria. For example, "No. 0002" in FIG. 5 defines that the track data is classified into 12 categories using the value extracted by the formula "Mid(DateTime(), 5, 2)." The formula "Mid(DateTime(), 5, 2)" extracts two characters from the fifth character from the left of DateTime(), which indicates the year, month, and day as an eight-digit number, moving to the right. For example, if the DateTime() for March 1, 2022, is expressed as "20220301," the formula "Mid(DateTime(), 5, 2)" extracts "03," which is the fifth and sixth character from the left of "20220301," i.e., "March." In this case, the formula "Mid(DateTime(), 5, 2)" extracts any one of "01" to "12," so the track data is classified into 12 categories. The classification content 54 is not limited to a calculation formula, and may be defined by a classification category such as thresholds, such as "No. 0003" to "No. 0009."

[0021] The classification content 54 of "No. 0001" indicates that the track data is classified using the K-means method based on the similarity corresponding to the distance or area between the track data. FIG. 6 is a diagram showing an example of the distance between the track data, and FIG. 7 is a diagram showing an example of the area between the track data. These diagrams show two-dimensional images represented by latitude and longitude, but altitude information may also be added to show a three-dimensional image. The following mainly describes the similarity corresponding to the distance between the track data in FIG. 6, but the similarity corresponding to the area between the track data in FIG. 7 is generally similar to the following.

[0022] In the example of Figure 6, the start points of the two track data 61a and 61b match, and the end points of the two track data 61a and 61b match. However, since the start point is the first observation data upon entering the airspace, and the end point is the last observation data before leaving the airspace or the end of the runway, in reality, the start points and end points of the two track data often do not match. In addition, the distance and number of observation points of the two track data often differ. In these cases, the distance between time series data of different lengths can be calculated using dynamic time warping (DTW).

[0023] 6, the distance between track data 61a and track data 61b is calculated by DTW. If track data 61a includes data A having passing points a0, a1, ..., am, and track data 61b includes data B having passing points b0, b1, ..., bn, the distance D(A, B) between track data 61a and track data 61b can be calculated by equation (1).

[0024]

number

[0025] The smaller this distance, the greater the similarity. The area between track data 71a and 71b in Figure 7 corresponds to the distance between track data 61a and 61b in Figure 6 when the number of passing points is infinite, and the smaller the area, the greater the similarity. As described above, the similarity and the distance in Figure 6 or the area in Figure 7 have an inverse correspondence relationship.

[0026] Alternatively, the distance between the start point and end point of the trajectory data may be divided into N equal parts by distance or time, and the nearest waypoint to each of these N equal parts may be calculated. Then, the Euclidean distance may be calculated between the calculated waypoints a1, a2, ..., aN and the calculated waypoints b1, b2, ..., bN, and the sum or average of the N distances may be calculated as the distance between the trajectory data. Furthermore, in calculating this distance, weighting may be applied to the first through Nth waypoints. For example, since a landing aircraft flies toward an airport (runway), waypoints closer to the airport may be weighted more heavily than waypoints farther from the airport in calculating the similarity between the trajectory data.

[0027] The performance data classification unit 23 in FIG. 2 classifies the movement performance data for each performance cluster dynamically reflecting the movement performance data based on one or more classification criteria defined by the classification criteria definition unit 22. For example, if the classification criteria are defined by classification categories such as thresholds of the classification content 54, such as "No. 0002" to "No. 0009" in FIG. 5, the performance data classification unit 23 classifies the corresponding track data according to the defined classification category. Alternatively, if the classification criteria are defined as "No. 0001" in FIG. 5, the performance data classification unit 23 classifies the movement performance data into 10 performance clusters using the K-means method, which is unsupervised learning, based on the similarity between the track data. Unsupervised learning refers to a learning device learning the characteristics of the training data by providing the learning device with training data that does not include results (labels). An example of such learning is described below.

[0028] FIG. 8 is a flowchart showing the operation of the performance data classification unit 23 to classify the track data in accordance with the K-means method based on the similarity between the track data.

[0029] First, in step S1, the performance data classification unit 23 randomly assigns each track data to a performance cluster c (c is 1 to 10). For example, when step S1 is performed for the first time, a default performance cluster is used as the performance cluster for step S1, and when step S1 is performed for the second time or later, the performance cluster determined during the previous processing is used as the performance cluster for step S1.

[0030] In step S2, the performance data classification unit 23 calculates the centroid means(c) of each performance cluster. Below, we will explain the case where multiple track data are assigned to one performance cluster as an example. The centroid means(c) of an performance cluster is one track data among the multiple track data that has the largest total similarity with the other track data excluding that one track data.

[0031] 6 and 7, when the similarity is expressed as a distance or an area, the similarity and the distance or the area have a correspondence relationship in which the magnitudes thereof are inverse to each other, as described above. If it is preferable to make the magnitudes of the similarity and the distance or the area the same, the reciprocal of the distance or the area may be defined as the similarity, for example.

[0032] FIG. 9 is a flowchart showing the operation of step S2 in FIG.

[0033] First, in step S11, the performance data classification unit 23 acquires a list of track data classified into performance cluster c. In step S12, the performance data classification unit 23 initializes the maximum value maxS of similarity and the central means by setting them to 0 and null.

[0034] Before step S13, the performance data classification unit 23 acquires one track data t1 from the track data list. In step S13, the performance data classification unit 23 initializes S(t1), which represents the similarity regarding the one track data t1.

[0035] Before step S14, the performance data classifying unit 23 acquires, from the list of track data, another track data t2 that is not the first track data t1 but is classified into the same performance cluster c as the first track data t1. In step S14, the performance data classifying unit 23 calculates a similarity s(t1, t2) corresponding to the distance or area between the first track data t1 and the second track data t2. In step S15, the performance data classifying unit 23 adds s(t1, t2) to S(t1).

[0036] The performance data classification unit 23 repeatedly performs the above steps S14 and S15 while changing the other track data t2. This repetition is performed until the other track data t2 is set to all of the track data classified into the performance cluster c, except for the one track data t1.

[0037] In step S16, the performance data classification unit 23 determines whether S(t1) is greater than maxS. If S(t1) is greater than maxS, the process proceeds to step S17, and if S(t1) is equal to or less than maxS, step S17 is skipped.

[0038] In step S17, the performance data classifying unit 23 sets maxS=S(t1) and holds one track data t1 as the center means(c) of the performance cluster c.

[0039] The performance data classification unit 23 repeatedly performs the above steps S13 to S17 while changing one track data t1. This repetition is performed until one track data t1 is set for all track data classified into performance cluster c. By the above processing of Fig. 9, the center means(c) of performance cluster c is obtained.

[0040] In step S3 of Fig. 8, the performance data classification unit 23 determines whether the center of the performance cluster has changed. If it is determined that the center of the performance cluster has changed, the operation of Fig. 8 ends, and if it is not determined that the center of the performance cluster has changed, the process proceeds to step S4.

[0041] Before step S4, the performance data classification unit 23 acquires one track data from all the track data "trace." In step S4, the performance data classification unit 23 calculates the similarity between the one track data and the center of each performance cluster. In step S5, the performance data classification unit 23 reassigns the one track data to the performance cluster with the greatest similarity.

[0042] The performance data classifying unit 23 repeatedly performs the above steps S4 and S5 while changing the one track data. The performance data classifying unit 23 changes the one track data until the one track data t1 is set for all track data.

[0043] In step S6, the performance data classification unit 23 determines whether there are no more changes to the performance clusters. If it is determined that there are no more changes to the performance clusters, the operation in Fig. 8 ends, and if it is not determined that there are no more changes to the performance clusters, the process returns to step S2.

[0044] To summarize the above, the performance data receiving unit 21 acquires movement performance data including the first track data and the second track data as learning data. Then, the performance data classifying unit 23 generates a learning model for inferring the feature quantities of performance clusters from the first track data and the second track data based on the learning data. Note that performance clusters and classification criteria can be treated equivalently, and the feature quantities of performance clusters correspond to, for example, the type and threshold of the classification criteria.

[0045] Furthermore, the performance data receiving unit 21 acquires movement performance data including the third track data. Then, the performance data classifying unit 23 classifies the third track data acquired by the performance data receiving unit 21 into performance clusters using a learning model for inferring the feature quantities of the performance clusters from the track data.

[0046] With this configuration, it is possible to reduce excesses and deficiencies in the extraction of flight path patterns. In the above description, the performance data receiving unit 21 performed non-hierarchical clustering using the K-means method, which is an example of unsupervised learning. However, hierarchical clustering, such as a shortest distance method, may also be performed. In the above description, the performance data receiving unit 21 performed unsupervised learning. However, it may also perform supervised learning using representative routes as training data, reinforcement learning using the degree of separation between performance clusters as a reward, or semi-supervised learning. In the above description, the performance data receiving unit 21 performed learning on the track data of the movement performance data. However, it may also perform learning on the track data and situation data of the movement performance data. The first performance cluster and the second performance cluster described below are included in the above performance clusters.

[0047] The representative route creation unit 24 in FIG. 2 creates a representative route, which is a route that represents each of the result clusters, for each result cluster based on the travel result data classified into the result clusters.

[0048] 10, 11, and 12 are diagrams for explaining examples of representative route creation by the representative route creation unit 24. FIG.

[0049] In the example of Fig. 10, the representative route creation unit 24 divides each of the track data 101a, 101b, 101c, and 101d in one achievement cluster into N (four in Fig. 10) pass points. Then, the representative route creation unit 24 creates, as the representative route 102, a route that passes through an average pass point obtained by averaging the corresponding pass points of the track data 101a, 101b, 101c, and 101d.

[0050] For example, the representative route creation unit 24 obtains the first average passing point p1 by averaging passing points a1 to d1 of the track data 101a to 101d, and obtains the second average passing point p2 by averaging passing points a2 to d2 of the track data 101a to 101d. The representative route creation unit 24 performs this calculation until the Nth average passing point pN is obtained. Then, the representative route creation unit 24 creates a route that passes through the first, second, ..., Nth average passing points p1, p2, ..., pN as the representative route 102. Thus, in the example of FIG. 10, the representative route creation unit 24 sets the route obtained by averaging multiple routes indicated by multiple pieces of travel history data in terms of at least one of latitude and longitude, speed, and altitude as the representative route of one history cluster.

[0051] In the example of FIG. 11, the representative route creation unit 24 divides each of the track data 111a, 111b, 111c, and 111d in one achievement cluster into N pass points. An example will be described below where N is 4. The representative route creation unit 24 calculates the total distance of pass point a1 as the sum of the distance between pass point a1 and pass point b1, the distance between pass point a1 and pass point c1, and the distance between pass point a1 and pass point d1. The representative route creation unit 24 calculates the total distance of pass point a2, pass point a3, and pass point a4 in the same way as the total distance of pass point a1, and calculates the sum of the total distances of pass points a1 to a4 as a value corresponding to the similarity of the track data 111a.

[0052] The representative route creation unit 24 calculates values corresponding to the similarity of the track data 111b, track data 111c, and track data 111d, similar to the value corresponding to the similarity of the track data 111a.The representative route creation unit 24 then creates the track data for which the value corresponding to the similarity is smallest, that is, the track data for which the similarity is greatest, as the representative route of one achievement cluster.As such, in the example of FIG. 11, the representative route creation unit 24 creates the route of one track data for which the similarity to the other track data excluding the one track data is largest, as the representative route of one achievement cluster.Note that this representative route is substantially the same as the center of the achievement cluster created by the above-mentioned K-means method.

[0053] In the example of Fig. 11, the representative route creation unit 24 uses the distance between the track data 111a to 111d as a value corresponding to the similarity. In contrast, in the example of Fig. 12, the representative route creation unit 24 uses the area between the track data 121a to 121d as a value corresponding to the similarity. The area of the hatched portion in Fig. 12 represents a value corresponding to the similarity between track data 121a and track data 121c. In other words, the example of Fig. 12 is essentially the same as Fig. 11, in which the number of passing points dividing the track data is set to infinity.

[0054] In the examples of Figures 10 to 12, the starting points of multiple track data match and the ending points of multiple track data match, but if at least one of the starting points or ending points does not match, the starting point and ending point can be treated as passing points and a representative route can be created as described above.

[0055] The reference path creation unit 25 in FIG. 2 extracts feature points from the representative path, and creates a reference path made up of the feature points based on the extracted feature points.

[0056] 13 is a diagram for explaining an example of creating a reference route 141 expressed in latitude and longitude from a representative route 131 expressed in latitude and longitude. FIG. 14 is a flowchart showing the operation of the reference route creating unit 25 to create the reference route 141.

[0057] First, before step S21, the reference route creation unit 25 sets the performance cluster f to 0 and acquires the representative route 131 of the performance cluster f. In step S21, the reference route creation unit 25 divides the acquired representative route 131 into fixed sections t. The criterion for dividing the sections may be distance or time. Dividing the representative route 131 into multiple sections corresponds to extracting a point cloud from the representative route 131.

[0058] In step S22, the reference route creation unit 25 calculates the heading v(0) of the nose of the aircraft at the starting point based on the representative route 131. The heading is expressed as a numerical value, for example, such that true north is 0° (=360°), true east is 90°, true south is 180°, and true west is 270°.

[0059] Before step S23, the reference route creation unit 25 sets the section t to 1. Note that the section t where t=0 corresponds to the start point. In step S23, the reference route creation unit 25 calculates the heading v(t) of the aircraft's nose in the section t based on the representative route 131.

[0060] In step S24, the reference path creation unit 25 determines whether the difference (absolute value) between v(t) and v(t-1) is equal to or greater than a threshold value. If it is determined that the difference is equal to or greater than the threshold value, the process proceeds to step S25, and if it is not determined that the difference is equal to or greater than the threshold value, step S25 is skipped.

[0061] In step S25, the reference trajectory creation unit 25 determines that the orientation of the nose has changed, i.e., that the direction of the aircraft's flight has changed due to a control instruction from an air traffic controller. The reference trajectory creation unit 25 then stores positions that represent the position p(t) in section t and the position p(t-1) in section t-1 as feature points. Here, the position p(t) is stored as the representative position (i.e., feature point), but this is not limiting; for example, the position p(t-1) may be stored, or an intermediate point between the positions p(t) and p(t-1) may be stored.

[0062] The reference path creation unit 25 repeatedly performs the above steps S23 to S25 while incrementing and changing the section t. In step S26, the reference path creation unit 25 connects the stored feature points to create a reference path 141 of the performance cluster f. As examples, Fig. 13 shows a reference path 141 drawn by a solid line using spline interpolation and a reference path 141 drawn by a dashed dot line using linear interpolation.

[0063] The reference path creating unit 25 repeatedly performs the above steps S21 to S26 while incrementing and changing the result cluster f, thereby creating reference paths 141 for the representative paths 131 of all result clusters.

[0064] The above describes a case where the reference route creation unit 25 extracts points where the orientation of the representative route 131 has changed by a threshold or more as feature points, and creates the reference route 141 based on the feature points, but the present invention is not limited to this. For example, the reference route creation unit 25 may extract points where the altitude or speed of the representative route 131 has changed by a threshold or more as feature points, similar to the orientation, and create the reference route 141 based on the feature points.

[0065] 15 is a diagram illustrating an example of creating a reference route expressed in latitude and longitude. The reference route creation unit 25 divides the representative route 151 into sections of unit time and calculates the change in altitude of the aircraft between two consecutive sections. If the difference in altitude change is equal to or greater than a threshold, the reference route creation unit 25 determines that an altitude change instruction has been issued as a control instruction by an air traffic controller for the two sections, and extracts a section that represents the two sections as a feature point. For example, if the difference between altitude difference 152b and altitude difference 152a is equal to or greater than a threshold, the reference route creation unit 25 determines point 153a as a feature point, and if the difference between altitude difference 152c and altitude difference 152b is less than the threshold, it does not determine point 153b as a feature point.

[0066] Feature points may also be found for speed, as with altitude. The reference route creation unit 25 divides the representative route 151 into sections of unit time, and finds the difference in speed of the aircraft between two consecutive sections. If the speed difference is equal to or greater than a threshold, the reference route creation unit 25 determines that a speed change instruction has been issued as a control instruction by an air traffic controller for those two sections, and extracts a section that represents those two sections as a feature point.

[0067] Furthermore, when the distance between two points in the point group extracted as feature points is equal to or less than a threshold, the reference path creation unit 25 may extract either one of the two points or an intermediate point between the two points as a feature point. Furthermore, when the time difference between two points in the point group extracted as feature points is equal to or less than a threshold, the reference path creation unit 25 may extract either one of the two points or an intermediate point between the two points as a feature point.

[0068] By performing the above-described processing, it is possible to extract a small number of feature points that serve as a guide for air traffic controllers to issue control instructions, thereby preventing the reference trajectory from becoming too complicated and reducing the control load on air traffic controllers.

[0069] The reference route creation unit 25 may acquire control instructions to the aircraft from the air traffic controller's voice or keyboard, rather than from the movement history of the aircraft, and extract feature points based on the location or time when the control instructions were issued. The reference route creation unit 25 may also estimate the aircraft that is the target of the control instructions (hereinafter also referred to as "aircraft to be controlled") based on the monitoring of the aircraft by the air traffic controller using a monitoring device. Specifically, the reference route creation unit 25 may estimate the aircraft to be controlled based on the gaze position on the control screen monitored by the air traffic controller. A method for estimating the aircraft to be controlled will be described below.

[0070] 16 is a diagram showing an example of a monitoring device used by an air traffic controller to monitor aircraft. When an air traffic controller 161 performs air traffic control duties using a control screen 162 of the monitoring device, the gaze position of the air traffic controller 161 is measured by a line-of-sight measurement device 163 provided on the control screen 162. Based on the gaze position and the position of the aircraft displayed on the screen, the reference trajectory creation unit 25 estimates the aircraft being gazed at as an aircraft to be controlled.

[0071] FIG. 17 is a flowchart showing the operation of the reference trajectory creation unit 25 to estimate the target aircraft based on the line-of-sight information of the air traffic controller.

[0072] First, in step S31, the reference trajectory creation unit 25 acquires a plurality of control instructions from an air traffic controller.

[0073] Before step S32, the reference trajectory creation unit 25 acquires one control instruction g from the plurality of control instructions. In step S32, the reference trajectory creation unit 25 acquires gaze coordinates corresponding to the gaze position on the gaze measurement device 163 based on the gaze position of the air traffic controller at time t when the one control instruction is issued.

[0074] In step S33, the reference route creation unit 25 converts the line-of-sight coordinates into latitude and longitude p(g, t) based on the airspace information displayed on the control screen 162 at time t. In step S34, the reference route creation unit 25 initializes the minimum distance D between the latitude and longitude p(g, t) and the aircraft displayed on the control screen 162. Here, the minimum distance D is initialized by setting the maximum value Dmax that is allowable as the minimum distance D to the minimum distance D.

[0075] Before step S35, the reference route creation unit 25 acquires one aircraft h from the multiple aircraft displayed on the control screen 162 at time t. In step S35, the reference route creation unit 25 acquires the latitude and longitude p(h, t) of aircraft h. In step S36, the reference route creation unit 25 calculates the distance d between the latitude and longitude p(g, t) of the line of sight coordinates of the air traffic controller 161 at time t and the latitude and longitude p(g, t) of aircraft h.

[0076] In step S37, the reference path creation unit 25 determines whether the distance d is smaller than the minimum distance D. If it is determined that the distance d is smaller than the minimum distance D, the process proceeds to step S38, and if it is not determined that the distance d is smaller than the minimum distance D, steps S38 and S39 are skipped.

[0077] In step S38, the reference path creation unit 25 updates the minimum distance D to the distance d. In step S39, the reference path creation unit 25 estimates and stores that the aircraft targeted by the control instruction g is aircraft h.

[0078] The reference path creation unit 25 repeatedly performs the above steps S35 to S39 while incrementing and changing the aircraft h. As a result, of the multiple aircraft displayed on the control screen 162 at time t, the aircraft whose latitude and longitude are closest to the line-of-sight coordinates is estimated as the aircraft that is the target of the control instruction g. The reference path creation unit 25 repeatedly performs the above steps S32 to S39 while incrementing and changing the control instruction g. As a result, the aircraft that is the target of the control instruction is estimated for each of all control instructions.

[0079] 2 transmits the reference route created by the reference route creation unit 25. The reference route created by the reference movement plan creation device 11 is presented to, for example, the mobile object control system 13 in FIG.

[0080] <Summary of the First Embodiment> According to the reference movement plan formulation device 11 according to the first embodiment as described above, the actual data classification unit 23 classifies the movement actual data for each actual cluster to which the movement actual data is reflected, based on the classification criteria. Therefore, the movement actual of the moving object, such as the operational status, weather conditions, and flight restrictions, is reflected in the actual cluster for classifying the movement actual data, so that the movement actual data can be appropriately classified.

[0081] Furthermore, in this first embodiment, the performance data classification unit 23 classifies the track data into performance clusters based on the similarity of the track data. With this configuration, it is possible to perform meaningful classification of the track data, and therefore it is possible to efficiently create a reference route with reduced excesses and deficiencies, as well as to create a reference route that takes into account operational conditions not recorded in the travel performance, such as flight restrictions and airspace congestion.

[0082] In the first embodiment, the classification criteria include at least one of the similarity between routes traveled by the mobile object, an index of the travel history of the mobile object, the type, the date and time, and an index of the surrounding conditions of the mobile object. With this configuration, major factors that affect the travel route of the mobile object can be set as classification criteria, so that the travel history data can be appropriately classified.

[0083] Furthermore, in the first embodiment, the representative route creation unit 24 creates a representative route for one achievement cluster by averaging multiple routes indicated by multiple pieces of travel history data with respect to at least one of latitude and longitude, speed, and altitude. Alternatively, the representative route creation unit 24 sets, among the multiple pieces of travel history data, a route of one piece of travel history data that has the greatest similarity to the other pieces of travel history data excluding the one piece of travel history data as the representative route for one achievement cluster. With this configuration, it is possible to appropriately create a representative route used to create a reference route, and therefore it is possible to appropriately create a reference route.

[0084] Furthermore, in the first embodiment, the reference route creation unit 25 extracts feature points for creating a reference route based on the location or time when an air traffic controller issues a control instruction to a moving object. This configuration makes it possible to present locations that serve as a guide for the air traffic controller to issue a control instruction, thereby reducing the burden on the air traffic controller of having to continuously monitor the flight status in order to issue a control instruction.

[0085] Furthermore, in the first embodiment, the reference trajectory creation unit 25 extracts a point that represents two points as a feature point based on the distance between the two points in the point group extracted as feature points or the time difference between the two points in the point group extracted as feature points. With this configuration, it is possible to efficiently extract points at which it is highly likely that an air traffic controller has issued a control instruction to a moving object, and it is possible to further reduce the burden on the air traffic controller of monitoring the flight status in order to issue a control instruction.

[0086] <Embodiment 2> 18 is a block diagram showing a configuration example of a reference movement plan formulation device 11 according to Embodiment 2. In the following, of the components according to Embodiment 2, components that are the same as or similar to the components described above will be assigned the same or similar reference numerals, and different components will be mainly described.

[0087] The configuration of FIG. 18 is the same as the configuration of FIG. 2 with an actual data cleansing unit 181 added. The actual data cleansing unit 181 excludes anomalous travel actual data from travel actual data classified into any of the actual clusters. Whether data is anomalous may be determined based on various distributions, such as the distribution of situation data values, the similarity of track data, and the distribution of flight actual values. The representative route creation unit 24 creates a representative route based on the travel actual data that was not excluded by the actual data cleansing unit 181.

[0088] An example of performing performance data cleansing based on the distribution of visibility values (one type of situation data) will be described below. Figure 19 is a diagram showing the frequency distribution of visibility when the visibility of "No. 0007" in Figure 5 is classified into more detailed categories of "0 to 8,000," which is defined as one classification category. The "0 to 8,000" in Figure 19 is an example of a visibility value, and values other than those in Figure 19 may also be used for the visibility value.

[0089] In the example of Fig. 19, there is almost no actual travel data with a visibility of less than 4,000 m, and therefore there is a high possibility that the actual travel data with a visibility of less than 4,000 m and the actual travel data with a visibility of 4,000 m or more have different trends. Here, the determination of whether there is almost no actual travel data is made based on whether the value is equal to or less than any value between 1% and 5% of the total actual travel data. For example, if the number of samples with a visibility of less than 4,000 m is less than 5% of the samples with a visibility of "0 to 8,000", the actual data cleansing unit 181 determines that there is almost no actual travel data with a visibility of less than 4,000 m.

[0090] When making such a judgment, the actual data cleansing unit 181 determines that the travel actual data with a visibility of less than 4,000 m is likely to have a different trend, and excludes the travel actual data with a visibility of less than 4,000 m from the travel actual data with a visibility of "0 to 8,000".

[0091] Next, we will explain performance data cleansing based on the distribution of flight performance values (one type of situation data). Figure 20 is a diagram showing the frequency distribution of flight times in a specific section. The "0 to 1200" in Figure 20 is an example of a flight time value, and values other than those in Figure 20 may be used for the flight time value. Note that the specific section includes, for example, the section from approach to landing in the case of a landing aircraft.

[0092] In the example of FIG. 20 , it is assumed that the travel history data with a flight time of less than 300 seconds or 900 seconds or more account for less than 5% of the travel history data with a flight time of "0 to 1200". In this case, the travel history data cleansing unit 181 determines that there is almost no travel history data with a flight time of less than 300 seconds or 900 seconds or more. Then, the travel history data cleansing unit 181 excludes the travel history data with a flight time of less than 300 seconds and the travel history data with a flight time of 900 seconds or more from the travel history data with a flight time of "0 to 1200". As a result, the representative route creation unit 24 creates a representative route based on the travel history data not excluded by the travel history data cleansing unit 181, that is, the travel history data that accounts for 5% to 95% of the travel history data with a flight time of "0 to 1200".

[0093] Here, the travel history data that is less than 5% on each side of the distribution is excluded, but this is not limited to this. For example, travel history data on only one side of the distribution may be excluded. Furthermore, the thresholds on one side and the other side of the distribution may be different, for example, excluding travel history data that is less than 5% on the side with a short flight time and excluding travel history data that is less than 1% on the side with a long flight time.

[0094] The exclusion of peculiar movement record data, and therefore peculiar track data, by the record data cleansing unit 181 is not limited to the above. For example, the record data cleansing unit 181 may calculate the similarity between one track data classified into one record cluster and the representative route of the one record cluster. Then, if the similarity is equal to or less than a threshold, the record data cleansing unit 181 may exclude the one track data from the one record cluster. The similarity here may be, for example, the similarity described with reference to FIGS. 6 and 7. Note that the record data cleansing unit 181 may repeatedly calculate the similarity, exclude the track data, and create the representative route until the similarity becomes greater than the threshold.

[0095] As another example of using similarity, the actual data cleansing unit 181 may calculate a similarity Sa between a piece of track data classified into a first actual cluster Ca and a representative route Ra of the first actual cluster Ca. Then, the actual data cleansing unit 181 may calculate a similarity Sb between the piece of track data and a representative route Rb of a second actual cluster Cb that is different from the first actual cluster Ca. Then, if the similarity Sa is lower than the similarity Sb, the actual data cleansing unit 181 may determine that the piece of track data should not belong to the first actual cluster Ca and may exclude the piece of track data from the first actual cluster Ca. In this case, the actual data cleansing unit 181 may add the piece of track data to the second actual cluster Cb.

[0096] <Summary of the second embodiment> According to the reference movement plan formulation device 11 according to the second embodiment, the actual data cleansing unit 181 removes peculiar movement history data from the movement history data classified into any of the actual clusters. With this configuration, an appropriate classification criterion can be obtained, and therefore a highly accurate representative route can be created.

[0097] <Third Embodiment> 21 is a block diagram showing a configuration example of a reference movement plan formulation device 11 according to Embodiment 3. Hereinafter, among the components according to Embodiment 3, components that are the same as or similar to the components described above will be assigned the same or similar reference numerals, and different components will be mainly described.

[0098] The configuration in Fig. 21 is the same as the configuration in Fig. 2, with an achievement cluster evaluation unit 211 added. The achievement cluster evaluation unit 211 changes the classification criteria based on the classification result of the travel achievement data by the achievement data classification unit 23. For example, the achievement cluster evaluation unit 211 creates a distribution of travel achievement data classified into one achievement cluster for attributes not used in classifying the travel achievement data, and sets the attribute as the classification criteria based on the distribution.

[0099] 22 and 23 are diagrams showing an example of a frequency distribution of time periods of track data classified into a certain performance cluster. This frequency distribution is obtained when, for example, the time period "No. 0003" in Fig. 5 is not used for classifying the performance cluster, and the performance cluster is created based on the similarity of the track data "No. 0001."

[0100] The performance cluster evaluation unit 211 determines whether the frequency distribution is biased or dispersed based on a statistical test and a predetermined threshold (for example, "accounting for 80% of the total data"). If the performance cluster evaluation unit 211 determines that the frequency distribution is biased, it sets the biased attribute as a classification criterion.

[0101] For example, the performance cluster evaluation unit 211 determines that the frequency distribution in Fig. 22 is biased toward late night and early morning, determines that there are performance clusters representing late night and early morning, and adds late night and early morning to the classification criteria. Conversely, the performance cluster evaluation unit 211 determines that the frequency distribution in Fig. 23 is dispersed, determines that there are no performance clusters representing any of the time periods, and does not add the time periods to the classification criteria. Note that the distribution described above is a distribution of occurrence frequency (number of samples), but is not limited to this and may be a distribution of proportions indicating, for example, what percentage of all track data from 12 to 15 o'clock is classified into the performance cluster.

[0102] As described above, by extracting attributes with small distributions (i.e., common attributes) and attributes with large distributions (i.e., attributes with large variations) for a performance cluster, it is possible to set appropriate classification criteria that characterize the performance cluster.

[0103] If there are multiple attributes with high occurrence frequencies, or if the attribute values differ greatly, there is a possibility that attributes with different tendencies are mixed together. An example of this will be described below.

[0104] FIG. 24 is a diagram showing an example of a frequency distribution of wind direction for track data classified into a certain achievement cluster. The wind directions include N (north), NE (northeast), E (east), SE (southeast), S (south), SW (southwest), W (west), and NW (northwest). In the example of FIG. 24, the occurrence frequency is high around N (N and NW) and around S (SE, S, and SW), and low around E and W. Furthermore, N and S are opposite directions. In this case, the trends around N and around S may differ. In such cases, the achievement cluster evaluation unit 211 may separate the classification criteria by separating the attributes of the achievement cluster.

[0105] In the field of natural language processing, "TF-IDF (Term Frequency-Inverse Document Frequency)" is known as a method for extracting keywords that characterize the content of a document. TF-IDF is a method for extracting words that appear frequently in a document and that appear infrequently in other documents as keywords for that document. Specifically, TF-IDF calculates the product of tf (TF), which represents the frequency of appearance of a certain word in a document, and idf (IDF), which represents the inverse of the proportion of documents that contain that word among all documents, and extracts keywords with a high value of this product. In other words, TF-IDF extracts keywords that have a high frequency of appearance of a certain word in a document and that are contained in a low proportion of all documents.

[0106] In view of this, the actual result cluster evaluation unit 211 may determine, with respect to an attribute not used in classifying the travel actual data, whether the appearance frequency of first travel actual data classified into the first actual cluster is equal to or greater than a first threshold and whether the appearance frequency of second travel actual data classified into a second actual cluster different from the first actual cluster is equal to or less than a second threshold. Then, when determining that the appearance frequency of the first travel actual data is equal to or greater than the first threshold and the appearance frequency of the second travel actual data is equal to or less than the second threshold, the actual result cluster evaluation unit 211 may set the attribute as a classification criterion.

[0107] FIG. 25 is a diagram showing an example of frequency distribution of wind direction for performance clusters a, b, and c, and FIG. 26 is a diagram showing the numerical values.

[0108] The td-idf of attribute j in achievement cluster i is calculated using equation (2), where f(i,j) is the number of times attribute j appears in achievement cluster i, df(i) is the number of achievement clusters in which attribute j appears, I is the number of clusters (here, the number of clusters in which attribute j appears 30 or more times), and J is the number of attribute classifications.

[0109]

number

[0110] In the examples of Figures 25 and 26, the tf-idf(a,0) of the performance cluster a for the wind direction N(j=0) and the tf-idf(b,4) of the performance cluster b for the wind direction S(j=4) are calculated as shown in equation (3).

[0111]

number

[0112] Since the frequency of occurrence of achievement cluster a for wind direction N is approximately the same as the frequency of occurrence of achievement cluster b for wind direction S, tf(a,0) and tf(b,4) have approximately the same value. On the other hand, since wind direction S appears frequently not only in achievement cluster b but also in achievement cluster c, idf(b,4) of achievement cluster b for wind direction S is smaller than idf(a,0) of wind direction N in achievement cluster a. As a result, tf-idf(b,4) is smaller than tf-idf(a,0), and wind direction N, which is the attribute with a high tf-idf, is set as the classification criterion for achievement clusters.

[0113] This corresponds to wind direction N being set as the classification criterion for the achievement clusters when the appearance frequency of achievement cluster a for wind direction N is equal to or greater than the first threshold and the appearance frequencies of achievement clusters b and c for wind direction N are equal to or less than the second threshold. Furthermore, this corresponds to wind direction S not being set as the classification criterion for the achievement clusters when the appearance frequency of achievement cluster b for wind direction S is equal to or greater than the first threshold and the appearance frequencies of achievement clusters a and c for wind direction S are greater than the second threshold.

[0114] The achievement cluster evaluation unit 211 may integrate two achievement clusters when the representative routes of the two achievement clusters are similar to each other. That is, the achievement cluster evaluation unit 211 may calculate the similarity between the representative route of the first achievement cluster and the representative route of the second achievement cluster, and integrate the first achievement cluster and the second achievement cluster when the similarity is equal to or greater than a threshold.

[0115] Here, because aircraft flying the same route can be controlled by their own flight speeds, air traffic controllers who monitor the distance between aircraft to ensure safe flight can easily predict the future distance between those aircraft. However, because aircraft flying different routes must be predicted based on many factors, such as direction, altitude, and speed, the control work of air traffic controllers becomes more complex. In contrast, the above-described configuration can appropriately reduce the number of representative routes as described above, thereby reducing the number of aircraft flying different routes. As a result, the monitoring work of air traffic controllers can be simplified and the workload on air traffic controllers can be reduced.

[0116] Furthermore, the achievement cluster evaluation unit 211 may create a representative route of a third achievement cluster obtained by temporarily integrating the first achievement cluster and the second achievement cluster. Then, the achievement cluster evaluation unit 211 may calculate the similarity between the representative route of the third achievement cluster and the representative route of the first achievement cluster, and may calculate the similarity between the representative route of the third achievement cluster and the representative route of the second achievement cluster. The achievement cluster evaluation unit 211 may integrate the first achievement cluster and the second achievement cluster when both similarities are equal to or greater than a threshold.

[0117] The fact that the representative route does not change significantly even when the first actual cluster and the second actual cluster are temporarily integrated means that there is little point in separating the first actual cluster and the second actual cluster. Therefore, by integrating the first actual cluster and the second actual cluster, the number of representative routes can be reduced, and the number of aircraft flying different routes can be reduced. As a result, the monitoring work of air traffic controllers can be simplified and the load on air traffic controllers can be reduced. Furthermore, because the number of flight path data samples that make up one actual cluster increases, the quality of the actual data can be improved, and the representative route can be made more accurate.

[0118] <Summary of the Third Embodiment> According to the standard movement plan formulation device 11 of the third embodiment as described above, the actual result cluster evaluation unit 211 changes the classification criteria based on the classification result of the movement history data by the actual result data classification unit 23. With such a configuration, an appropriate classification criteria can be obtained, and therefore a highly accurate representative route can be created.

[0119] <Modification> The performance data receiving unit 21, classification criteria defining unit 22, performance data classifying unit 23, representative route creating unit 24, reference route creating unit 25, and reference route transmitting unit 26 shown in FIG. 2 are hereinafter referred to as the "performance data receiving unit 21, etc." The performance data receiving unit 21, etc. are realized by a processing circuit 81 shown in FIG. 27. That is, the processing circuit 81 includes the performance data receiving unit 21 that acquires movement performance data of a moving object; the classification criteria defining unit 22 that defines one or more classification criteria that serve as standards for classifying the movement performance data; the performance data classifying unit 23 that classifies the movement performance data for each performance cluster to which the movement performance data is reflected based on the classification criteria; the representative route creating unit 24 that creates a representative route for each performance cluster based on the movement performance data classified into the performance cluster; the reference route creating unit 25 that creates a reference route based on feature points of the representative route; and the reference route transmitting unit 26 that transmits the reference route. The processing circuit 81 may be implemented by dedicated hardware or a processor that executes a program stored in memory. Processors include, for example, central processing units, processing units, arithmetic units, microprocessors, microcomputers, and DSPs (Digital Signal Processors).

[0120] When the processing circuit 81 is dedicated hardware, the processing circuit 81 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination of these. The functions of each unit such as the performance data receiving unit 21 may be realized by a circuit in which processing circuits are distributed, or the functions of each unit may be realized together by a single processing circuit.

[0121] When the processing circuit 81 is a processor, the functions of the performance data receiving unit 21 and the like are realized in combination with software and the like. Software and the like may include, for example, software, firmware, or both software and firmware. The software and the like are written as a program and stored in memory. As shown in FIG. 28 , the processor 82 applied to the processing circuit 81 realizes the functions of each unit by reading and executing a program stored in the memory 83. That is, the reference movement plan formulation device 11 includes a memory 83 for storing a program that, when executed by the processing circuit 81, results in the following steps: acquiring movement performance data of a moving object; defining one or more classification criteria as criteria for classifying the movement performance data; classifying the movement performance data for each performance cluster reflecting the movement performance data based on the classification criteria; creating a representative route for each performance cluster based on the movement performance data classified into the performance cluster; creating a reference route based on feature points in the point cloud of the representative route; and transmitting the reference route. In other words, this program can be said to cause a computer to execute the procedures and methods of the performance data receiving unit 21 and the like. Here, the memory 83 may be, for example, a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), an HDD (Hard Disk Drive), a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, a DVD (Digital Versatile Disc), a drive device for any of these, or any storage medium to be used in the future.

[0122] The above describes a configuration in which each function of the performance data receiving unit 21, etc. is realized either by hardware or software, etc. However, the present invention is not limited to this, and a configuration in which part of the performance data receiving unit 21, etc. is realized by dedicated hardware and another part is realized by software, etc. For example, the performance data receiving unit 21's functions can be realized by dedicated hardware such as a processing circuit 81 and acquisition circuitry such as an interface, and the remaining functions can be realized by the processing circuit 81 as a processor 82 reading and executing programs stored in a memory 83.

[0123] As described above, the processing circuitry 81 can realize the above-mentioned functions by hardware, software, or a combination of these.

[0124] The reference movement plan development device described above can also be applied to a reference movement plan development system constructed as a system by appropriately combining a processing device, a communication terminal including a mobile terminal such as a mobile phone, a smartphone, or a tablet, the functions of an application installed in at least one of the processing device and the communication terminal, and a server. In this case, the functions or components of the reference movement plan development device described above may be distributed among the devices that construct the system, or may be concentrated in one of the devices.

[0125] It should be noted that the embodiments and modifications may be freely combined, and the embodiments and modifications may be modified or omitted as appropriate.

[0126] The above description is illustrative in all respects and is not restrictive. It is understood that countless variations not illustrated can be envisioned. [Explanation of symbols]

[0127] 11 Reference movement plan formulation device, 21 Performance data receiving unit, 22 Classification criteria definition unit, 23 Performance data classification unit, 24 Representative route creation unit, 25 Reference route creation unit, 26 Reference route transmission unit, 181 Performance data cleansing unit, 211 Performance cluster evaluation unit.

Claims

1. an achievement data receiving unit that acquires movement achievement data of the moving object; a classification criterion definition unit that defines one or more classification criteria that are used to classify the travel history data; an achievement data classification unit that classifies the travel achievement data for each achievement cluster to which the travel achievement data is reflected based on the classification criteria; a representative route creation unit that creates, for each of the achievement clusters, a representative route that is a route that represents the achievement cluster based on the travel achievement data classified into the achievement clusters; a reference path creation unit that creates a reference path based on the feature points of the representative path; a reference path transmitting unit that transmits the reference path; an achievement cluster evaluation unit that changes the classification criteria based on the classification result of the movement achievement data; A reference movement plan formulation device comprising:

2. The reference movement plan formulation device according to claim 1, The reference path creation unit A reference movement plan formulation device that extracts the characteristic points based on the point or time at which a control instruction is issued from a controller to the moving object.

3. The reference movement plan formulation device according to claim 1, The reference path creation unit a reference movement plan formulation device that extracts, as the feature point, a point that represents two points among the point cloud extracted as the feature points, based on the distance between the two points among the point cloud extracted as the feature points, or the time difference between the two points among the point cloud extracted as the feature points.

4. The reference movement plan formulation device according to claim 1, The classification criteria include at least one of the similarity between the routes traveled by the moving body, an indicator of the moving body's movement history, type, date and time, and an indicator of the surrounding conditions of the moving body.

5. The reference movement plan formulation device according to claim 1, The performance cluster evaluation unit a reference movement plan formulation device that creates a distribution of the movement history data classified into one of the history clusters for an attribute not used in classifying the movement history data, and sets the attribute as the classification criterion based on the distribution.

6. The reference movement plan formulation device according to claim 1, the performance clusters include a first performance cluster and a second performance cluster; The performance cluster evaluation unit a reference movement plan formulation device that sets, with respect to an attribute not used in classifying the movement history data, the attribute as the classification criterion when the frequency of occurrence of the movement history data classified into the first performance cluster is equal to or greater than a first threshold and the frequency of occurrence of the movement history data classified into the second performance cluster is equal to or less than a second threshold.

7. The reference movement plan formulation device according to claim 1, the performance clusters include a first performance cluster and a second performance cluster; The performance cluster evaluation unit a reference movement plan formulation device that, when the representative route of the first performance cluster and the representative route of the second performance cluster are similar, integrates the first performance cluster and the second performance cluster.

8. The reference movement plan formulation device according to claim 1, the performance clusters include a first performance cluster and a second performance cluster; The performance cluster evaluation unit a reference movement plan formulation device that integrates the first performance cluster and the second performance cluster when the representative route of a third performance cluster obtained by provisionally integrating the first performance cluster and the second performance cluster is similar to the representative route of the first performance cluster and the representative route of the second performance cluster, respectively.

9. The reference movement plan formulation device according to claim 1, The representative route creation unit a reference movement plan formulation device that sets a route obtained by averaging multiple routes indicated by multiple pieces of movement history data classified into one of the performance clusters in terms of at least one of latitude and longitude, speed, and altitude as the representative route of the one performance cluster.

10. The reference movement plan formulation device according to claim 1, The representative route creation unit a reference movement plan formulation device that sets, among the plurality of movement history data classified into one of the performance clusters, a route of the one of the movement history data that has the greatest similarity to the other movement history data excluding the one of the movement history data as the representative route of the one of the performance clusters.

11. The reference movement plan formulation device according to claim 1, The reference path creation unit A reference movement plan formulation device extracts, as the characteristic point, a point on the representative route where any one of the direction, speed, and altitude of the moving object has changed by a threshold or more.

12. The reference movement plan formulation device according to claim 2, The reference path creation unit A reference movement plan formulation device that estimates the moving object that is the target of the control instruction based on monitoring of the moving object performed by the controller using a monitoring device.

13. an achievement data receiving unit that acquires movement achievement data of the moving object; a classification criterion definition unit that defines one or more classification criteria that are used to classify the travel history data; an achievement data classification unit that classifies the travel achievement data for each achievement cluster to which the travel achievement data is reflected based on the classification criteria; a representative route creation unit that creates, for each of the achievement clusters, a representative route that is a route that represents the achievement cluster based on the travel achievement data classified into the achievement clusters; a reference path creation unit that creates a reference path based on feature points in the point cloud of the representative path; a reference path transmitting unit that transmits the reference path; an achievement cluster evaluation unit that changes the classification criteria based on the classification result of the movement achievement data; A reference movement planning system comprising:

14. Acquire movement performance data of the moving object, defining one or more classification criteria that are used to classify the travel history data; classifying the travel history data for each history cluster reflecting the travel history data based on the classification criteria; creating a representative route that is a route that represents each of the performance clusters based on the travel performance data classified into the performance clusters; creating a reference route based on feature points in the point cloud of the representative route; transmitting the reference path; A reference movement plan formulation method, which changes the classification criteria based on the classification results of the movement performance data.

Images