Signal analysis device, signal analysis method, and program

Abstract

Provided is a signal analysis device capable of properly detecting an event. A signal analysis device (1) comprises an estimation unit (2) and an event detection unit (4). The estimation unit (2) estimates the speed of a vehicle traveling on a road at each position on the road at each time using signals obtained by measuring the road. The event detection unit (4) detects an event occurring on the road on the basis of at least one of a corrected speed obtained by performing a smoothing process on the estimated speed and an inaccuracy level indicating the degree of inaccuracy of the estimated speed.

Description

Technical Field

[0001] The present invention relates to a signal analysis device, a signal analysis method, and Program .

Background Art

[0002] Techniques for estimating the speed of a vehicle traveling on a road using signals obtained by measurement with a sensing device are known. In relation to this technique, Patent Document 1 discloses a method for estimating traffic flow characteristics (average traffic speed, number of vehicles, speed of each vehicle, etc.) using optical fibers existing along many roads.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] It is desirable to detect events such as traffic jams or accidents using the estimated speed. Here, the signals acquired by measurement in the sensing device may include elements that adversely affect the signals, such as noise. In a method of estimating the speed of a vehicle using such signals including noise or the like, it is difficult to accurately estimate the speed due to the influence of noise or the like. Therefore, there is a risk that events cannot be detected appropriately.

[0005] An object of the present disclosure is to solve such problems, and to provide a signal analysis device, a signal analysis method, and Program .

Means for Solving the Problems

[0006] The signal analysis device according to the present disclosure includes an estimation unit that estimates the speed of a vehicle traveling on a road at each time at each position on the road by using a signal obtained by measuring the road, a corrected speed obtained by performing a smoothing process on the estimated speed which is the estimated speed, and an event detection unit that detects an event that occurred on the road based on at least one of the inaccuracy indicating the degree of inaccuracy of the estimated speed.

[0007] Further, the signal analysis method according to the present disclosure uses a signal obtained by measuring a road to estimate the speed of a vehicle traveling on the road at each time at each position on the road, and detects an event that occurred on the road based on at least one of a corrected speed obtained by performing a smoothing process on the estimated speed which is the estimated speed, and the inaccuracy indicating the degree of inaccuracy of the estimated speed.

[0008] Further, the program according to the present disclosure causes a computer to execute a step of estimating the speed of a vehicle traveling on a road at each time at each position on the road by using a signal obtained by measuring the road, and a step of detecting an event that occurred on the road based on at least one of a corrected speed obtained by performing a smoothing process on the estimated speed which is the estimated speed, and the inaccuracy indicating the degree of inaccuracy of the estimated speed.

Effect of the Invention

[0009] According to the present disclosure, a signal analysis device, a signal analysis method, and Program that can appropriately detect an event can be provided.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Embodiments for Carrying Out the Invention

[0011] (Outline of Embodiments According to the Present Disclosure) Prior to the description of the embodiments of the present disclosure, an overview of the embodiments according to the present disclosure will be described. FIG. 1 is a diagram showing an overview of a signal analysis apparatus 1 according to an embodiment of the present disclosure. Further, FIG. 2 is a diagram showing an overview of a signal analysis method executed by the signal analysis apparatus 1 according to an embodiment of the present disclosure.

[0012] The signal analysis apparatus 1 includes an estimation unit 2 and an event detection unit 4. The estimation unit 2 has a function as an estimation means. The event detection unit 4 has a function as an event detection means. The signal analysis apparatus 1 can be realized by, for example, a computer.

[0013] The estimation unit 2 estimates the speed of a vehicle traveling on a road (step S12). Specifically, the estimation unit 2 estimates the speed of a vehicle traveling on the road at each time at each position on the road using the signal obtained by measuring the road. Note that the signal obtained by measurement can be obtained by, for example, a sensing device such as optical fiber sensing described later. Further, the estimated speed, which is the estimated speed, can be calculated for each position on the road at each time. Details will be described later.

[0014] The event detection unit 4 detects an event using the estimated speed (step S14). Specifically, the event detection unit 4 detects an event that has occurred on the road based on at least one of a corrected speed obtained by performing a smoothing process on the estimated speed and an inaccuracy degree indicating the degree of inaccuracy of the estimated speed. Here, an "event" is any event that occurs on the road, and in particular, an event that causes a change in the speed of a vehicle traveling on the road. For example, in the present embodiment, an "event" is an event that causes a decrease in the speed of a vehicle, but is not limited thereto. An "event" is, for example, traffic congestion or an accident on the road, but is not limited thereto.

[0015] In addition, the "degree of inaccuracy" means the degree to which the corresponding estimated speed can be considered inaccurate. For example, when the estimated speed is unnaturally slow (or fast) compared to the estimated speeds of neighboring positions at that time, the degree of inaccuracy regarding that estimated speed can increase. Also, when the estimated speed is unnaturally slow (or fast) compared to the estimated speeds of neighboring times at that position, the degree of inaccuracy regarding that estimated speed can increase. Details will be described later. Note that the degree of inaccuracy may indicate the degree of abnormality of the estimated speed. Alternatively, the degree of inaccuracy may indicate the degree of inappropriateness of the estimated speed. Also, the degree of inaccuracy may indicate the validity of the estimated speed. In this case, the higher the validity of the estimated speed, the smaller the degree of inaccuracy can be. Also, the degree of inaccuracy may indicate the reliability of the estimated speed. In this case, the higher the reliability of the estimated speed, the smaller the degree of inaccuracy can be.

[0016] (Comparative Example) Here, the optical fiber sensing according to the comparative example will be described. FIG. 3 is a diagram for explaining the optical fiber sensing according to the comparative example. Optical fiber sensing is used to widely monitor a road. An optical fiber sensing system 50 for realizing optical fiber sensing includes a sensing device 52 and an optical fiber cable 54. The optical fiber cable 54 is laid along a road 80. The sensing device 52 is connected to one end of the optical fiber cable 54.

[0017] The sensing device 52 can be realized, for example, by DAS (Distributed Acoustic Sensing) technology. The sensing device 52 can detect vibrations generated at the position where the optical fiber cable 54 is provided. Specifically, as shown by an arrow P, the sensing device 52 makes pulsed light (sensing signal) enter the optical fiber cable 54 toward the end 54e of the optical fiber cable 54. The end 54e is subjected to end processing to suppress reflection of the pulsed light. Alternatively, the optical fiber cable 54 may not be subjected to end processing.

[0018] Here, when pulsed light is incident on the optical fiber cable 54, as indicated by the arrow R, return light called backward scattered light is generated. That is, due to the non-uniformity of the optical fiber cable 54, when pulsed light is incident on the optical fiber cable 54, return light is generated at every position of the optical fiber cable 54. The sensing device 52 observes the return light in time series.

[0019] Then, when vibration is applied to the vicinity of a position X on the optical fiber cable 54 by a vehicle traveling on the road 80 nearby, the quality (amplitude or light intensity, etc.) of the return light generated at that position changes. The sensing device 52 can calculate the generation position X of the return light with the changed quality based on the round-trip time of light. Specifically, let the distance from the sensing device 52 to the generation position X of the return light with the changed quality be L, c be the speed of light in a vacuum, and n be the refractive index of the optical fiber. In this case, the time from when the sensing device 52 emits pulsed light until the return light generated at position X returns to the sensing device 52 is represented by 2Ln / c. Thus, by measuring the time from when the sensing device 52 emits pulsed light until the return light returns, the distance L from the sensing device 52 to position X can be calculated. Note that by ensuring that the sensing device 52 does not emit the next pulsed light from when it emits pulsed light until the return light returns, the mixing of pulsed light in the optical fiber cable 54 can be suppressed.

[0020] In this way, the sensing device 52 measures the return light (signal) at each position of the optical fiber cable 54 in time series by measuring the road 80, and obtains measurement data. Then, the optical fiber sensing system 50 can detect the position and time at which vibration is generated by the vehicle traveling on the road 80 by analyzing the change in the quality such as the intensity or amplitude of the return light. Thereby, the optical fiber sensing system 50 can detect the traveling position of the vehicle at a certain time. Furthermore, the sensing device 52 can obtain the traveling trajectory of the vehicle traveling on the road 80 by performing this position detection in time series. Details will be described later.

[0021] In FIG. 3, the locus Tr11 indicates the traveling locus of the vehicle Ve11. The locus Tr12 indicates the traveling locus of the vehicle Ve12. The locus Tr21 indicates the traveling locus of the vehicle Ve21. The locus Tr22 indicates the traveling locus of the vehicle Ve22. The locus Tr23 indicates the traveling locus of the vehicle Ve23.

[0022] Here, the locus Tr is represented by a graph with the horizontal axis being the position (distance from the sensing device 52) and the vertical axis being the time. In FIG. 3, the right direction of the horizontal axis indicates the distance from the sensing device 52. That is, the left side of the horizontal axis indicates a position closer to the sensing device 52, and the right side indicates a position farther from the sensing device 52. Also, in FIG. 3, the downward direction of the vertical axis indicates the passage of time. That is, the lower side of the vertical axis indicates a time closer to the present, and the upper side indicates a past time.

[0023] Therefore, when the slope of the locus Tr is from the upper right to the lower left (downward to the left), the locus Tr indicates that the corresponding vehicle Ve is traveling in a direction approaching the sensing device 52. On the other hand, when the slope of the locus Tr is from the upper left to the lower right (downward to the right), the locus Tr indicates that the corresponding vehicle Ve is traveling in a direction away from the sensing device 52. Therefore, the vehicle Ve11 corresponding to the downward-left locus Tr11 is traveling in a direction approaching the sensing device 52. Also, the vehicle Ve12 corresponding to the downward-right locus Tr12 is traveling in a direction away from the sensing device 52. Also, the vehicle Ve21 corresponding to the downward-left locus Tr21 is traveling in a direction approaching the sensing device 52. Similarly, the vehicles Ve22 and Ve23 corresponding to the downward-left loci Tr22 and Tr23 respectively are traveling in a direction approaching the sensing device 52.

[0024] Also, the slope of the trajectory Tr corresponds to the speed of the corresponding vehicle Ve. Specifically, a gentle slope of the trajectory Tr indicates that the running of the corresponding vehicle Ve is smooth, that is, it is highly likely that the vehicle Ve is running at a normal running speed. On the other hand, a steep slope of the trajectory Tr indicates that the running of the corresponding vehicle Ve is stalled, that is, it is highly likely that the vehicle Ve is running at a speed slower than the normal running speed. In other words, a steep slope of the trajectory Tr indicates that the corresponding vehicle Ve is highly likely to be involved in a traffic jam or encounter a trouble such as an accident. Therefore, the vehicles Ve11 and Ve12 corresponding to the gently sloping trajectories Tr11 and Tr12 are highly likely to be running smoothly. On the other hand, the vehicles Ve21, Ve22, and Ve23 corresponding to the steep trajectories Tr21, Tr22, and Tr23 are highly likely to be involved in a traffic jam or the like.

[0025] The optical fiber sensing system 50 estimates the running trajectory and the speed of the vehicle by analyzing the signals (measurement data) obtained by the sensing device 52 (step S900). Here, the measurement data is the time waveform data (time series data) of the phase change of the return light generated at each position of the optical fiber cable 54. This phase change corresponds to the intensity of the vibration captured at each position of the optical fiber cable 54. Then, the optical fiber sensing system 50 uses the estimated speed to detect an event such as a traffic jam or an accident (step S920). That is, an event can be detected when the estimated speed is extremely slow at a certain time at a certain position.

[0026] FIG. 4 is a diagram for explaining a method (S900) by which the optical fiber sensing system 50 according to the comparative example estimates the speed of a vehicle. The optical fiber sensing system 50 acquires travel locus data 500 showing the travel locus of each vehicle by using measurement data obtained by the sensing device 52 (step S902). Each line of the travel locus data 500 indicates the travel locus Tr of each vehicle traveling on the road 80. Here, as described above, the measurement data is time-series data of the phase change (vibration intensity) of the return light generated at each position of the optical fiber cable 54, and the travel locus data 500 is obtained by drawing points with an intensity equal to or higher than a predetermined threshold value of this measurement data in space-time. That is, for each position, the time points at which the intensity in the measurement data becomes equal to or higher than the predetermined threshold value are plotted on a graph (map) with the horizontal axis representing the position (distance from the sensing device 52) and the vertical axis representing the time. Thereby, the travel locus data 500 with the horizontal axis representing the position and the vertical axis representing the time is obtained. That is, the travel locus data 500 is a map composed of the position (distance from the sensing device 52) and time.

[0027] As described above, the travel locus data 500 in FIG. 4 is data of a road with the direction toward the sensing device 52 being the traveling direction of the vehicle, and the right side is farther away from the sensing device 52 and the left side is closer to the sensing device 52. That is, the left side of the travel locus data 500 corresponds to the front side of the road (downstream side in the traveling direction of the vehicle), and the right side of the travel locus data 500 corresponds to the rear side of the road (upstream side in the traveling direction of the vehicle). Also, the lower side of the travel locus data 500 is closer to the current time and the upper side is closer to the past time. Since the travel locus data 500 appears to move downward as time passes, it is also called waterfall data.

[0028] Here, depending on the condition of road 80, there are areas that vibrate constantly, such as bridges, or areas with strong vibrations. Alternatively, due to the influence of noise or the like, the intensity may be increased in the measurement data even though the actual vibration is not strong. In this case, the travel trajectory data 500 may not appropriately represent the trajectory of each vehicle. Therefore, in order to remove these influences, the optical fiber sensing system 50 performs a normalization process on the travel trajectory data 500 (step S904). As a result, the slanted lines in the travel trajectory data 500 will represent the trajectory Tr of each vehicle to a certain extent well.

[0029] Next, the optical fiber sensing system 50 divides the travel trajectory data 500 into a plurality of patches (step S906). For example, as shown in FIG. 4, the optical fiber sensing system 50 divides the travel trajectory data 500 into patches 502 (sections) having a size with a length in the horizontal direction (spatial direction) of 1 km and a length in the vertical direction (time direction) of 1 minute (min). Here, the patch 502 corresponds to the unit for calculating the estimated speed. That is, for each patch 502, from the inclination of each trajectory included in the patch 502, the average estimated speed of the vehicle traveling at the position corresponding to the patch 502 at the time corresponding to the patch 502 can be obtained.

[0030] Arrow Pa shows an image of a specific example of the travel trajectory data 500 divided into patches 502. Here, in the actual travel trajectory data 500, as shown by the ellipse, the trajectory (hatched line) may be interrupted due to the influence of noise or the like. Therefore, the optical fiber sensing system 50 according to the comparative example uses the analysis engine 92 to remove the noise included in each patch 502 (step S908). The analysis engine 92 can be realized by a machine learning algorithm such as a DNN (Deep Neural Network). The analysis engine 92 is trained by a large number of teacher data corresponding to the travel trajectory data represented by position (distance) and time to output travel trajectory data in which the slope of the hatched line appropriately represents the speed. The analysis engine 92 takes the travel trajectory data (image data) of each patch 502 as input and outputs travel trajectory data (image data) in which the influence of noise is removed and the hatched line represents the speed. The optical fiber sensing system 50 calculates the average estimated speed corresponding to each patch 502 from the slope of each travel trajectory (hatched line) in the travel trajectory data output from the analysis engine 92 (step S910). Specifically, the optical fiber sensing system 50 calculates the speed from the slope of each hatched line (trajectory) included in the patch 502, and calculates the average estimated speed in the patch 502 by averaging the calculated speeds.

[0031] FIG. 5 is a diagram for explaining event detection according to a comparative example. FIG. 5 shows the temporal change of the speed at a certain position. As the speed data indicated by arrow A1, when there are no accidents, traffic jams, etc., actually, a continuous speed change can occur. However, in the above-described analysis engine 92, there is a possibility that the influence of noise, etc. cannot be completely removed. That is, the noise generated in the travel locus data 500 may depend on the usage environment of the optical fiber cable 54, such as the state of the road 80 (such as a bridge or a tunnel), and the state of the vehicle Ve traveling on the road 80 (such as the weight of the vehicle). And since the analysis engine 92 can be learned to be commonly used for any environment, it may be difficult to appropriately remove the influence of noise, etc. for various environments (special environments) as described above only by using the analysis engine 92.

[0032] Therefore, like the speed data indicated by arrow A2, in the method according to the comparative example, there is a possibility that an apparently unnatural deceleration different from the actual situation may be estimated due to the influence of noise, etc. In other words, this unnatural deceleration is an estimated speed due to an erroneous estimation caused by the influence of noise, etc. In this case, if the estimated speed due to the erroneous estimation is lower than the threshold for low-speed detection, there is a possibility that an erroneous detection may occur, such as an event being detected even though an event such as a traffic jam has not actually occurred. Therefore, there is a possibility that an erroneous report may be made, such as a report indicating that an event has occurred even though an event such as a traffic jam has not actually occurred.

[0033] In contrast, the signal analysis device 1 according to the present embodiment detects an event based on at least one of a corrected speed obtained by performing a smoothing process on the estimated speed and an inaccuracy degree indicating the degree of inaccuracy of the estimated speed. Here, the corrected speed is one in which the estimated speed due to misestimation is corrected. Therefore, by detecting an event based on this corrected speed, false detection of the event is suppressed, and thus it becomes possible to appropriately detect the event. Also, for the estimated speed due to misestimation, the corresponding degree of inaccuracy can increase. Therefore, it is possible to prevent event detection from being performed on an estimated speed with a large degree of inaccuracy. Therefore, by detecting an event based on the degree of inaccuracy, false detection of the event is suppressed, and thus it becomes possible to appropriately detect the event.

[0034] Note that it is also possible to appropriately detect an event by using a signal analysis system having the signal analysis device 1, the sensing device 52, and the optical fiber cable 54. Also, it is possible to appropriately detect an event by using the signal analysis method realized by the signal analysis device 1 and a program that executes the signal analysis method.

[0035] (Embodiment 1) Hereinafter, embodiments will be described with reference to the drawings. For clarity of explanation, the following description and drawings are appropriately omitted and simplified. Also, in each drawing, the same elements are denoted by the same reference numerals, and duplicate explanations are omitted as necessary.

[0036] FIG. 6 is a diagram showing a signal analysis system 10 according to Embodiment 1. The signal analysis system 10 includes a sensing device 52, an optical fiber cable 54, and a signal analysis device 100. The signal analysis device 100 is communicably connected to the sensing device 52 via a wired or wireless network 20.

[0037] As described above, the sensing device 52 emits pulsed light into the optical fiber cable 54 and receives the return light. Thereby, the sensing device 52 acquires measurement data of the return light (signal) at each position of the optical fiber cable 54. The sensing device 52 transmits the measurement data (signal) at each position of the optical fiber cable 54 to the signal analysis device 100.

[0038] The signal analysis device 100 corresponds to the signal analysis device 1 shown in FIG. 1. The signal analysis device 100 is, for example, a computer such as a server or a personal computer. The signal analysis device 100 uses the signal obtained by the measurement by the sensing device 52 to estimate the speed of the vehicle traveling on the road 80, and uses the estimated speed (estimated speed) to detect an event occurring on the road. Details will be described later.

[0039] FIG. 7 is a diagram showing the configuration of the signal analysis device 100 according to the first embodiment. As shown in FIG. 7, the signal analysis device 100 includes, as main hardware components, a control unit 102, a storage unit 104, a communication unit 106, and an interface unit 108 (IF; Interface). The control unit 102, the storage unit 104, the communication unit 106, and the interface unit 108 are interconnected via a data bus or the like. Note that the sensing device 52 may also have the hardware configuration of the signal analysis device 100 shown in FIG. 7.

[0040] The control unit 102 is a processor such as a CPU (Central Processing Unit). The control unit 102 has a function as an arithmetic unit that performs control processing, arithmetic processing, and the like. Note that the control unit 102 may have a plurality of processors. The storage unit 104 is a storage device such as a memory or a hard disk. The storage unit 104 is, for example, a ROM (Read Only Memory) or a RAM (Random Access Memory). The storage unit 104 has a function of storing a control program, an arithmetic program, and the like executed by the control unit 102. That is, the storage unit 104 (memory) stores one or more instructions. Further, the storage unit 104 has a function of temporarily storing processing data and the like. The storage unit 104 may include a database. Also, the storage unit 104 may have a plurality of memories.

[0041] The communication unit 106 performs processing necessary for communicating with other devices such as the sensing device 52 via a network. The communication unit 106 may include a communication port, a router, a firewall, and the like. The interface unit 108 (IF; Interface) is, for example, a user interface (UI). The interface unit 108 has an input device such as a keyboard, a touch panel, or a mouse, and an output device such as a display or a speaker. The interface unit 108 may be configured such that an input device and an output device are integrated, such as a touch screen (touch panel). The interface unit 108 receives an operation of inputting data by a user (operator) and outputs information to the user. The interface unit 108 outputs, for example, that an event has occurred when the event is detected.

[0042] The signal analysis device 100 according to Embodiment 1 includes, as components, a signal acquisition unit 110, a trajectory acquisition unit 120, a speed estimation unit 130, an estimated speed processing unit 140, an event detection unit 150, and an event notification unit 160. The estimated speed processing unit 140 includes an estimated speed data storage unit 142, a corrected speed calculation unit 144, and an inaccuracy degree calculation unit 146.

[0043] The signal acquisition unit 110 has a function as a signal acquisition means. The trajectory acquisition unit 120 has a function as a trajectory acquisition means. The speed estimation unit 130 corresponds to the estimation unit 2 shown in FIG. 1. The speed estimation unit 130 has a function as a speed estimation means (estimation means). The estimated speed processing unit 140 has a function as an estimated speed processing means. The estimated speed data storage unit 142 has a function as an estimated speed data storage means. The corrected speed calculation unit 144 has a function as a corrected speed calculation means. The inaccuracy degree calculation unit 146 has a function as an inaccuracy degree calculation means. The event detection unit 150 corresponds to the event detection unit 4 shown in FIG. 1. The event detection unit 150 has a function as an event detection means. The event notification unit 160 has a function as an event notification means.

[0044] Note that each of the above-described components can be realized, for example, by causing a program to be executed under the control of the control unit 102. More specifically, each component can be realized by the control unit 102 executing a program (instruction) stored in the storage unit 104. Further, by recording a necessary program on an arbitrary non-volatile recording medium and installing it as necessary, each component may be realized. Also, each component is not limited to being realized by software by a program, and may be realized by any combination of hardware, firmware, and software. Further, each component may be realized using a user-programmable integrated circuit such as, for example, an FPGA (field-programmable gate array) or a microcomputer. In this case, a program composed of the above-described components may be realized using this integrated circuit. Note that the specific functions of each component will be described later with reference to FIG. 8 and the like.

[0045] FIG. 8 is a flowchart showing a signal analysis method executed by the signal analysis apparatus 100 according to the first embodiment. The signal acquisition unit 110 acquires the measured signal (step S102). Specifically, the signal acquisition unit 110 acquires measurement data (signal) from the sensing device 52.

[0046] The trajectory acquisition unit 120 acquires trajectory data (step S104). Specifically, the trajectory acquisition unit 120 uses the measurement data acquired from the sensing device 52, like the optical fiber sensing system 50 described above, to acquire trajectory data (travel trajectory data 500) indicating the travel trajectories of the respective vehicles. As described above, the trajectory data is a map composed of position (distance from the sensing device 52) and time. Also, the processing of the trajectory acquisition unit 120 corresponds to the processing of S902 described above.

[0047] The speed estimation unit 130 estimates the speed of the vehicles traveling on the road (step S106). Specifically, the speed estimation unit 130 uses the trajectory data (travel trajectory data 500) to estimate the speed of the vehicles traveling on the road at each time at each position on the road. More specifically, the speed estimation unit 130, like the optical fiber sensing system 50 described above, divides the travel trajectory data 500 into patches 502 of a predetermined distance and a predetermined time size, and calculates the estimated speed of the vehicle corresponding to each divided patch 502. The processing of the speed estimation unit 130 corresponds to the processing of S904 - S910 described above.

[0048] The estimated speed processing unit 140 performs processing on the calculated estimated speed (S110 - S114). The estimated speed data storage unit 142 stores the estimated speed data (step S110). Specifically, the estimated speed data storage unit 142 stores the estimated speed data indicating the values of the estimated speed for each position and time (i.e., for each patch) calculated by the speed estimation unit 130. The estimated speed data storage unit 142 can be realized by the storage unit 104. The estimated speed data may be, for example, data in CSV (Comma Separated Value) format, or may be two - dimensional matrix - formatted data composed of position components and time components.

[0049] FIG. 9 is a diagram illustrating an estimated speed map 200 according to Embodiment 1. The estimated speed map may be configured by the estimated speed data stored in the estimated speed data storage unit 142. In the estimated speed map 200 illustrated in FIG. 9, the horizontal axis represents the position (distance from the sensing device 52), and the vertical axis represents the time. The right direction of the horizontal axis of the estimated speed map 200 corresponds to the direction away from the sensing device 52. Here, when the estimated speed map 200 illustrated in FIG. 9 corresponds to a road where the direction toward the sensing device 52 is the traveling direction of the vehicle, the left direction of the estimated speed map 200 corresponds to the front of the road (downstream direction in the traveling direction of the vehicle). On the other hand, the right direction of the estimated speed map 200 corresponds to the rear of the road (upstream direction in the traveling direction of the vehicle). Also, the downward direction of the vertical axis of the estimated speed map 200 corresponds to the direction of time elapse. Note that the interface unit 108, which is a display, may display this estimated speed map 200.

[0050] Further, the estimated speed map 200 illustrated in FIG. 9 is divided into a plurality of patches 202. The patch 202 corresponds to the patch 502 shown in FIG. 4. Therefore, the numerical values described in each patch 202 of the estimated speed map 200 in FIG. 9 indicate the estimated speed at the position and time of the corresponding patch 202. Note that the "position of the patch 202" does not indicate a precise point in space (on the road), and may correspond to a spatial region of a predetermined range (patch size: 1 km, etc.) along the road. Similarly, the "time of the patch 202" does not indicate a precise time on the time axis, and may correspond to a time region of a predetermined range (patch size: 1 minute, etc.) along the time axis.

[0051] For example, in FIG. 9, the estimated speed at the position (first position) and time (first time) corresponding to the patch 202A is 20 (km / h). Also, the patch 202B on the left side of the patch 202A corresponds to a position near the position (first position) corresponding to the patch 202A at the time (first time) corresponding to the patch 202A. Similarly, the patch 202C on the right side of the patch 202A corresponds to a position near the position (first position) corresponding to the patch 202A at the time (first time) corresponding to the patch 202A.

[0052] And, patch 202B corresponds to a position closer to the sensing device 52 by one patch (e.g., 1 km) than the position (first position) corresponding to patch 202A at the same time as the time (first time) corresponding to patch 202A. The estimated speed at the position and time corresponding to patch 202B is 60 (km / h). Also, patch 202C corresponds to a position farther from the sensing device 52 by one patch (e.g., 1 km) than the position (first position) corresponding to patch 202A at the same time as the time (first time) corresponding to patch 202A. The estimated speed at the position and time corresponding to patch 202C is 50 (km / h).

[0053] Also, patch 202D above patch 202A corresponds to a time in the vicinity of the time (first time) corresponding to patch 202A at the position (first position) corresponding to patch 202A. Similarly, patch 202E below patch 202A corresponds to a time in the vicinity of the time (first time) corresponding to patch 202A at the position (first position) corresponding to patch 202A. Further, patch 202F two above patch 202A may also correspond to a time in the vicinity of the time (first time) corresponding to patch 202A at the position (first position) corresponding to patch 202A, together with patch 202D.

[0054] Patch 202D corresponds to a time that is one patch (e.g., 1 minute) before the time corresponding to Patch 202A (the first time) at the same position as the position corresponding to Patch 202A (the first position). The estimated speed at the position and time corresponding to Patch 202D is 70 (km / h). Also, Patch 202E corresponds to a time that is one patch (e.g., 1 minute) after the time corresponding to Patch 202A (the first time) at the same position as the position corresponding to Patch 202A (the first position). The estimated speed at the position and time corresponding to Patch 202E is 50 (km / h). Also, Patch 202F corresponds to a time that is two patches (e.g., 2 minutes) before the time corresponding to Patch 202A (the first time) at the same position as the position corresponding to Patch 202A (the first position). The estimated speed at the position and time corresponding to Patch 202F is 80 (km / h).

[0055] Returning to the description of FIG. 8, the corrected speed calculation unit 144 calculates a corrected speed by performing a smoothing process on the corresponding estimated speed for each patch 202 (step S112). Specifically, the corrected speed calculation unit 144 performs a smoothing process using the estimated speeds of the patches 202 around (in the vicinity of) the patch 202 (patch X) for which the corrected speed is to be calculated, thereby calculating a corrected speed obtained by correcting the estimated speed of patch X. That is, the corrected speed calculation unit 144 performs a smoothing process using at least one of the estimated speed of the patch 202 at a position near the position of patch X at the time of patch X and the estimated speed of the patch 202 at a time near the time of patch X at the position of patch X.

[0056] Here, assume that patch X corresponds to the first time at the first position, and the estimated speed of patch X is the first estimated speed. In this case, the corrected speed calculation unit 144 performs a smoothing process on the first estimated speed using at least one of the estimated speed at the first time at a position near the first position and the estimated speed at the first position at a time near the first time. Thereby, the corrected speed calculation unit 144 calculates a corrected speed regarding the estimated speed (the first estimated speed) of patch X.

[0057] FIG. 10 is a diagram for explaining the processing of the correction speed calculation unit 144 according to Embodiment 1. FIG. 10 shows an example of a method for calculating a correction speed obtained by correcting the estimated speed regarding the patch 202A shown in FIG. 9. In the example of FIG. 10, the correction speed calculation unit 144 performs smoothing in the spatial direction and smoothing in the temporal direction on the estimated speed of the patch 202A. Specifically, the correction speed calculation unit 144 performs a smoothing process using the estimated speed regarding the patch 202A, the estimated speed regarding the patch 202B, the estimated speed regarding the patch 202C, the estimated speed regarding the patch 202D, and the estimated speed regarding the patch 202F. That is, the correction speed calculation unit 144 performs smoothing in the spatial direction using one patch 202B and 202C before and after the patch 202A at the same time as the patch 202A. On the other hand, the correction speed calculation unit 144 performs smoothing in the temporal direction using the past two patches 202D and 202F of the patch 202A at the same position as the patch 202A.

[0058] More specifically, the correction speed calculation unit 144 calculates the average value of the estimated speeds regarding each of the patches 202A, 202B, 202C, 202D, and 202F as a smoothing process. The calculated average value corresponds to the correction speed. That is, the correction speed calculation unit 144 calculates the correction speed by calculating the average value of the estimated speed of the patch X for which the correction speed is to be calculated and the estimated speeds of the patches around the patch X. In the example of FIG. 10, the correction speed calculation unit 144 calculates the correction speed as (20 + 60 + 50 + 70 + 80) / 5 = 56 (km / h) with respect to the estimated speed (20 km / h) of the patch 202A. The correction speed calculation unit 144 generates the correction speed map 220 illustrated in FIG. 10 by performing the same process for all the patches 202. Note that only the correction speed (56 km / h) regarding the patch 222A corresponding to the patch 202A of the estimated speed map 200 is shown in the correction speed map 220 illustrated in FIG. 10, but actually, the correction speed is calculated for all the patches 222. Note that the interface unit 108 which is a display may display this correction speed map 220.

[0059] In the above example, spatial smoothing and temporal smoothing are performed on the estimated velocity of patch X for which the correction velocity is to be calculated, but it is not limited thereto. That is, only spatial smoothing or only temporal smoothing may be performed on the estimated velocity of patch X for which the correction velocity is to be calculated. For example, when the spatial resolution is not good, that is, when the size of patch 202 in the spatial direction (position direction; horizontal direction) is large (e.g., about 10 km), even if the estimated velocities of adjacent patches 202 are significantly different from each other, it may not be unnatural. Therefore, in this case, only temporal smoothing may be performed.

[0060] Also, in the above example, when performing the smoothing process, the estimated velocity of patch X and the estimated velocities of four surrounding patches X are used, but it is not limited to this. The number of patches X used when performing the smoothing process is arbitrary. Also, in the above example, as the smoothing process, the average (simple average) of the estimated velocity of patch X and the estimated velocities of four surrounding patches X is calculated, but it is not limited to this. Any smoothing process can be performed. For example, the smoothing process may be performed using a weighted average.

[0061] Also, in the above example, for temporal smoothing, the estimated velocities of patches before (i.e., past) the time of patch X for which the correction velocity is to be calculated are used. However, if the estimated velocities for patches after the time of patch X for which the correction velocity is to be calculated have already been calculated, the smoothing process may be performed using the estimated velocities of patches after the time of patch X for which the correction velocity is to be calculated. On the other hand, by performing the smoothing process using the estimated velocities of patches before the time of patch X for which the correction velocity is to be calculated, immediately after the estimated velocity of patch X for which the correction velocity is to be calculated is calculated, the correction velocity can be calculated. Therefore, it becomes possible to ensure the immediacy of correction velocity calculation and the immediacy of event detection described later.

[0062] Return to the description of FIG. 8. The inaccuracy calculation unit 146 calculates the inaccuracy of the estimated speed for each patch 202 (step S114). Specifically, the inaccuracy calculation unit 146 calculates the inaccuracy of the estimated speed of patch X using the estimated speeds of the patches 202 around (in the vicinity of) the patch 202 (patch X) for which the inaccuracy is to be calculated. That is, the inaccuracy calculation unit 146 calculates the inaccuracy using at least one of the estimated speed of the patch 202 at a position near the position of patch X at the time of patch X and the estimated speed of the patch 202 at a time near the time of patch X at the position of patch X.

[0063] Here, assume that patch X corresponds to the first position at the first time, and the estimated speed of patch X is the first estimated speed. In this case, the inaccuracy calculation unit 146 calculates the inaccuracy using at least one of the estimated speed at the first time at a position near the first position and the estimated speed at the first position at a time near the first time with respect to the first estimated speed. Thereby, the inaccuracy calculation unit 146 calculates the inaccuracy regarding the estimated speed (the first estimated speed) of patch X.

[0064] The greater the inaccuracy calculated with respect to the estimated speed of patch X, the higher the likelihood that the estimated speed of that patch X deviates from the actual speed due to the influence of noise or the like. That is, the greater the inaccuracy calculated with respect to the estimated speed of patch X, the lower the validity of the estimated speed of that patch X. Note that the inaccuracy may correspond to, for example, the variation in the estimated speed of patch X and the estimated speeds of patches 202 around patch X. That is, the inaccuracy may be set to increase as the variation between the first estimated speed (the estimated speed of patch X), the estimated speed at the first time at a position near the first position, and the estimated speed at the first position at a time near the first time increases. The inaccuracy may be, for example, an index representing variation, such as variance, standard deviation, or mean deviation. Also, the inaccuracy may be, for example, the average value of the differences between the estimated speed of patch X and the estimated speeds of a plurality of patches 202 around patch X as an index representing variation. That is, the inaccuracy calculation unit 146 may calculate, as the inaccuracy, an index representing the variation between the first estimated speed and at least one of the estimated speed at the first time at a position near the first position and the estimated speed at the first position at a time near the first time.

[0065] FIG. 11 is a diagram for explaining the processing of the inaccuracy calculation unit 146 according to Embodiment 1. FIG. 11 shows an example of a method for calculating the inaccuracy rate with respect to the estimated speed for the patch 202A shown in FIG. 9. Similar to the example of FIG. 10, in the example of FIG. 11, the inaccuracy calculation unit 146 calculates the inaccuracy rate from the variations in the spatial direction and the variations in the temporal direction with respect to the estimated speed of the patch 202A. Specifically, the inaccuracy calculation unit 146 calculates the inaccuracy rate using the estimated speed related to the patch 202A, the estimated speed related to the patch 202B, the estimated speed related to the patch 202C, the estimated speed related to the patch 202D, and the estimated speed related to the patch 202F. That is, the inaccuracy calculation unit 146 calculates the inaccuracy rate using one patch 202B and 202C before and after the patch 202A at the same time in the spatial direction with respect to the patch 202A. On the other hand, the inaccuracy calculation unit 146 calculates the inaccuracy rate using the past two patches 202D and 202F of the patch 202A at the same position as the patch 202A in the temporal direction.

[0066] More specifically, in the example of FIG. 11, the inaccuracy calculation unit 146 calculates the variance of the estimated speeds related to the patches 202A, 202B, 202C, 202D, and 202F as the inaccuracy rate with respect to the estimated speed of the patch 202A. That is, the inaccuracy calculation unit 146 calculates the inaccuracy rate by calculating the variance between the estimated speed of the patch X for which the inaccuracy rate is to be calculated and the estimated speeds of the patches around the patch X. In the example of FIG. 11, the inaccuracy calculation unit 146 calculates the variance as the inaccuracy rate with respect to the estimated speed (20 km / h) of the patch 202A as {(20 - 56) 2 +(60 - 56) 2 +(50 - 56) 2 +(70 - 56) 2 +(80 - 56) 2It is calculated as} / 5 = 424. The inaccuracy calculation unit 146 generates an inaccuracy map 240 illustrated in FIG. 11 by performing the same process for all the patches 202. Note that only the inaccuracy (424) regarding the patch 242A corresponding to the patch 202A of the estimated speed map 200 is shown in the inaccuracy map 240 illustrated in FIG. 11, but actually, the inaccuracy is calculated for all the patches 242. The interface unit 108, which is a display, may display this inaccuracy map 240.

[0067] Note that in the above example, the inaccuracy associated with the variation in the spatial direction and the variation in the temporal direction is calculated for the estimated speed of the patch X for which the inaccuracy is to be calculated, but it is not limited to this. That is, the inaccuracy associated only with the variation in the spatial direction may be calculated for the estimated speed of the patch X for which the inaccuracy is to be calculated, or the inaccuracy associated only with the variation in the temporal direction may be calculated. As described above, when the spatial resolution is not good, that is, when the size of the patch 202 in the spatial direction (position direction; horizontal direction) is large (for example, about 10 km), even if the estimated speeds of adjacent patches 202 are significantly different from each other, it may not be unnatural. Therefore, in this case, the inaccuracy may be calculated considering only the variation in the temporal direction.

[0068] Also, in the above example, when calculating the inaccuracy degree, the estimated speed of patch X and the estimated speeds of the four patches X around it are used, but it is not limited to this. The number of patches X used when calculating the inaccuracy degree is arbitrary. Also, in the above example, the patch 202 used when calculating the inaccuracy degree with respect to the estimated speed of patch X is the same as the patch 202 used when performing the smoothing process on the estimated speed of patch X, but it is not limited to this. Also, in the above example, as the inaccuracy degree with respect to the estimated speed of patch X, the variance between the estimated speed of patch X and the estimated speeds of the patches X around it is calculated, but it is not limited to this. The inaccuracy degree can be calculated using an index representing any variation. For example, as the inaccuracy degree with respect to the estimated speed of patch X, the standard deviation between the estimated speed of patch X and the estimated speeds of the four patches X around it may be calculated, or the average deviation between them may be calculated.

[0069] Also, in the above example, regarding the variation in the time direction, the estimated speeds of patches before (i.e., in the past) the time of patch X for which the inaccuracy degree is to be calculated are used. However, if the estimated speeds regarding patches after the time of patch X for which the inaccuracy degree is to be calculated have already been calculated, the inaccuracy degree may be calculated using the estimated speeds of the patches after the time of patch X for which the inaccuracy degree is to be calculated. On the other hand, by calculating the inaccuracy degree using the estimated speeds of the patches before the time of patch X for which the inaccuracy degree is to be calculated, immediately after the estimated speed of patch X for which the inaccuracy degree is to be calculated is calculated, the inaccuracy degree can be calculated. Therefore, it becomes possible to ensure the immediacy of inaccuracy degree calculation and the immediacy of event detection described later.

[0070] Return to the description of FIG. 8. The event detection unit 150 performs event detection for each patch 202 (step S120). That is, the event detection unit 150 determines whether or not an event (such as traffic congestion) that causes a decrease in the vehicle speed has been detected for each patch 202. In other words, the event detection unit 150 determines whether or not an event has occurred at the position of the road corresponding to the patch 202 at the time corresponding to the patch 202, using the correction speed and the inaccuracy degree corresponding to the patch 202 for each patch 202.

[0071] Specifically, for each patch 202, the event detection unit 150 determines whether the corresponding correction speed is equal to or lower than a predetermined threshold value Vth (first threshold value) and whether the corresponding inaccuracy degree is equal to or lower than a predetermined threshold value Dth (second threshold value). When this determination is true, the event detection unit 150 determines that an event has occurred at the position corresponding to the patch 202 at the time corresponding to the patch 202. On the other hand, when this determination is false, the event detection unit 150 does not determine that an event is occurring at the position corresponding to the patch 202 at the time corresponding to the patch 202.

[0072] When an event is detected (YES in S120), the event notification unit 160 issues a report indicating that the event has occurred at the position corresponding to the patch 202 and at the time corresponding to the patch 202 (step S130). For example, the event notification unit 160 may perform control so that in the estimated speed map 200 displayed on the interface unit 108, the display mode of the patch 202 where the event is detected is made more prominent than the display modes of the other patches 202. Alternatively, the event notification unit 160 may perform control to output a display indicating that the event has occurred to the interface unit 108, which is a display, at the position corresponding to the patch 202 and at the time corresponding to the patch 202. Alternatively, the event notification unit 160 may perform control to output a sound indicating that the event has occurred to the interface unit 108, which is a speaker, at the position corresponding to the patch 202 and at the time corresponding to the patch 202. Then, the processing flow returns to S112, and the same processing is performed for the other patches 202.

[0073] On the other hand, when an event is not detected (NO in S120), the process of S130 is not performed. Then, the processing flow returns to S112, and the same processing is performed for the other patches 202.

[0074] Figures 12 to 16 are diagrams for explaining the effect of performing event detection using the corrected speed and the inaccuracy degree. Regarding the traffic flow on a road, it is usually extremely rare for the vehicle speed at a certain position at a certain time to be significantly different (substantially lower) compared to the vehicle speeds at the surrounding positions of that position. Similarly, it is usually extremely rare for the vehicle speed at a certain position at a certain time to be significantly different (substantially lower) compared to the vehicle speeds at the surrounding times of that time. Therefore, such an estimated speed that is significantly different compared to the surrounding positions and times is likely to be a speed that has been erroneously estimated due to the influence of noise or the like.

[0075] For example, in the estimated speed maps 200 of FIGS. 9 to 11, the estimated speed (20 km / h) of patch 202A is considerably lower than the estimated speeds of the surrounding patches 202 (202B, 202C, 202D, 202E, 202F). Therefore, the estimated speed of patch 202A may be an incorrect speed affected by noise or the like. In such a case, as shown by the ellipse B2 of the speed data indicated by arrow A2 in FIG. 12, actually, although the vehicle speed at that position and time is not so low, the estimated speed may be below the threshold (Vth) for low-speed detection. In this case, if an event is detected, an alarm indicating the occurrence of the event may be issued.

[0076] On the other hand, by performing the smoothing process as described above, as illustrated in FIG. 10, the estimated speed of patch 202A is corrected using the estimated speeds of the surrounding patches 202. As a result, the difference between the corrected speed (56 km / h) of patch 202A and the corrected speeds in the surrounding patches 202 can be reduced. Thereby, the estimated speed of patch 202A becomes a corrected speed with the influence of noise or the like suppressed, and can be close to the actual vehicle speed. As a result, as shown by the ellipse B3 of the speed data indicated by arrow A3 in FIG. 12, the speed change is suppressed. Therefore, when the vehicle speed is not actually decreasing, it is suppressed that the vehicle speed (corrected speed) for event detection falls below the threshold for low-speed detection. Therefore, it is suppressed that a false alarm indicating that an event has occurred is issued although an event such as traffic congestion has not actually occurred. Therefore, it becomes possible to appropriately perform event detection.

[0077] Also, as shown by the speed data indicated by arrow A4 in FIG. 13, when the influence of noise or the like is considerably large, the error between the actual vehicle speed and the estimated speed may become considerably large. In this case, the estimated speed may be much lower than the threshold. In this case, as shown by the speed data indicated by arrow A5, even if the estimated speed is corrected by smoothing, as shown by the ellipse B5, the corrected speed may still be below the threshold.

[0078] On the other hand, by calculating the inaccuracy level (variation) as described above, it is possible to determine the inaccuracy of the estimated speed. That is, for an estimated speed with a large inaccuracy level, it is highly likely that it is greatly affected by noise or the like and is due to an incorrect estimation, so it can be made untrusted. Specifically, in the inaccuracy level data indicated by arrow A6, when the inaccuracy level becomes equal to or higher than the threshold value (Dth) as in ellipse B6, event detection is not performed using the corresponding speed (corrected speed and estimated speed). In this case, even if the speed (corrected speed and estimated speed) at that time becomes lower than the threshold value, it is not determined that an event has been detected. Therefore, when an incorrect estimation occurs where the estimated speed suddenly drops as in the example of FIG. 13, the reporting of an event is suppressed. Therefore, even if the corrected speed becomes lower than the threshold value even though the estimated speed is corrected by smoothing due to the large influence of noise or the like when in fact no event has occurred, false reporting is suppressed. Therefore, it becomes possible to perform event detection more appropriately.

[0079] FIG. 14 shows a specific example of the estimated speed map 200. FIG. 15 shows a specific example of the corrected speed map 220 corresponding to the estimated speed map 200 illustrated in FIG. 14. FIG. 16 shows a specific example of the inaccuracy level map 240 corresponding to the estimated speed map 200 illustrated in FIG. 14.

[0080] The estimated speed map 200 illustrated in FIG. 14 has the horizontal axis representing the position (distance from the sensing device 52) and the vertical axis representing the time. The right direction of the horizontal axis of the estimated speed map 200 corresponds to the direction away from the sensing device 52. Here, the estimated speed map 200 illustrated in FIG. 14 corresponds to a road where the direction toward the sensing device 52 is taken as the traveling direction of the vehicle. Therefore, the left direction of the estimated speed map 200 illustrated in FIG. 14 corresponds to the front of the road (downstream direction of the traveling direction of the vehicle). On the other hand, the right direction of the estimated speed map 200 corresponds to the rear of the road (upstream direction of the traveling direction of the vehicle). Also, the downward direction of the vertical axis of the estimated speed map 200 corresponds to the direction of the passage of time. The same applies to the corrected speed map 220 illustrated in FIG. 15 and the inaccuracy map 240 illustrated in FIG. 16 with respect to these vertical and horizontal axes.

[0081] Here, in the examples of FIGS. 14 to 16, the threshold value of the corrected speed is set to Vth = 40 km / h. That is, when an event such as a traffic jam occurs, at the position and time where the event occurs, the vehicle speed is 40 km / h or less. In other words, when the vehicle speed exceeds 40 km / h, it can be considered that no event has occurred. Also, in the examples of FIGS. 14 to 16, the threshold value of the inaccuracy is set to Dth = 20. That is, the estimated speed of the patch 202 corresponding to the patch 242 with an inaccuracy of 20 or less can be regarded as being accurate to a reliable extent. On the other hand, the estimated speed of the patch 202 corresponding to the patch 242 with an inaccuracy exceeding 20 can be regarded as being inaccurate to such an extent that it cannot be used for event detection.

[0082] In the estimated speed map 200 illustrated in FIG. 14, the hatched patch 202 is a patch 202 whose estimated speed is equal to or lower than the threshold value Vth. Among these, in the patch 202 indicated by the ellipse C1 in FIG. 14, a true low-speed event is occurring. That is, at the time and position corresponding to that patch 202, actually, the vehicle speed (average speed) has decreased due to a low-speed event such as a traffic jam.

[0083] Here, when a low-speed event such as traffic congestion actually occurs, as time passes, the location where the vehicle speed (average vehicle speed) becomes low on the road propagates backward (upstream side in the vehicle's traveling direction). For example, assume that the cause of traffic congestion occurs at the position corresponding to patch 202Y at the time corresponding to patch 202Y. In this case, in the estimated speed map 200, in the time direction, the speed becomes low in patch 202 corresponding to the time after that time, and in the space direction, the speed becomes low in patch 202 corresponding to the position behind that position. Therefore, when the cause of traffic congestion occurs at the position and time corresponding to patch 202Y, the low-speed event propagates to the time and position behind that patch 202Y.

[0084] Therefore, in patch 202 (for example, patch 202Z) where a true low-speed event is occurring, the estimated speed (vehicle speed) can decrease in the surrounding patches 202. And in the patch 202 indicated by ellipse C1 in FIG. 14, the estimated speed has decreased below the threshold value, which can indicate that the estimated speed has not decreased due to misestimation caused by the influence of noise or the like, but rather the estimated speed has decreased as the actual vehicle speed has decreased.

[0085] As described above, in patch 202 where a true low-speed event is occurring, not only the estimated speed of that patch 202 but also the estimated speeds of the surrounding patches 202 have decreased. Therefore, when the correction speed calculation unit 144 calculates the correction speed using the estimated speeds of the surrounding patches 202 with respect to the estimated speed of patch X where a true low-speed event is occurring, a correction speed with a low speed value can be calculated. In the correction speed map 220 illustrated in FIG. 15, in patch 222 (for example, patch 222Z) indicated by ellipse D1 corresponding to patch 202 indicated by ellipse C1 in FIG. 14, the correction speed has also decreased below the threshold value.

[0086] Also, in patch 202 where a true low-speed event is occurring, not only the estimated speed of that patch 202 but also the estimated speeds of the surrounding patches 202 are decreasing. Therefore, when the inaccuracy calculation unit 146 calculates the inaccuracy for the estimated speed of patch X where a true low-speed event is occurring using the estimated speeds of the surrounding patches 202, since the variation in these estimated speeds is small, a low inaccuracy can be calculated. In the inaccuracy map 240 illustrated in FIG. 16, for patch 242 (e.g., patch 242Z) indicated by ellipse E1 corresponding to patch 202 indicated by ellipse C1 in FIG. 14, the inaccuracy is below the threshold. Therefore, the event detection unit 150 can accurately detect that an event has occurred for patch 202 where a true low-speed event is occurring. Therefore, the event notification unit 160 can accurately report that an event has occurred for patch 202 where a true low-speed event is occurring.

[0087] On the other hand, in patches 202G, 202H, and 202I indicated by arrow C2 in FIG. 14, the estimated speed is below the threshold Vth, but in the surrounding patches 202 of these, the estimated speed has not decreased below the threshold. Therefore, in patches 202G, 202H, and 202I, it is highly likely that an event is not actually occurring. That is, the estimated speeds for patches 202G, 202H, and 202I are highly likely to have decreased due to misestimation caused by the influence of noise or the like.

[0088] Here, when the correction speed calculation unit 144 calculates the correction speed using the estimated speeds of the surrounding patches 202 with respect to the estimated speed of the patch 202G, the calculated correction speed exceeds the threshold value Vth, like the correction speed of the patch 222G in the correction speed map 220 shown in FIG. 15. Therefore, the event detection unit 150 does not determine that an event has occurred at the position and time corresponding to the patch 202G. That is, the event detection unit 150 does not detect an event for the patch 202G where no event has occurred. In this way, the event detection unit 150 can accurately perform event detection. As a result, the event notification unit 160 can be prevented from making a false alarm of reporting that an event has occurred for a patch 202 where no event has occurred.

[0089] Also, when the correction speed calculation unit 144 calculates the correction speed using the estimated speeds of the surrounding patches 202 with respect to the estimated speed of the patch 202H, the calculated correction speed exceeds the threshold value Vth, like the correction speed of the patch 222H in the correction speed map 220 shown in FIG. 15. Similarly, when the correction speed calculation unit 144 calculates the correction speed using the estimated speeds of the surrounding patches 202 with respect to the estimated speed of the patch 202I, the calculated correction speed exceeds the threshold value Vth, like the correction speed of the patch 222I in the correction speed map 220 shown in FIG. 15. Therefore, the event detection unit 150 does not determine that an event has occurred at the position and time corresponding to the patches 202H and 202I. That is, the event detection unit 150 does not detect an event for the patches 202H and 202I where no event has occurred. In this way, the event detection unit 150 can accurately perform event detection.

[0090] On the other hand, when the correction speed calculation unit 144 calculates the correction speed for the estimated speed of the patch 202J that does not exceed the threshold value, the calculated correction speed becomes equal to or less than the threshold value Vth, like the correction speed of the patch 222J in the correction speed map 220 shown in FIG. 15. This is because the estimated speeds of the patches 202H and 202I around the patch 202J have decreased due to incorrect estimation. Here, in the inaccuracy map 240 shown in FIG. 16, in the patch 242J corresponding to the patch 222J in FIG. 15, the degree of inaccuracy exceeds the threshold value (Dth = 20). Therefore, the event detection unit 150 does not determine that an event has occurred at the position and time corresponding to the patch 202J. That is, the event detection unit 150 does not detect an event for the patch 202J where no event has occurred. In this way, the event detection unit 150 can accurately perform event detection.

[0091] Also, even when the correction speed obtained for a patch 202 where no event actually occurs, such as the patch 202J, becomes equal to or less than the threshold value, by performing event detection using the degree of inaccuracy, it is possible to suppress erroneously detecting an event. Therefore, by using the correction speed and the degree of inaccuracy, it becomes possible to perform event detection more accurately. Therefore, it is possible to further suppress false alarms.

[0092] (Modification example) Note that the present invention is not limited to the above-described embodiments and can be appropriately modified without departing from the spirit thereof. For example, the order of each step (process) of the flowchart described above can be appropriately changed. Also, one or more of the steps (processes) of the flowchart can be appropriately omitted. For example, in the flowchart of FIG. 8, the order of the processes of S112 and S114 may be reversed. Alternatively, the processes of S112 and S114 may be executed in parallel.

[0093] Alternatively, only one of the processes of S112 and S114 may be executed. When only the process of S112 is executed, in the process of S120, the event detection unit 150 determines whether the corresponding correction speed is equal to or lower than the threshold value Vth for each patch 202. When the correction speed is equal to or lower than the threshold value Vth, the event detection unit 150 determines that an event has occurred at the position corresponding to the patch 202 and at the time corresponding to the patch 202. On the other hand, when the correction speed is not equal to or lower than the threshold value Vth, the event detection unit 150 determines that no event has occurred at the position corresponding to the patch 202 and at the time corresponding to the patch 202.

[0094] In contrast, when only the process of S114 is executed, in the process of S120, the event detection unit 150 determines whether the corresponding estimated speed is equal to or lower than the threshold value Vth and the corresponding inaccuracy degree is equal to or lower than the threshold value Dth for each patch 202. When this determination is true, the event detection unit 150 determines that an event has occurred at the position corresponding to the patch 202 and at the time corresponding to the patch 202. That is, in this case, the estimated speed corresponding to this patch 202 is reliable (i.e., not due to misestimation), and since the estimated speed has decreased to be equal to or lower than the threshold value Vth, it is determined that an event has occurred. On the other hand, when this determination is false, the event detection unit 150 does not determine that an event has occurred at the position corresponding to the patch 202 and at the time corresponding to the patch 202. That is, if the inaccuracy degree is not equal to or lower than the threshold value Dth, the corresponding estimated speed is not reliable, so event detection is not performed based on the unreliable estimated speed. Also, if the estimated speed is not equal to or lower than the threshold value Vth, it is highly likely that no speed reduction event has occurred, so it is not determined that an event has occurred.

[0095] Also, in the above-described embodiments, it was assumed that the patch size is uniform. However, the patch size may not be uniform. For example, the size of the patches divided in the travel locus data 500 may vary depending on the position of the patches. The same applies to the patch sizes of the estimated speed map 200, the corrected speed map 220, and the inaccuracy degree map 240.

[0096] Also, in the above-described embodiments, it was assumed that the estimated speed of a vehicle traveling on a road is obtained using a signal obtained by optical fiber sensing. However, the estimated speed of the vehicle may be obtained using a signal obtained by a method other than optical fiber sensing.

[0097] The above-described program, when loaded into a computer, includes a set of instructions (or software code) for causing the computer to perform one or more functions described in the embodiments. The program may be stored in a non-transitory computer-readable medium or a tangible storage medium. By way of example and not limitation, the computer-readable medium or tangible storage medium includes random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD), or other memory technologies, CD-ROM, digital versatile disk (DVD), Blu-ray (registered trademark) disk, or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage, or other magnetic storage devices. The program may be transmitted on a transitory computer-readable medium or a communication medium. By way of example and not limitation, the transitory computer-readable medium or communication medium includes electrical, optical, acoustic, or other forms of propagated signals.

[0098] Some or all of the above embodiments may be described as follows in the following supplementary notes, but are not limited thereto. (Supplementary Note 1) Estimation means for estimating the speed of a vehicle traveling on the road at each time at each position on the road using a signal obtained by measuring the road; Event detection means for detecting an event occurring on the road based on at least one of a corrected speed obtained by performing a smoothing process on an estimated speed which is the estimated speed, and an inaccuracy degree indicating the degree of inaccuracy of the estimated speed; A signal analysis device having the above. (Appendix 2) The event detection means detects an event when the corrected speed is less than or equal to a predetermined first threshold value. The signal analysis device according to Appendix 1. (Appendix 3) The event detection means detects an event when the corrected speed is less than or equal to the first threshold value and the inaccuracy degree is less than or equal to a predetermined second threshold value. The signal analysis device according to Appendix 2. (Appendix 4) Correction speed calculation means for calculating a correction speed related to the first estimated speed by performing a smoothing process using at least one of the estimated speed at the first time at a position near the first position and the estimated speed at the first position at a time near the first time with respect to the first estimated speed at the first time at the first position; The signal analysis device according to any one of Appendices 1 to 3, further having the above. (Appendix 5) The correction speed calculation means calculates a correction speed related to the first estimated speed by performing a smoothing process using the estimated speed at the first position at a time before the first time. The signal analysis device according to Appendix 4. (Appendix 6) Inaccuracy degree calculation means for calculating the inaccuracy degree using at least one of the estimated speed at the first time at a position near the first position and the estimated speed at the first position at a time near the first time with respect to the first estimated speed at the first time at the first position; The signal analysis device according to any one of Appendices 1 to 5, further comprising (Appendix 7) The inaccuracy calculation means calculates the inaccuracy using the estimated speed of the first position at a time before the first time. The signal analysis device according to Appendix 6. (Appendix 8) The inaccuracy calculation means calculates, as the inaccuracy, an index representing the variation between the first estimated speed and at least one of the estimated speed of a position near the first position at the first time and the estimated speed of the first position at a time near the first time. The signal analysis device according to Appendix 6 or 7. (Appendix 9) The inaccuracy calculation means calculates the inaccuracy such that the greater the variation among the first estimated speed, the estimated speed of a position near the first position at the first time, and the estimated speed of the first position at a time near the first time, the greater the inaccuracy of the first estimated speed. The signal analysis device according to Appendix 6 or 7. (Appendix 10) The estimation means estimates the speed of the vehicle using a signal detected using an optical fiber provided along the road. The signal analysis device according to any one of Appendices 1 to 9. (Appendix 11) Using a signal obtained by measuring a road, estimate the speed of a vehicle traveling on the road at each position and each time on the road. Detect an event occurring on the road based on at least one of a corrected speed obtained by performing a smoothing process on an estimated speed that is the estimated speed and an inaccuracy indicating the degree of inaccuracy of the estimated speed. Signal analysis method. (Appendix 12) When the corrected speed is equal to or less than a predetermined first threshold value, detect an event. The signal analysis method according to Appendix 11. (Appendix 13) When the correction speed is equal to or less than the first threshold value and the inaccuracy degree is equal to or less than a predetermined second threshold value, an event is detected. The signal analysis method according to Appendix 12. (Appendix 14) For the first estimated speed at the first time at the first position, by performing a smoothing process using at least one of the estimated speeds at the first time at positions near the first position and the estimated speeds at the first position at times near the first time, a correction speed regarding the first estimated speed is calculated. The signal analysis method according to any one of Appendices 11 to 13. (Appendix 15) By performing a smoothing process using the estimated speed at the first position at a time before the first time, a correction speed regarding the first estimated speed is calculated. The signal analysis method according to Appendix 14. (Appendix 16) For the first estimated speed at the first time at the first position, the inaccuracy degree is calculated using at least one of the estimated speeds at the first time at positions near the first position and the estimated speeds at the first position at times near the first time. The signal analysis method according to any one of Appendices 11 to 15. (Appendix 17) The inaccuracy degree is calculated using the estimated speed at the first position at a time before the first time. The signal analysis method according to Appendix 16. (Appendix 18) An index representing the variation between the first estimated speed and at least one of the estimated speed at the first time at positions near the first position and the estimated speeds at the first position at times near the first time is calculated as the inaccuracy degree. The signal analysis method according to Appendix 16 or 17. (Appendix 19) The inaccuracy degree is calculated such that the greater the variation in the first estimated speed, the estimated speed at the position near the first position at the first time, and the estimated speed of the first position at the time near the first time, the greater the inaccuracy degree of the first estimated speed. The signal analysis method according to Appendix 16 or 17. (Appendix 20) Using the signal detected by using the optical fiber provided along the road to estimate the speed of the vehicle. The signal analysis method according to any one of Appendices 11 to 19. (Appendix 21) Estimating the speed of a vehicle traveling on the road at each time at each position on the road by using a signal obtained by measuring the road; Detecting an event occurring on the road based on at least one of a corrected speed obtained by performing a smoothing process on an estimated speed that is the estimated speed, and an inaccuracy degree indicating the degree of inaccuracy of the estimated speed. A non-transitory computer-readable medium storing a program for causing a computer to execute the above.

Explanation of Signs

[0099] 1 Signal analysis device 2 Estimation unit 4 Event detection unit 10 Signal analysis system 50 Optical fiber sensing system 52 Sensing device 54 Optical fiber cable 80 Road 92 Analysis engine 100 Signal analysis device 110 Signal acquisition unit 120 Trajectory acquisition unit 130 Speed estimation unit 140 Estimated speed processing unit 142 Estimated speed data storage unit 144 Corrected speed calculation unit 146 Inaccuracy degree calculation unit 150 Event detection unit 160 Event notification unit 200 Estimated speed map 202 Patch 220 Corrected speed map 222 Patch 240 Inaccuracy map 242 Patch 500 Travel trajectory data 502 Patch

Claims

1. Estimation means for estimating the speed of a vehicle traveling on the road at each of a plurality of times corresponding to a predetermined range of time along the time axis for each of a plurality of positions corresponding to a spatial region of a predetermined range along the road on the road, using signals obtained by measuring the road; Correction speed calculation means for calculating a correction speed related to the first estimated speed by performing a smoothing process using the estimated speed at the first time of the first position among the plurality of positions, the estimated speed at at least one second position near the first position at the first time among the plurality of positions, and the estimated speed of the first position at at least one second time near the first time among the plurality of times; Inaccuracy calculation means for calculating an inaccuracy indicating the degree of inaccuracy of the first estimated speed using the estimated speed at the first time of the second position and the estimated speed of the first position at the second time with respect to the first estimated speed; Event detection means for detecting an event occurring on the road based on at least one of the correction speed obtained by performing a smoothing process on the first estimated speed and the inaccuracy indicating the degree of inaccuracy of the first estimated speed; A signal analysis device having the above.

2. The event detection means detects an event when the correction speed is less than or equal to a predetermined first threshold value. The signal analysis device according to Claim 1.

3. The event detection means detects an event when the correction speed is less than or equal to the first threshold value and the inaccuracy is less than or equal to a predetermined second threshold value. The signal analysis device according to Claim 2.

4. The correction speed calculation means calculates a correction speed related to the first estimated speed by performing a smoothing process using the estimated speed of the first position at the second time which is a time before the first time. The signal analysis device according to Claim 1.

5. The inaccuracy calculation means calculates the inaccuracy using the estimated speed of the first position at the second time which is a time before the first time. The signal analysis device according to Claim 1.

6. The inaccuracy calculation means calculates, as the degree of inaccuracy, an index representing the variation between the first estimated speed and at least one of the estimated speed at the first time of the second position near the first position and the estimated speed of the first position at the second time near the first time. The signal analysis device according to any one of claims 1 to 5. **Claim 7** Using signals obtained by measuring a road, estimate the speed of a vehicle traveling on the road at each of a plurality of times corresponding to a predetermined time range along the time axis for each of a plurality of positions corresponding to a predetermined space range along the road on the road. With respect to the first estimated speed, which is the speed estimated at the first time of the first position among the plurality of positions, smoothing processing is performed using the estimated speed at the first time of at least one second position near the first position among the plurality of positions and the estimated speed of the first position at at least one second time near the first time among the plurality of times, thereby calculating a corrected speed regarding the first estimated speed. Using the estimated speed at the first time of the second position and the estimated speed of the first position at the second time with respect to the first estimated speed, calculate the degree of inaccuracy indicating the degree of inaccuracy of the first estimated speed. Detect an event occurring on the road based on at least one of the corrected speed obtained by performing smoothing processing on the first estimated speed and the degree of inaccuracy indicating the degree of inaccuracy of the first estimated speed. A signal analysis method executed by a computer. **Claim 8** A step of estimating the speed of a vehicle traveling on the road at each of a plurality of times corresponding to a predetermined time range along the time axis for each of a plurality of positions corresponding to a predetermined space range along the road on the road, using signals obtained by measuring the road. A step of calculating a correction speed related to the first estimated speed by performing a smoothing process using the estimated speed at a first time of a first position among the plurality of positions, the estimated speed at at least one second position in the vicinity of the first position at the plurality of positions at the first time, and the estimated speed at the first position at at least one second time in the vicinity of the first time among the plurality of times. A step of calculating an inaccuracy degree indicating the degree of inaccuracy of the first estimated speed by using the estimated speed at the first time of the second position and the estimated speed at the first position at the second time with respect to the first estimated speed. A step of detecting an event that occurred on the road based on at least one of the correction speed obtained by performing a smoothing process on the first estimated speed and the inaccuracy degree indicating the degree of inaccuracy of the first estimated speed. A program for causing a computer to execute.

Images