Detection system, sensor system, detection method, and program recording medium

Abstract

The detection system (10) has an acquisition unit (11), a generation unit (12), a classification unit (13), a center-of-gravity calculation unit (14), a correlation unit (15), and a mixing unit (16). The acquisition unit (11) acquires a reception signal (Sig1) from a radio wave sensor (2). The generation unit (12) generates a scatter point including position information of an object (3) based on the reception signal (Sig1). The classification unit (13) classifies a set of scatter points, i.e., a scatter point group, generated by the generation unit (12) over a prescribed time into one or more clusters. The center-of-gravity calculation unit (14) calculates a center of gravity for each of the one or more clusters classified by the classification unit (13). The correlation unit (15) correlates a certain one of the one or more clusters classified by the classification unit (13) with the object (3). The mixing unit (16) mixes the center of gravity in the previous cycle and the cycles before the previous cycle calculated by the center-of-gravity calculation unit (14) into the scatter point group as a scatter point.

Description

Technical Field

[0001] This invention relates to a detection system, sensor system, detection method, and program recording medium for detecting objects based on signals output from an electromagnetic wave sensor. Background Technology

[0002] Patent Document 1 discloses a Doppler radar system for detecting the position of a target object. This Doppler radar system includes a transmitting antenna, multiple receiving antennas, a micro-Doppler arithmetic unit, a first centroid calculation unit, a second centroid calculation unit, and a centroid determination unit. The micro-Doppler arithmetic unit calculates multiple measurement points representing the positions of multiple reflection points included in the target object. The first centroid calculation unit calculates the first centroid of the measurement points using a first dividing frame having a first interval. The second centroid calculation unit calculates the second centroid of the measurement points using a second dividing frame having a second interval narrower than the first interval. The centroid determination unit determines at least one second centroid as the position of the target object from among the multiple second centroids based on the positional relationship between the second centroid and the first centroid.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2013-96828 Summary of the Invention

[0006] The problem the invention aims to solve

[0007] This invention provides a detection system, sensor system, detection method, and program recording medium that can easily improve the detection accuracy of the position of an object.

[0008] Solution for solving the problem

[0009] One aspect of the present invention relates to a detection system comprising an acquisition unit, a generation unit, a classification unit, a centroid calculation unit, an association unit, and a mixing unit. The acquisition unit acquires received signals from an electromagnetic wave sensor, which radiates electromagnetic waves from a transmitting antenna toward an object and receives reflected electromagnetic waves from the object via multiple receiving antennas. The generation unit generates scattered points containing location information of the object based on the received signals acquired by the acquisition unit. The classification unit classifies the set of scattered points generated by the generation unit over a predetermined time period, i.e., a scattered point group, into one or more clusters. The centroid calculation unit calculates the centroid for each of the one or more clusters obtained by the classification unit. The association unit associates one of the one or more clusters obtained by the classification unit with the object. The mixing unit mixes the centroids calculated by the centroid calculation unit for the previous period and earlier into the scattered point group.

[0010] One aspect of the present invention relates to a sensor system comprising the detection system and the radio wave sensor controlled by the detection system.

[0011] One aspect of the present invention relates to a detection method comprising an acquisition step, a generation step, a classification step, a centroid calculation step, an association step, and a mixing step. In the acquisition step, a received signal is acquired from an electromagnetic wave sensor, which radiates electromagnetic waves from a transmitting antenna toward an object and receives reflected electromagnetic waves from the object through multiple receiving antennas. In the generation step, scattered points containing location information of the object are generated based on the received signal acquired in the acquisition step. In the classification step, the set of scattered points generated over a predetermined time period in the generation step, i.e., a scattered point group, is classified into one or more clusters. In the centroid calculation step, a centroid is calculated for each of the one or more clusters classified in the classification step. In the association step, one of the one or more clusters classified in the classification step is associated with the object. In the mixing step, the centroids calculated in the centroid calculation step for the previous period and earlier periods are mixed into the scattered point group as scattered points.

[0012] One aspect of the present invention involves a program recording medium that records a program for causing one or more processors to execute the detection method.

[0013] The effects of the invention

[0014] The detection system, sensor system, detection method, and program recording medium of the present invention have the advantage of easily improving the detection accuracy of the position of the object. Attached Figure Description

[0015] Figure 1 This is a block diagram illustrating the overall structure of the detection system involved in the embodiments.

[0016] Figure 2 This is a diagram illustrating an outline of object detection performed by the radio wave sensor according to the embodiment.

[0017] Figure 3 This is a diagram showing an example of scattered points generated by the generation unit of the detection system according to the embodiment.

[0018] Figure 4 This is an explanatory diagram of a scattered point group, which is the classification unit of the detection system involved in the implementation method as the classification object.

[0019] Figure 5 This is a diagram showing an outline of the classification section of the detection system according to the embodiment, which is a group of scattered points as the classification object.

[0020] Figure 6 This diagram illustrates a basic example of the association between an object and a cluster performed by the association unit of the detection system according to the embodiment.

[0021] Figure 7 This is an explanatory diagram illustrating the association between objects and clusters performed by the association unit of the detection system involved in the implementation method.

[0022] Figure 8 This is an explanatory diagram of the first operating example of the hybrid section of the detection system according to the embodiment.

[0023] Figure 9 This is an explanatory diagram of cluster migration in the first operational example of the mixing section of the detection system according to the embodiment.

[0024] Figure 10 This is an explanatory diagram of a second operating example of the hybrid section of the detection system according to the embodiment.

[0025] Figure 11 This is an explanatory diagram of the third operation example of the hybrid unit of the detection system according to the embodiment.

[0026] Figure 12 This is a flowchart illustrating an example of the operation of the detection system involved in the implementation.

[0027] Figure 13A This is a diagram illustrating an example of the detection results of the comparative detection system.

[0028] Figure 13B This is a diagram illustrating an example of the detection results of the detection system involved in the embodiment.

[0029] Figure 14 This is a block diagram illustrating the overall structure of the detection system involved in the variations of the embodiments.

[0030] Figure 15 This is a diagram showing an outline of a scattered point group as a classification object in the classification section of a detection system involved in a variation of the embodiment. Detailed Implementation

[0031] (Insights that form the basis of this invention)

[0032] First, the inventor's perspective will be explained below.

[0033] Previously, there was a need in spaces where people reside, such as residences or offices, to detect the presence of people without using cameras in order to protect privacy. Furthermore, to meet this need, it is known to use radio wave sensors to detect the presence of people in a space.

[0034] Radio wave sensors can not only detect people in space, but also acquire biometric information (such as breathing or heart rate) of the detected people in a non-contact manner. If biometric information can be acquired non-contactly using radio wave sensors, it has the following advantages compared to acquiring biometric information through wearable devices such as wristwatches: it liberates the person from the awareness of being detected, and the person is less likely to feel the inconvenience or pressure of wearing such devices.

[0035] Furthermore, to acquire biometric information of multiple individuals simultaneously using a single radio wave sensor, signal separation techniques such as beamforming can be considered. When using such signal separation techniques, accurately determining the source of the biometric information—that is, the location of each individual—is crucial. In particular, for high-precision acquisition of biometric information, stable location detection for each individual is essential.

[0036] In view of the above circumstances, the inventors ultimately created this invention.

[0037] The embodiments will now be described in detail with reference to the accompanying drawings. Furthermore, the embodiments described below are general or specific examples. The numerical values, shapes, materials, constituent elements, the arrangement and connection methods of constituent elements, steps, and the order of steps shown in the following embodiments are examples and are not intended to limit the invention. Additionally, constituent elements in the following embodiments that are not described in the independent claims are described as optional constituent elements.

[0038] Furthermore, the figures are schematic diagrams and not necessarily strictly illustrated. Additionally, substantially identical structures are labeled with the same reference numerals across the figures, and sometimes repeated descriptions are omitted or simplified.

[0039] (Implementation Method)

[0040] [1. Structure]

[0041] Next, use Figure 1 The structure of the detection system 10 involved in the implementation method will be explained. Figure 1 This is a block diagram showing the overall structure of the detection system 10 according to the embodiment. The detection system 10 is a system for detecting an object 3 based on a signal (received signal Sig1) output from the radio wave sensor 2. In the embodiment, the object 3 is a person. Furthermore, if the method is to reflect the radio waves radiated by the radio wave sensor 2, the object 3 may also be a living organism other than a person, or it may not be a living organism at all.

[0042] The radio wave sensor 2 radiates radio waves from the transmitting antenna 211 towards the object 3, and the reflected radio waves from the object 3 are received by multiple receiving antennas 221. Specifically, the radio wave sensor 2 is, for example, a SIMO (single-input and multiple-output) radar, which includes a transmitting unit 21, a receiving unit 22, a transmitting antenna 211, and multiple receiving antennas 221. The multiple receiving antennas 221 constitute an array antenna. Alternatively, the radio wave sensor 2 can also be a MIMO (multiple-input and multiple-output) radar with multiple transmitting antennas 211.

[0043] The transmitting unit 21 generates a transmission signal and outputs the generated transmission signal to the transmitting antenna 211. The transmitting unit 21 is controlled, for example, by the detection system 10. When the transmitting antenna 211 receives the transmission signal output from the transmitting unit 21, it radiates the transmission signal in the form of radio waves.

[0044] Here, if the object 3 has a complex shape with uneven surfaces, such as a person, and the wavelength of the radio wave is longer than that of the radio wave, then multiple points (i.e., reflection points) satisfying the reflection conditions exist on the surface of the object 3. Multiple receiving antennas 221 respectively receive the reflected radio waves transmitted from the transmitting antenna 211 after being reflected by the multiple reflection points on the surface of the object 3. The receiving unit 22 generates a receiving signal Sig1 corresponding to each of the multiple reflected radio waves received by the multiple receiving antennas 221, and outputs the generated receiving signal Sig1 to the detection system 10.

[0045] In this implementation, the radio wave sensor 2 can switch between any of two modes: a normal mode and a specific mode, whereby the specific mode involves concentrated radiation or selective reception of radio waves in a specific direction. The normal mode is used to detect the position of one or more objects 3 in a space by radiating radio waves across the entire space being detected using the radio wave sensor 2. The specific mode is used to acquire biological information (e.g., respiration or heart rate) of a specific object 3 detected in the normal mode by using beamforming to concentrate radiation or selective reception of radio waves.

[0046] The detection system 10 is implemented, for example, by a processing circuit equipped with a processor and memory, such as a microcomputer. That is, the processor in the processing circuit executes a program stored in memory to perform various functions as the detection system 10. Figure 1 As shown, the detection system 10 includes an acquisition unit 11, a generation unit 12, a classification unit 13, a center of gravity calculation unit 14, a correlation unit 15, and a mixing unit 16.

[0047] The acquisition unit 11 acquires the received signal Sig1 from the radio wave sensor 2. The acquisition unit 11 is the execution body for the acquisition step ST1 in the detection method. In this embodiment, the acquisition unit 11 is connected to the output terminal of the radio wave sensor 2 via a signal line. Furthermore, the acquisition unit 11 periodically acquires the received signal Sig1 output from the output terminal of the radio wave sensor 2 via the signal line.

[0048] The generation unit 12 generates scattered points 41 containing the position information of the object 3 based on the received signal Sig1 acquired by the acquisition unit 11. The generation unit 12 is the execution body of the generation step ST2 in the detection method. Here, "generating scattered points 41" means calculating (estimating) the reflection points on the surface of the object 3.

[0049] Specifically, the generation unit 12 calculates the distance D1 between the object 3 and the radio wave sensor 2 (see reference) by performing processing based on the received signal Sig1 using frequency-modulated continuous wave radar technology. Figure 2 Additionally, the generation unit 12 calculates the arrival angle θ1 of the reflected radio wave (see reference) by performing a process based on the received signal Sig1 to estimate the direction of arrival using the reflected radio wave (i.e., the received signal Sig1) received by the array antenna (multiple receiving antennas 221). Figure 2 That is, the direction of the object 3 relative to the radio wave sensor 2. Furthermore, the generation unit 12 measures the intensity of the reflected radio wave, for example, by measuring the RSSI (Received Signal Strength Indicator) value of the received signal Sig1, or by extracting peak values ​​from the spectrum (called the range spectrum) obtained by performing a Fast Fourier Transform on the received signal Sig1 according to the principle of frequency-modulated continuous wave radar. In other words, the scattered points 41 contain information related to at least one of the following (both in this embodiment): the distance between the object 3 and the radio wave sensor 2, and the direction of the object 3 relative to the radio wave sensor 2, as well as information related to the intensity of the reflected radio wave.

[0050] The following describes a specific example of the generation of scattered points 41 by the generation unit 12. Figure 2 This is a diagram showing an outline of the detection of an object 3 by the radio wave sensor 2 according to the embodiment. Figure 2 The curved, dashed line in the image represents the radio waves transmitted from radio wave sensor 2. Figure 2 In the example shown, within the detection range of the radio wave sensor 2, there are three people (first object 3A, second object 3B, and third object 3C) as object 3.

[0051] The generation unit 12 generates scattered points 41 by performing the aforementioned processing based on the received signal Sig1. Specifically, regarding the first object 3A, the generation unit 12 generates a first scattered point 41A with a distance D1A and an arrival angle θ1A. Furthermore, regarding the second object 3B, the generation unit 12 generates a second scattered point 41B with a distance D1B and an arrival angle θ1B. Furthermore, regarding the third object 3C, the generation unit 12 generates a third scattered point 41C with a distance D1C and an arrival angle θ1C. Additionally, in... Figure 2 In the example shown, the first scattered point 41A of the first object 3A, the second scattered point 41B of the second object 3B, and the third scattered point 41C of the third object 3C are all single points, but in reality, there may be a large number of points. Furthermore, the intensity of the reflected radio waves from each scattered point 41A to 41C is omitted here.

[0052] Figure 3 This is a diagram showing an example of scattered points 41 generated by the generation unit 12 of the detection system 10 according to the embodiment. Figure 3 The circular symbols shown correspond to the scattered points 41 generated in the generation unit 12, representing the planar coordinates (xy coordinates) of the scattered points 41 with the radio wave sensor 2 as the origin in the plane containing the radio wave sensor 2 and each object 3. Figure 3 In the example shown, the location of set 4A of the first scattered points 41A roughly corresponds to the location of the first object 3A. Similarly, the location of set 4B of the second scattered points 41B roughly corresponds to the location of the second object 3B. Furthermore, the location of set 4C of the third scattered points 41C roughly corresponds to the location of the third object 3C.

[0053] Classification unit 13 classifies the set of scattered points 41 generated by generation unit 12 over a specified time T1, i.e., scattered point group 4, into one or more clusters 5 (see reference). Figure 5 Classification section 13 is the main body responsible for executing the classification step ST3 in the detection method. For example, in the scattered point group 4, such as... Figure 3 In the case shown, which includes set 4A of the first scattered point 41A, set 4B of the second scattered point 41B, and set 4C of the third scattered point 41C, the classification unit 13 classifies these sets 4A to 4C into three clusters 5 respectively.

[0054] Figure 4 This is an explanatory diagram of the classification unit 13 of the detection system 10 according to the embodiment, which is a group of scattered points 4 as the classification object. Figure 4In the example shown, "#1" to "#n (n is a natural number greater than 2)," "#n+1," and "#n+2" represent the sets of scattered points 41 generated by the generation unit 12 in each frame (e.g., a fraction of a second). Furthermore, the classification unit 13 sets the set of scattered points 41 accumulated over a specified time T1 (here, n frames) into a scattered point group 4, and performs clustering on this scattered point group 4 using a specified algorithm.

[0055] exist Figure 4 In the example shown, "#1*" represents a set of scattered points 4 accumulated in n frames from "#1" to "#n". Similarly, "#2*" represents a set of scattered points 41 accumulated in n frames from "#2" to "#n+1". Furthermore, "#3*" represents a set of scattered points 41 accumulated in n frames from "#3" to "#n+2". The classification unit 13 classifies each of these scattered point sets 4 ("#1*", "#2*", and "#3*") into one or more clusters 5. That is, the classification unit 13 classifies the scattered point sets 4 into one or more clusters 5 every period P1 (one frame).

[0056] In this implementation, the classification unit 13 uses an algorithm based on the density of scattered points 41 to remove isolated points, thereby classifying the scattered point group 4 into more than one cluster 5. Specifically, the classification unit 13 uses DBSCAN (Density-Based Spatial Clustering of Applications with Noise) to classify the scattered point group 4 into more than one cluster 5.

[0057] In DBSCAN, if any random point 41 is a core point, then this core point and all reachable points reachable from this core point form a cluster 5. Furthermore, in DBSCAN, random points 41 that do not correspond to either a core point or reachable points are outliers (isolated points). Isolated points are not classified into cluster 5 (i.e., they are removed). Here, a "core point" is defined as any "point p" that has at least "minPts" points within a "distance ε" from that "point p". Additionally, "reachable points" include points that are directly reachable from the core point because they exist within a "distance ε" from the core point, and points that, although not existing within a "distance ε" from the core point, are reachable from the core point via other core points. Both "distance ε" and "minPts" are parameters that can be appropriately set.

[0058] Figure 5 This is a diagram showing an outline of the classification unit 13 of the detection system 10 according to the embodiment, which is a group of scattered points 4 for classification. Figure 5 In the example shown, the area enclosed by the dashed circle represents cluster 5, which was classified by classification department 13, and the circle symbol represents scattered points 41. Additionally, in Figure 5 In the example shown, the × symbol indicates that the centroid 42 in the previous and preceding periods P1 calculated by the centroid calculation unit 14 (described later) is included as a scattered point 41 in the scattered point group 4. Furthermore, either only one centroid 42 may be included as a scattered point 41 in the scattered point group 4, or multiple centroids 42 may be included as scattered points 41 in the scattered point group. When the scattered point group 4 contains multiple centroids 42, the multiple centroids 42 overlap at one location.

[0059] Here, "the previous period P1" refers to one or more periods P1 preceding the current period P1, where the period P1 in which the classification department 13 classifies the scattered point group 4 into one or more clusters 5 is taken as the current period P1. Figure 5 In the example shown, the area enclosed by the dashed circle represents cluster 5 in the current period P1, the circle symbol represents the scattered point 41 in the current period P1, and the × symbol represents the centroid 42 in the previous period P1.

[0060] Furthermore, at the point in time when the center of gravity calculation unit 14 has not yet calculated the center of gravity 42, the center of gravity 42 will naturally not be included in the scattered point group 4. That is to say, in the scattered point group 4, the center of gravity 42 is included as a scattered point 41 in the period P1 after the center of gravity calculation unit 14 calculates the period P1 of the center of gravity 42.

[0061] The centroid calculation unit 14 calculates the centroid 42 for each of the more than one clusters 5 classified by the classification unit 13. The centroid calculation unit 14 is the execution body of the centroid calculation step ST4 in the detection method. In the embodiment, the centroid calculation unit 14 calculates the plane coordinates (xy coordinates) of the centroid 42 for each of the more than one clusters 5 using the plane coordinates (xy coordinates) of each scattered point 41 (including the centroid 42 of the previous period P1) contained in the cluster 5 and the intensity of the reflected radio wave of each scattered point 41, using the following equations (1) and (2). The following equation (1) is the formula for calculating the x coordinate of the centroid 42, and equation (2) is the formula for calculating the y coordinate of the centroid 42. In either of the following equations (1) and (2), "p" represents the intensity of the reflected radio wave.

[0062] [Number 1]

[0063]

[0064] [Number 2]

[0065]

[0066] The association unit 15 associates one of the more than one clusters 5 obtained by the classification unit 13 with the object 3. The association unit 15 is the execution body of the association step ST5 in the detection method. By associating one or more clusters 5 with the object 3 by the association unit 15, the position of the associated cluster 5 is detected as the position of the object 3.

[0067] Figure 6 This is a diagram illustrating a basic example of the association between the object 3 and the cluster 5 performed by the association unit 15 of the detection system 10 according to the embodiment. Figure 6 The results of classification by the classification unit 13 are shown when two people (objects 3) are present within the detection range of the radio wave sensor 2. Furthermore, in Figure 6 In the example shown, we assume that two people (object 3) are stationary and have hardly moved.

[0068] exist Figure 6 In the expression "Cls(a, b)", "a" represents the cluster number arbitrarily assigned by taxonomy 13, and "b" represents period P1. Figure 6 In the example shown, the lower left cluster 5, Cls(1, n-2), in the (n-2)th period P1, the lower left cluster 5, Cls(2, n-1), in the (n-1)th period P1, and the lower left cluster 5, Cls(2, n), in the nth period P1 are all in approximately the same position. Therefore, the association unit 15 associates the lower left cluster 5 with one of the two people (object 3).

[0069] In addition, Figure 6 In the example shown, the upper right cluster 5, Cls(2, n-2), in the (n-2)th period P1, the upper right cluster 5, Cls(3, n-1), in the (n-1)th period P1, and the upper right cluster 5, Cls(1, n), in the nth period P1 are all in approximately the same position. Therefore, the association unit 15 associates the upper right cluster 5 with one of the two people (object 3).

[0070] In addition, Figure 6 In the example shown, the upper left cluster 5, Cls(1, n-1), in the (n-1)th period P1 does not exist in the (n-2)th period P1 and the nth period P1. Therefore, the association unit 15 does not associate this cluster 5 with the object 3.

[0071] In the implementation, the association unit 15 associates one of the more than one clusters 5 with the object 3 based on whether the centroid 42 in the previous and earlier cycles P1 is included in one or more clusters 5 classified by the classification unit 13.

[0072] Figure 7This is an explanatory diagram illustrating the association between the object 3 and the cluster 5 by the association unit 15 of the detection system 10 according to the embodiment. Figure 7 The diagrams illustrate the process by which the classification unit 13 classifies scattered point groups 4 into one or more clusters 5 in period P1, and the centroid calculation unit 14 calculates the centroid 42 for each of the more than one clusters 5. Furthermore, Figure 7 The area enclosed by the dashed ellipse shown represents the process of determining whether cluster 5, Cls(α, n), in the current period P1 contains the centroid 42, G(β, n-1), of cluster 5, Cls(β, n-1), in the previous period P1. Furthermore, "α" and "β" represent the cluster numbers arbitrarily assigned to cluster 5 by the classification unit 13, and "n" and "n-1" represent the period P1.

[0073] When Cls(α, n) contains G(β, n-1), the association unit 15 determines that cluster 5 in the current period P1, i.e., Cls(α, n), originates from the same object 3 as cluster 5 in the previous period P1, i.e., Cls(β, n-1). In other words, the association unit 15 can capture object 3 by associating the current cluster 5 with the past cluster 5. Furthermore, when Cls(α, n) does not contain G(β, n-1), the association unit 15 determines that cluster 5 in the current period P1, i.e., Cls(α, n), originates from different object 3 as cluster 5 in the previous period P1, i.e., Cls(β, n-1).

[0074] The mixing unit 16 mixes (includes) the centroid 42 calculated by the centroid calculation unit 14 in the previous and preceding periods P1 as scattered points 41 into the scattered point group 4. The mixing unit 16 is the execution body of the mixing step ST6 in the detection method. (As already described...) Figure 5 As shown, the mixing unit 16 mixes one or more centroids 42 calculated by the centroid calculation unit 14 in the previous and previous cycles P1 as one or more scattered points 41 into the set of scattered points 41 generated by the generation unit 12 in the current cycle P1.

[0075] In this embodiment, the mixing unit 16 adjusts the mixing amount or delay amount of the centroids 42 mixed into the previous and preceding cycles P1 of the scattered point group 4. Here, "mixing amount of centroids 42" refers to the number of centroids 42 mixed into the set of scattered points 41. That is, the mixing unit 16 adjusts whether to mix one centroid 42 or multiple centroids 42 into the set of scattered points 41 in the current cycle P1.

[0076] Furthermore, "the delay of the center of gravity 42" means that the center of gravity 42 of the set of scattered points 41 is the center of gravity 42 of the previous period P1. In other words, the mixing unit 16 adjusts whether the set of scattered points 41 in the current period P1 is mixed with the center of gravity 42 of the previous period P1, or further mixed with the center of gravity 42 of the previous period P1.

[0077] By adjusting the mixing amount or delay amount of the centroid 42 in the previous and previous cycles P1 of the scattered point group 4 by the mixing unit 16, it is easy to associate the current cluster 5 with the past cluster 5, or to smooth the position change of the cluster 5 over time.

[0078] In particular, in this embodiment, the classification unit 13 uses DBSCAN to classify the scattered point group 4 into more than one cluster 5. Therefore, the mixing unit 16 can prevent the cluster 5 from disappearing by adjusting the number (mixing amount) of centroids 42 contained in the scattered point group 4 so that the core points do not disappear (that is, so that the core points do not become isolated points), and as a result, it can continuously capture the object 3.

[0079] Below, examples of the first to third operations of the mixing unit 16 are listed. In the embodiment, the mixing unit 16 applies either the first or second operation example, as well as the third operation example, but it may also apply only one operation example. Furthermore, the operation example of the mixing unit 16 shown below is only one example, and the mixing unit 16 may apply other operation examples.

[0080] In the first example, for each of the more than one clusters 5, if it does not contain any scattered points 41 other than the centroid 42, the mixing unit 16 increments the counter; if it contains any scattered points 41 other than the centroid 42, the mixing unit 16 resets the counter. Furthermore, if the counter reaches a threshold (first threshold) Th1, the mixing unit 16 discards the centroid 42 (that is, it does not include the centroid 42 as a scattered point 41 in the scattered point group 4). Thus, clusters 5 that have continuously remained without scattered points 41 are eliminated when the counter reaches the threshold Th1.

[0081] Next, use Figure 8 and Figure 9 Let's illustrate the first action example. Figure 8 This is an explanatory diagram of the first operation example of the mixing unit 16 of the detection system 10 according to the embodiment. Figure 9 This is an explanatory diagram of the migration of cluster 5 in the first operation example of the mixing unit 16 of the detection system 10 according to the embodiment. Figure 8 (a), (b), (c) and (d) in the text are respectively related to Figure 9 The corresponding terms are (a), (b), (c), and (d). Figure 9In the diagram, the area enclosed by a solid circle represents a cluster 5 containing both scattered point 41 and centroid 42, while the area enclosed by a dashed circle represents a cluster 5 containing centroid 42 but not scattered point 41. Furthermore, in... Figure 9 The diagram of scattered point 41 is omitted.

[0082] In state (a), the classification unit 13 classifies the scattered point group 4 into three clusters 5 (first cluster 5A, second cluster 5B, and third cluster 5C). In this state, each cluster 5A to 5C contains scattered points 41. Therefore, the mixing unit 16 prevents the counters corresponding to any cluster 5A to 5C from incrementing.

[0083] In state (b), the first cluster 5A still contains scattered points 41, but the scattered points 41 disappear in the second cluster 5B and the third cluster 5C. Therefore, the mixing unit 16 increments the counters corresponding to the second cluster 5B and the third cluster 5C respectively. Furthermore, the counters increment by P1 per cycle.

[0084] In state (c), the scattered points 41 that temporarily disappeared in the second cluster 5B are restored, while the scattered points 41 still disappear in the third cluster 5C. Therefore, the mixing unit 16 resets the counter corresponding to the second cluster 5B, and on the other hand, causes the counter corresponding to the third cluster 5C to continue incrementing.

[0085] In state (d), the counter corresponding to the third cluster 5C reaches the threshold (first threshold) Th1. Therefore, the mixing unit 16 adjusts the mixing amount of the centroid 42 to discard the centroid 42 corresponding to the third cluster 5C, that is, to use the centroid 42 not contained in the third cluster 5C as the scatter point 41. As a result, there is no scatter point 41 and centroid 42 in the third cluster 5C, and the third cluster 5C disappears.

[0086] As described above, the first action example has the following advantages: it can prevent the cluster 5 from remaining even though there is no scattered point 41, that is, no object 3, thereby easily improving the detection accuracy of object 3.

[0087] In the second action example, for each cluster 5 in more than one cluster 5, if it does not contain scattered points 41 other than the centroid 42, the mixing unit 16 reduces the mixing amount of the centroid 42; if it contains scattered points 41 other than the centroid 42, the mixing unit 16 restores the mixing amount of the centroid 42. Thus, clusters 5 that have always been absent from scattered points 41 are eliminated as the core points disappear when the number of centroids 42 falls below the threshold (second threshold) Th2 corresponding to "minPts" in DBSCAN due to the reduction in the mixing amount of the centroids 42.

[0088] Next, use Figure 9 and Figure 10Let's illustrate the second action example. Figure 10 This is an explanatory diagram of a second operation example of the mixing unit 16 of the detection system 10 according to the embodiment. Figure 10 (a), (b), (c) and (d) in the text are respectively related to Figure 9 The corresponding terms are (a), (b), (c), and (d).

[0089] In state (a), each cluster 5A to 5C contains scattered points 41. Therefore, the mixing unit 16 ensures that the mixing amount of the centroid 42 corresponding to any cluster 5A to 5C does not decrease.

[0090] In state (b), the first cluster 5A still contains scattered points 41, but the scattered points 41 disappear in the second cluster 5B and the third cluster 5C. Therefore, the mixing section 16 reduces the mixing amount of the centroids 42 corresponding to the second cluster 5B and the third cluster 5C, respectively. Furthermore, the mixing amount of the centroids 42 decreases every P1 cycle.

[0091] In state (c), the scattered points 41 that temporarily disappeared in the second cluster 5B are restored, while the scattered points 41 still disappear in the third cluster 5C. Therefore, the mixing unit 16 restores the mixing amount of the centroid 42 corresponding to the second cluster 5B, while continuing to reduce the mixing amount of the centroid 42 corresponding to the third cluster 5C.

[0092] In state (d), the mixing amount of the centroid 42 corresponding to the third cluster 5C reaches the threshold (second threshold) Th2. Therefore, the core point disappears in the third cluster 5C, and the third cluster 5C disappears.

[0093] As described above, the second action example has the following advantages: it can prevent the cluster 5 from remaining even though there is no scattered point 41, that is, no object 3, thereby easily improving the detection accuracy of object 3.

[0094] In the third example of operation, when the radio wave sensor 2 operates in a specific mode, compared to when the radio wave sensor 2 operates in a normal mode, the mixing unit 16 increases the mixing amount of the center of gravity 42. As a result, when the radio wave sensor 2 operates in a specific mode, compared to when the radio wave sensor 2 operates in a normal mode, cluster 5 is more likely to depend on the position of the center of gravity 42.

[0095] Next, use Figure 11 Let's illustrate the third action example. Figure 11 This is an explanatory diagram of a third operation example of the mixing unit 16 of the detection system 10 according to the embodiment. Figure 11In the diagram, the size of the × symbol representing the center of gravity 42 indicates the amount of mixing of the center of gravity 42. That is, the larger the × symbol representing the center of gravity 42, the greater the amount of mixing of the center of gravity 42 (that is, the more centers of gravity 42 there are).

[0096] In the third example of operation, when the detection system 10 determines that the change in the detection position of the object 3 is small, it switches the operation mode of the radio wave sensor 2 from the normal mode to a specific mode. The state of small change in the detection position of the object 3 might occur, for example, when a person (object 3) is asleep. For instance, when the number of scattered points 41 in the cluster 5 corresponding to the object 3 is below a threshold, the detection system 10 switches the operation mode of the radio wave sensor 2 from the normal mode to the specific mode; when it is above the threshold, the detection system 10 switches the operation mode of the radio wave sensor 2 from the specific mode back to the normal mode. Alternatively, for example, when the stability of the intensity of the reflected radio wave is above a threshold, the detection system 10 switches the operation mode of the radio wave sensor 2 from the normal mode to the specific mode; when the stability of the intensity of the reflected radio wave is below a threshold, the detection system 10 switches the operation mode of the radio wave sensor 2 from the specific mode back to the normal mode.

[0097] As described above, in the third example of operation, when the radio wave sensor 2 operates in a specific mode, the number of centroids 42 contained in the cluster 5 increases, thus the position of the cluster 5, i.e., the detection position of the object 3, is less likely to change. Therefore, it has the following advantages: it is easy to use beamforming to concentrate the radiation or selectively receive radio waves for the object 3, thereby easily obtaining biological information about the object 3.

[0098] [2. Action]

[0099] Next, use Figure 12 Here is an example illustrating the operation of the detection system 10 involved in the implementation method. Figure 12 This is a flowchart illustrating an example of the operation of the detection system 10 according to the embodiment.

[0100] First, the acquisition unit 11 periodically acquires the received signal Sig1 (S1) output from the radio wave sensor 2. Processing S1 corresponds to the acquisition step ST1 in the detection method. Next, the generation unit 12 generates scattered points 41 (S2) based on the received signal Sig1 acquired by the acquisition unit 11. Processing S2 corresponds to the generation step ST2 in the detection method.

[0101] Next, the classification unit 13 classifies the set of scattered points 41 generated by the generation unit 12 over a predetermined time T1 (e.g., n frames), i.e., scattered point group 4, into one or more clusters 5 (S3). Processing S3 is equivalent to the classification step ST3 in the detection method. Here, if the centroid calculation unit 14 calculates the centroid 42 in the previous period P1, as described later, the centroid 42 is included in the scattered point group 4. Next, the centroid calculation unit 14 calculates the centroid 42 for each of the one or more clusters 5 classified by the classification unit 13 (S4). Processing S4 is equivalent to the centroid calculation step ST4 in the detection method.

[0102] Next, the association unit 15 associates one of the more than one clusters 5 classified by the classification unit 13 with the object 3 (S5). Processing S5 is equivalent to the association step ST5 in the detection method. Here, if the centroid calculation unit 14 calculates the centroid 42 in the previous period P1, the association unit 15 associates one of the more than one clusters 5 with the object 3 based on whether the centroid 42 is included in more than one cluster 5 in the current period P1.

[0103] Then, the mixing unit 16 mixes the centroid 42 calculated by the centroid calculation unit 14 in the previous and preceding periods P1 as scattered points 41 into the scattered point group 4 (S6). Process S6 is equivalent to the mixing step ST6 in the detection method. In addition, the order of processes S5 and S6 can be reversed, and processes S5 and S6 can also be executed in parallel. Next, processes S1 to S6 are repeated.

[0104] [3. Advantages]

[0105] As described above, the detection system 10 according to the embodiment includes clustering the centroids 42 in the previous and preceding periods P1 calculated by the centroid calculation unit 14. Therefore, the detection system 10 according to the embodiment has the following advantages: it is easy to associate past clusters 5 with the current clusters 5, thereby making it easier to capture the object 3, and thus it is easy to improve the detection accuracy of the position of the object 3.

[0106] Here, regarding whether past clusters and current clusters originate from the same object, we consider determining this based on, for example, the distance between the centroids of past and current clusters, the distance between the centroids of current and past clusters, the distance between the centroids of past and current clusters, or the similarity between past and current clusters. However, depending on how these distances or similarities are set, the results of these determinations may vary greatly, thus presenting a problem of lacking stability in the association.

[0107] In contrast, in the detection system 10 according to the embodiment, the association unit 15 associates one of the more than one cluster 5 with the object 3 based on whether the centroid 42 in the previous and earlier periods P1 is included in one or more clusters 5 classified by the classification unit 13. Therefore, the detection system 10 according to the embodiment has the following advantages: it is easier to associate past clusters 5 with current clusters 5, and it is easier to capture the object 3.

[0108] Next, use Figure 13A and Figure 13B To illustrate the advantages of the detection system 10 involved in the implementation method. Figure 13A This is a diagram illustrating an example of the detection results of the comparative detection system. Figure 13B This is a diagram illustrating an example of the detection results of the detection system 10 according to the embodiment. Figure 13A and Figure 13B In the diagram, the vertical axis represents the planar coordinates (x-coordinate) of the centroid 42 of cluster 5, and the horizontal axis represents time. Additionally, in... Figure 13A and Figure 13B In the diagram, “□”, “△”, and “〇” each represent three distinct objects 3. Furthermore, “×” does not correspond to any object 3.

[0109] The difference between the comparative example detection system and the detection system 10 involved in the implementation method is that the comparative example detection system does not have a mixing section 16, and it does not perform correlation on whether the centroid 42 in the previous and earlier periods P1 is included in more than one cluster 5. Figure 13A As shown, in the comparative example detection system, the x-coordinates of the objects 3 corresponding to “□”, “△”, and “〇” are all unstable, making it impossible to capture the objects 3. In contrast, in the detection system 10 of the embodiment, the x-coordinates of the objects 3 corresponding to “□”, “△”, and “〇” are all stable, enabling the capture of the objects 3.

[0110] (Modified Example)

[0111] The above describes the implementation methods, but the present invention is not limited to the above-described implementation methods. Below, variations of the implementation methods are listed. The variations described below can also be appropriately combined.

[0112] In the implementation method, such as Figure 14 As shown, the detection system 10A may also include a prediction unit 17. Figure 14 This is a block diagram showing the overall structure of the detection system 10A, including variations of the embodiments.

[0113] The prediction unit 17 predicts the centroid 43 in the next and subsequent periods P1 for each of the more than one clusters 5 classified by the classification unit 13. Specifically, the prediction unit 17 predicts the centroid 43 in the next period P1 by using time series data of the centroid 42 in the current and previous periods P1 calculated by the centroid calculation unit 14, for example, using an α-β filter or a Kalman filter.

[0114] Moreover, in the detection system 10A, such as Figure 15 As shown, the mixing unit 16 mixes the centroid 43 in the next and subsequent cycles P1 predicted by the prediction unit 17 as a random point 41 into the random point group 4. Figure 15 This is a diagram showing a summary of the classification unit 13 of the detection system 10A according to a modified embodiment, which is a group of scattered points 4 for classification. Figure 15 In the example shown, the centroid 42 of the current period P1, represented by the solid line × symbol, is used to predict the centroid 43 of the next period P1, represented by the dashed line × symbol, and the predicted centroid 43 is included as a random point 41 in the random point group 4.

[0115] As described above, in the detection system 10A, the scattered point group 4 is clustered with the centroid 43 in the next and subsequent periods P1 predicted by the prediction unit 17, thereby enabling the estimation of the future position of the object 3. Therefore, the detection system 10A has the advantage that the position of the object 3 can be easily detected with high accuracy even when the object 3 is moving.

[0116] In this embodiment, the radio wave sensor 2 can switch its operation between a normal mode and a specific mode, but is not limited thereto. For example, the radio wave sensor 2 may also be configured to operate only in the normal mode. In this case, the third operation example is not applied in the mixing unit 16.

[0117] In this implementation, the classification unit 13 uses DBSCAN as the clustering algorithm, but is not limited to it. For example, the classification unit 13 may also use Ward's method or k-means clustering as the clustering algorithm.

[0118] In one embodiment, the centroid calculation unit 14 calculates the centroid 42 for each of the more than one clusters 5 using the intensity of the reflected radio waves as a weighting factor, but is not limited thereto. For example, the centroid calculation unit 14 may also calculate the centroid 42 of each of the more than one clusters 5 by referring only to the position information of the scattered points 41.

[0119] Furthermore, in this embodiment, the centroid calculation unit 14 calculates the centroid 42 for each of the more than one clusters 5, including the centroid 42 of the previous period P1 as a scattered point 4, but it is not limited to this. For example, the centroid calculation unit 14 may also calculate the centroid 42 for each of the more than one clusters 5 without including the centroid 42 of the previous period P1 as a scattered point 41.

[0120] In this embodiment, the centroid calculation unit 14 calculates the planar coordinates (xy coordinates) of the centroid 42 for each of the more than one cluster 5, but is not limited thereto. For example, if the scattered points 41 contain information about the height direction (z direction), the centroid calculation unit 14 may also calculate the three-dimensional coordinates (xyz coordinates) of the centroid 42 for each of the more than one cluster 5. In this case, for calculating the z coordinate, simply replace "x" with "z" in the above equation (1). In addition, the centroid calculation unit 14 may also calculate the average value of the intensity of the reflected radio waves as information representing the centroid 42.

[0121] In one embodiment, the detection system 10 is implemented by a single device, but it can also be implemented by multiple devices. When the detection system 10 is implemented by multiple devices, the components of the detection system 10 can be distributed among the multiple devices in any way. Alternatively, for example, the components of the detection system 10 in the above embodiment can also be provided in a server device, which is located at or far from the facility. In this case, the detection system 10 can acquire the received signal Sig1 output from the radio wave sensor 2 simply by wirelessly communicating with the radio wave sensor 2 via a network such as the Internet. That is, the present invention can be implemented either through cloud computing or edge computing.

[0122] For example, there are no particular limitations on the communication method between devices in the above embodiments. Additionally, communication between devices may also be conducted via a relay device (not shown).

[0123] Furthermore, in the above embodiments, each component can be implemented by executing a software program suitable for that component. Each component can be implemented by reading and executing a software program recorded on a recording medium such as a hard disk or semiconductor memory using a program execution unit such as a CPU (Central Processing Unit) or processor.

[0124] Furthermore, each component can also be implemented in hardware. For example, each component can be a circuit (or integrated circuit). These circuits can either form a single circuit as a whole or be separate circuits. Additionally, these circuits can be general-purpose circuits or dedicated circuits.

[0125] Furthermore, the present invention, in its entirety or in specific forms, can also be implemented by a system, apparatus, method, integrated circuit, computer program, or a recording medium such as a computer-readable CD-ROM. Additionally, it can be implemented by any combination of the system, apparatus, method, integrated circuit, computer program, and recording medium.

[0126] For example, the present invention can be implemented as a management method executed by a computer, such as a detection system 10, or as a program for causing a computer to execute such a management method, or as a computer-readable non-transitory recording medium containing such a program.

[0127] In the implementation method, such as Figure 1 As shown, the detection system 10 can also be combined with the radio wave sensor 2 to form a sensor system 100. That is, the sensor system 100 includes the detection system 10 and the radio wave sensor 2 controlled by the detection system 10.

[0128] Furthermore, the sensor system 100 may also be composed of the detection system 10A and the radio wave sensor 2 as described in the modified embodiment. Additionally, in the sensor system 100, the detection system 10 and the radio wave sensor 2 may be configured independently or as a single unit. For example, the sensor system 100 may also be configured by embedding a signal processing circuit equivalent to the detection system 10 within the housing of the radio wave sensor 2.

[0129] In addition, this invention also includes various modifications that can be conceived by those skilled in the art to implement the embodiments, or the implementation by arbitrarily combining the constituent elements and functions of each embodiment without departing from the spirit of the invention.

[0130] (Summarize)

[0131] As described above, the detection system 10 includes an acquisition unit 11, a generation unit 12, a classification unit 13, a centroid calculation unit 14, an association unit 15, and a mixing unit 16. The acquisition unit 11 acquires a received signal Sig1 from an electromagnetic wave sensor 2. The electromagnetic wave sensor 2 radiates electromagnetic waves from a transmitting antenna 211 towards the object 3 and receives reflected electromagnetic waves from the object 3 through multiple receiving antennas 221. The generation unit 12 generates scattered points 41 containing the location information of the object 3 based on the received signal Sig1 acquired by the acquisition unit 11. The classification unit 13 classifies the set of scattered points 41 generated by the generation unit 12 over a predetermined time T1, i.e., scattered point groups 4, into one or more clusters 5. The centroid calculation unit 14 calculates the centroid 42 for each of the one or more clusters 5 classified by the classification unit 13. The association unit 15 associates one of the one or more clusters 5 classified by the classification unit 13 with the object 3. The mixing unit 16 mixes the centroid 42 in the previous and earlier periods P1 calculated by the centroid calculation unit 14 as scattered point 4 into the scattered point group 4.

[0132] The detection system 10 has the following advantages: it is easy to associate past clusters 5 with current clusters 5, thereby making it easy to capture object 3, and thus it is easy to improve the detection accuracy of the position of object 3.

[0133] Additionally, for example, in the detection system 10, the scattered points 41 contain information related to at least one of the distance between the object 3 and the radio wave sensor 2 and the direction of the object 3 relative to the radio wave sensor 2, as well as information related to the intensity of the reflected radio waves.

[0134] The detection system 10 has the following advantages: it makes it easier to improve the detection accuracy of the position of the object 3.

[0135] Additionally, for example, in the detection system 10, the association unit 15 associates one of the more than one cluster 5 with the object 3 based on whether the centroid 42 in the previous and earlier cycles P1 is included in more than one cluster 5 classified by the classification unit 13.

[0136] According to such a detection system 10, it has the following advantages: it is easier to associate the past cluster 5 with the current cluster 5, thus making it easier to capture the object 3.

[0137] Additionally, for example, in the detection system 102, the classification unit 13 uses an algorithm based on the density of scattered points 41 to remove isolated points and classify the scattered point group 4 into more than one cluster 5.

[0138] According to such a detection system 10, the following advantages are available: noise unrelated to the object 3 can be easily removed and classified into more than one cluster 5, thus easily improving the detection accuracy of the object 3.

[0139] Additionally, for example, in the detection system 10, the mixing unit 16 adjusts the mixing amount or delay amount of the centroid 42 in the previous and earlier cycles P1 that are mixed into the scattered point group 4.

[0140] According to such a detection system 10, the following advantages are available: it is possible to easily associate the current cluster 5 with the past cluster 5, or to smooth the positional changes of cluster 5 over time.

[0141] Additionally, for example, in the detection system 10, for each of the more than one clusters 5, if it does not contain scattered points 41 other than the centroid 42, the mixing unit 16 increments the counter; if it contains scattered points 41 other than the centroid 42, the mixing unit 16 resets the counter; if the counter reaches a threshold (first threshold) Th1, the mixing unit 16 discards the centroid 42.

[0142] The detection system 10 has the following advantages: it can prevent the continuous presence of clusters 5 even though there are no scattered points 41, that is, no object 3, thereby easily improving the detection accuracy of object 3.

[0143] Additionally, for example, in the detection system 10, for each of the more than one clusters 5, if it does not contain scattered points 41 other than the centroid 42, the mixing unit 16 reduces the mixing amount of the centroid 42; if it contains scattered points 41 other than the centroid 42, the mixing unit 16 restores the mixing amount of the centroid 42.

[0144] The detection system 10 has the following advantages: it can prevent the continuous presence of clusters 5 even though there are no scattered points 41, that is, no object 3, thereby easily improving the detection accuracy of object 3.

[0145] Furthermore, for example, in the detection system 10, the radio wave sensor 2 can switch its operation to any mode between a normal mode and a specific mode, wherein the specific mode is a mode that radiates radio waves in a concentrated manner or selectively receives radio waves in a specific direction. When the radio wave sensor 2 operates in the specific mode, compared with the case where the radio wave sensor 2 operates in the normal mode, the mixing unit 16 increases the mixing amount at the center of gravity 42.

[0146] According to such a detection system 10, the following advantages are available: the detection position of the object 3 is not easily changed, so it is easy to use, for example, beamforming to concentrate the radiation or selectively receive radio waves for the object 3, thereby easily obtaining biological information of the object 3.

[0147] Additionally, for example, the detection system 10A also includes a prediction unit 17, which predicts the centroid 42 in the next and subsequent periods P1 for each of the more than one cluster 5 classified by the classification unit 13. The mixing unit 16 mixes the centroid 42 predicted by the prediction unit 17 in the next and subsequent periods as a random point 41 into the random point group 4.

[0148] The detection system 10A has the following advantages: the position of object 3 can be easily detected with high accuracy even when object 3 is moving.

[0149] Additionally, for example, the sensor system 100 includes a detection system 10 and an electromagnetic wave sensor 2 controlled by the detection system 10.

[0150] According to such a sensor system 100, the following advantages are available: it is easy to associate the past cluster 5 with the current cluster 5, thereby making it easy to capture the object 3, and thus it is easy to improve the detection accuracy of the position of the object 3.

[0151] Additionally, for example, the detection method includes an acquisition step ST1, a generation step ST2, a classification step ST3, a centroid calculation step ST4, an association step ST5, and a mixing step ST6. In the acquisition step ST1, a received signal Sig1 is acquired from an electromagnetic sensor 2, which radiates electromagnetic waves from a transmitting antenna 211 towards the object 3 and receives reflected electromagnetic waves from the object 3 through multiple receiving antennas 221. In the generation step ST2, scattered points 41 containing the location information of the object 3 are generated based on the received signal Sig1 acquired in the acquisition step ST1. In the classification step ST3, the set of scattered points 41 generated in the generation step ST2 over a predetermined time T1, i.e., scattered point group 4, is classified into one or more clusters 5. In the centroid calculation step ST4, the centroid 42 is calculated for each of the one or more clusters 5 classified in the classification step ST3. In the association step ST5, one of the clusters 5 classified in the classification step ST3 is associated with the object 3. In the mixing step ST6, the centroid 42 in the previous and earlier periods P1 calculated in the centroid calculation step ST4 is mixed into the scattered point group 4 as a scattered point 41.

[0152] According to this detection method, the following advantages are available: it is easy to associate the past cluster 5 with the current cluster 5, thereby making it easy to capture the object 3, and thus it is easy to improve the detection accuracy of the position of the object 3.

[0153] Alternatively, for example, the program recording medium may contain a program for causing one or more processors to execute the aforementioned detection method.

[0154] The recording medium based on such a program has the following advantages: it is easy to associate the past cluster 5 with the current cluster 5, thereby making it easy to capture the object 3, and thus it is easy to improve the detection accuracy of the position of the object 3.

[0155] Explanation of reference numerals in the attached figures

[0156] 10, 10A: Detection system; 11: Acquisition unit; 12: Generation unit; 13: Classification unit; 14: Center of gravity calculation unit; 15: Correlation unit; 16: Mixing unit; 17: Prediction unit; 2: Radio wave sensor; 211: Transmitting antenna; 221: Receiving antenna; 3: Object; 100: Sensor system; 4: Scattered point group; 41: Scattered point; 42, 43: Center of gravity; 5: Cluster; T1: Specified time; P1: Period; Sig1: Received signal; ST1: Acquisition step; ST2: Generation step; ST3: Classification step; ST4: Center of gravity calculation step; ST5: Correlation step; ST6: Mixing step; Th1: First threshold (threshold).

Claims

1. A detection system, comprising: The acquisition unit acquires received signals from an electromagnetic wave sensor, which radiates electromagnetic waves from a transmitting antenna toward the object and receives reflected electromagnetic waves from the object through multiple receiving antennas. The generation unit generates scattered points containing the location information of the object based on the received signal acquired by the acquisition unit. The classification department classifies the set of scattered points generated by the generation department over a specified time into one or more clusters. The centroid calculation unit calculates the centroid of the coordinates of each scattered point contained in each of the more than one clusters obtained by the classification part. The association section associates one of the more than one clusters obtained from the classification section with the object. as well as The mixing unit mixes the centroids calculated by the centroid calculation unit for the previous and earlier periods into the scattered point group. The association unit associates one of the more than one clusters with the object based on whether the center of gravity in the previous and earlier cycles is included in the more than one clusters obtained by the classification part.

2. The detection system according to claim 1, characterized in that, The scattered points include information relating to at least one of the distance between the object and the radio wave sensor and the direction of the object relative to the radio wave sensor, as well as information relating to the intensity of the reflected radio wave.

3. The detection system according to claim 1 or 2, characterized in that, The classification department uses an algorithm based on the density of the scattered points to remove isolated points and classifies the scattered point group into one or more clusters.

4. The detection system according to claim 1 or 2, characterized in that, The mixing unit adjusts the mixing amount or delay amount of the center of gravity in the previous and earlier cycles that are mixed into the scattered point group.

5. The detection system according to claim 4, characterized in that, Regarding each of the more than one clusters, if it does not contain any scattered points other than the centroid, the mixing unit increments the counter; if it contains any scattered points other than the centroid, the mixing unit resets the counter; and if the counter reaches a threshold, the mixing unit discards the centroid.

6. The detection system according to claim 4, characterized in that, Regarding each of the more than one clusters, if it does not contain any scattered points other than the centroid, the mixing part reduces the mixing amount of the centroid; if it contains any scattered points other than the centroid, the mixing part restores the mixing amount of the centroid.

7. The detection system according to claim 4, characterized in that, The radio wave sensor can switch its operation between a normal mode and a specific mode, wherein the specific mode is a mode that concentrates the radiation of radio waves in a specific direction or selectively receives them. When the radio wave sensor operates in the specific mode, compared to when the radio wave sensor operates in the normal mode, the mixing unit increases the mixing amount of the center of gravity.

8. The detection system according to claim 1 or 2, characterized in that, It also includes a prediction unit that predicts the centroid for the next and subsequent cycles for each of the more than one clusters obtained by the classification part. The mixing unit mixes the centroids predicted by the prediction unit for the next and subsequent cycles into the group of scattered points.

9. A sensor system comprising a detection system according to claim 1 or 2, and the radio wave sensor controlled by the detection system.

10. A detection method, comprising: The acquisition step involves acquiring a received signal from an electromagnetic wave sensor, which radiates electromagnetic waves from a transmitting antenna toward the object and receives reflected electromagnetic waves from the object through multiple receiving antennas. The generation step generates scattered points containing the location information of the object based on the received signal obtained in the acquisition step. The classification step classifies the set of scattered points generated over a specified time in the generation step, i.e., the scattered point group, into one or more clusters. The centroid calculation step involves calculating the centroid of the coordinates of each of the scattered points contained in the cluster for each of the more than one clusters obtained in the classification step. The association step involves associating one of the more than one clusters obtained in the classification step with the object. as well as The mixing step involves incorporating the centroids calculated in the centroid calculation step from the previous and earlier periods into the scattered point group as the scattered points. In the association step, one of the more than one clusters is associated with the object based on whether the centroid in the previous and earlier cycles is included in the more than one clusters classified in the classification step.

11. A program recording medium having a program for causing one or more processors to execute the detection method according to claim 10.

Images