Anomaly detection device and anomaly detection method

Abstract

The present invention comprises: a data acquisition unit (11) that acquires detection data, which relates to a mobile body existing in a target region and which is generated on the basis of sensor data acquired by a sensor (2) detecting an object existing in the target region; a selection unit (131) that selects whether to perform abandonment detection in the target region or intrusion detection for the target region; an abandonment detection unit (132) that, when the selection unit (131) has selected to perform the abandonment detection in the target region, detects the presence or absence of abandonment in the target region on the basis of the detection data acquired by the data acquisition unit (11); and an intrusion detection unit (133) that, when the selection unit (131) has selected to perform the intrusion detection for the target region, detects the presence or absence of intrusion into the target region on the basis of the detection data acquired by the data acquisition unit (11).

Description

Anomaly detection device and anomaly detection method

[0001] The present disclosure relates to an anomaly detection device and an anomaly detection method.

[0002] In recent years, the problem of infants and other children being left behind in areas that are subject to abnormality detection, such as vehicle interiors (hereinafter referred to as "target areas"), has become a social issue. Meanwhile, technologies for preventing unauthorized entry into target areas have been studied. For example, a technology is known that utilizes an inherent anti-theft function to detect abandonment in the target areas and prevent such abandonment from occurring (e.g., Patent Document 1).

[0003] JP 2010-70060 A

[0004] The detection of an intrusion into the target area and the detection of an abandoned child or the like in the target area do not necessarily have to be performed at the same time. The conventional technology disclosed in Patent Document 1 does not take this into consideration, and there is a problem in that the detection of an intrusion into the target area and the detection of an abandoned child or the like in the target area are continuously performed. As a result, there is a problem in that power is consumed by the execution of unnecessary processes.

[0005] The present disclosure has been made to solve the above-mentioned problems, and aims to provide an anomaly detection device that can simultaneously detect whether or not an intrusion into a target area has occurred and whether or not an infant or other person has been left behind in the target area, while suppressing the execution of unnecessary processing.

[0006] The anomaly detection device according to the present disclosure includes a data acquisition unit that acquires detection data regarding moving objects present in a target area, which is generated based on sensor data acquired by a sensor detecting an object present in the target area; a selection unit that selects whether to perform abandonment detection in the target area or intrusion detection into the target area; an abandonment detection unit that, if the selection unit selects to perform abandonment detection in the target area, detects whether an object has been abandoned in the target area based on the detection data acquired by the data acquisition unit; and an intrusion detection unit that, if the selection unit selects to perform intrusion detection into the target area, detects whether an intrusion has occurred into the target area based on the detection data acquired by the data acquisition unit.

[0007] According to the present disclosure, the anomaly detection device can simultaneously detect whether or not an intrusion into a target area has occurred and whether or not a child or other person has been left behind in the target area, while suppressing the execution of unnecessary processes.

[0008] FIG. 5A is a diagram illustrating an example of the configuration of an anomaly detection device according to embodiment 1. FIG. 5B is a diagram illustrating an example of the configuration of a data processing unit. FIG. 3A and FIG. 3B are diagrams illustrating an example of a three-dimensional spatial distribution generated by a three-dimensional spatial distribution generation unit according to embodiment 1. FIG. 3A shows an example of a top view of the three-dimensional spatial distribution, and FIG. 3B shows an example of a side view of the three-dimensional spatial distribution. FIG. 5A is a diagram illustrating an example of the concept of an abandonment detection unit detecting whether an object has been abandoned in a vehicle cabin using a three-dimensional spatial distribution according to embodiment 1. FIG. 5B is a diagram illustrating an example of the concept of an intrusion detection unit detecting whether an intrusion has occurred in the vehicle cabin using a three-dimensional spatial distribution. FIG. 5B is a diagram illustrating an example of the detailed configuration of a detection result output unit of an anomaly detection device according to embodiment 1. 9A, 9B, and 9C are diagrams illustrating an example of the operation of the anomaly detection device according to the first embodiment. FIGS. 8A, 8B, and 8C are diagrams illustrating an example of the control content when the selection unit is capable of controlling the operation of the sensor in the first embodiment. FIGS. 9A, 9B, and 9C are diagrams illustrating the concept of the selection unit changing the anomaly detection cycle in the first embodiment. FIG. 9B is a diagram illustrating an example of a three-dimensional spatial distribution generated based on sensor data obtained by a sensor with a smaller number of antennas detecting an object present in the vehicle cabin. FIG. 9A is a diagram illustrating an example of a three-dimensional spatial distribution generated based on sensor data obtained by a sensor with a larger number of antennas than FIG. 9B detecting an object present in the vehicle cabin. FIG. 9C is a diagram illustrating a three-dimensional spatial distribution obtained by superimposing the distribution of moving objects on the three-dimensional spatial distribution illustrated in FIG. 9B. FIGS. 10A and 10B are diagrams illustrating an example of the hardware configuration of the anomaly detection device according to the first embodiment.

[0009] In the present disclosure, the anomaly detection device detects an abnormality in an area targeted for anomaly detection (hereinafter referred to as the "target area"). In the present disclosure, abnormalities detected by the anomaly detection device include the abandonment of a person requiring assistance in the target area and intrusion into the target area. That is, the target area is, more specifically, an area targeted for detecting whether a person requiring assistance has been abandoned or whether an intrusion into the target area has occurred. In the present disclosure, the "person requiring assistance" refers to a living being, such as an infant, that has a physical constitution that makes it difficult for the person to leave the target area on their own if left in the target area. Note that in the present disclosure, the "person requiring assistance" also includes living beings such as pets. In the following embodiments, as an example, the "target area" is the vehicle interior, and the "person requiring assistance" is an infant. The anomaly detection device detects both whether an infant has been left in the vehicle interior and whether a person has intruded into the vehicle interior. Note that the vehicle interior refers to, for example, the interior of a regular passenger car. In the following embodiments, when the term "abandonment" is simply used, it refers to the abandonment of a person requiring assistance. In the following embodiments, when the term "intrusion" is used simply, it means the intrusion of a person. More specifically, it means the intrusion of a moving object that is considered to be a person. In the following embodiments, the detection of whether an object is left inside the vehicle cabin, which is performed by the anomaly detection device, is also simply referred to as "detection of abandonment inside the vehicle cabin," and the detection of whether an intrusion into the vehicle cabin is also simply referred to as "detection of intrusion into the vehicle cabin."

[0010] 1 is a diagram showing an example of the configuration of an anomaly detection device 1 according to embodiment 1. The anomaly detection device 1 is mounted on, for example, a vehicle 100, and is connected to a sensor 2 and an output device 3.

[0011] The sensor 2 detects an object in the vehicle cabin. In the first embodiment, the sensor 2 is assumed to be a radio wave sensor such as a millimeter wave radar. The sensor 2 acquires, as sensor data, radio waves that are emitted toward the vehicle cabin and reflected by objects in the vehicle cabin. The sensor 2 is equipped with, for example, an antenna with wide-angle directivity, and is installed in the vehicle cabin so that radio waves are irradiated to occupants who may be present in the vehicle cabin.

[0012] Sensor 2 is a general radio wave sensor that detects objects. An example of how sensor data is acquired by sensor 2 will be described. While various modulation methods are available for the sensing signals of radio wave sensors, this example will use the FM-CW (Frequency Modulation - Continuous Wave) method, which is commonly used in automotive applications. A radio wave transmitter / receiver (not shown) included in sensor 2 periodically generates an FM signal (called a chirp wave) whose frequency increases and decreases. The radio wave transmitter / receiver amplifies the signal power to obtain the power required for radio wave emission, and emits radio waves into the interior of the vehicle via a transmitting antenna (not shown). When the radio waves radiated into the interior of the vehicle reach a target object within the radio wave emission range of the radio wave transmitter / receiver, a portion of the radio waves is reflected by the surface of the target object and returns to the radio wave transmitter / receiver. Here, the target object refers to an object that reflects radio waves, such as a passenger in the vehicle or a vehicle structure.

[0013] The radio wave transmitting / receiving unit receives radio waves (reflected waves) reflected from the surface of the target object via a receiving antenna (not shown). A signal similar to the FM transmitted wave is input as a received FM signal to the radio wave transmitting / receiving unit. The received signal is input to the radio wave transmitting / receiving unit with a time lag corresponding to the time it takes for the radio waves to reach the target object and return. The radio wave transmitting / receiving unit extracts the frequency difference between the frequency of the generated FM signal and the frequency of the received signal, and generates an intermediate frequency (IF) signal having the frequency difference. An A / D conversion unit (not shown) included in the sensor 2 converts the intermediate frequency signal from an analog signal to a digital signal and outputs the digital signal to the anomaly detection device 1 as sensor data.

[0014] 1 illustrates only one sensor 2, but this is merely an example. A plurality of sensors 2 may be provided in the vehicle 100, and the plurality of sensors 2 may be connected to the abnormality detection device 1. For example, one sensor 2 may be provided in a position from which the entire interior of the vehicle can be viewed, such as an overhead console, so that one sensor 2 can cover the entire area of ​​the vehicle interior. Alternatively, a plurality of sensors 2 may be provided in an area corresponding to a seat or in each of a plurality of areas in the vehicle interior that have been previously divided into a plurality of areas. Furthermore, the operation of the sensor 2 may be controlled by the sensor 2 alone using a trigger inside the sensor 2, or the operation of the sensor 2 may be controlled based on a trigger from outside the sensor 2.

[0015] The anomaly detection device 1 detects whether an object has been left in the vehicle cabin and whether a person has entered the vehicle cabin based on the sensor data output from the sensor 2. In the first embodiment, the anomaly detection device 1 detects whether an object has been left in the vehicle cabin and whether a person has entered the vehicle cabin based on the sensor data output from the common sensor 2. Details of an example configuration of the anomaly detection device 1 will be described later.

[0016] The output device 3 is, for example, a display device such as a display or an audio output device such as a speaker provided on a mobile terminal carried by the owner of the vehicle 100, or a lamp or a horn provided on the vehicle 100. Furthermore, for example, the output device 3 may be a driving control device that controls the driving of the vehicle 100. Furthermore, for example, the output device 3 may be a server provided in an emergency system. Note that, in FIG. 1 , as an example, the output device 3 is provided on the vehicle 100, but this is merely an example. As described above, the output device 3 may be provided outside the vehicle 100 and connected to the abnormality detection device 1 outside the vehicle 100.

[0017] The output device 3 performs an operation based on information (hereinafter referred to as "output control information") generated based on the result of the abnormality detection performed by the abnormality detection device 1 (specifically, detection of whether an object has been left in the vehicle cabin or whether an intrusion has occurred into the vehicle cabin). For example, if the output device 3 is a display device, the display device displays a warning that an object has been left in the vehicle cabin or a warning that an intrusion has occurred into the vehicle cabin based on the output control information. For example, if the output device 3 is an audio output device, the audio output device outputs an alarm that an object has been left in the vehicle cabin or a warning that an intrusion has occurred into the vehicle cabin based on the output control information. For example, if the output device 3 is a driving control device, the driving control device controls the vehicle 100 based on the output control information. Control of the vehicle 100 includes stopping the vehicle 100, opening and closing windows, controlling air conditioning, etc. For example, if the output device 3 is a server included in an emergency system, the server requests an ambulance based on the output control information.

[0018] 1 illustrates only one output device 3, but this is merely an example. For example, a plurality of output devices 3 may be provided, and the plurality of output devices 3 may be connected to the anomaly detection device 1. Furthermore, for example, a plurality of types of output devices 3, such as an audio output device and an operation control device, may be connected to the anomaly detection device 1, and each of the plurality of types of output devices 3 may perform an operation based on the output control information.

[0019] An example configuration of an anomaly detection device 1 according to the first embodiment will be described. As shown in FIG. 1 , the anomaly detection device 1 includes a data acquisition unit 11, an anomaly detection unit 13, and a detection result output unit 14. The data acquisition unit 11 includes a data processing unit 12. The anomaly detection unit 13 includes a selection unit 131, an abandonment detection unit 132, and an intrusion detection unit 133. Note that, although the data acquisition unit 11 is shown here as including the data processing unit 12, this is merely an example. For example, the data processing unit 12 may be provided outside the data acquisition unit 11 in the anomaly detection device 1.

[0020] The data acquisition unit 11 acquires data on a moving object present in the vehicle cabin (hereinafter referred to as "detection data"), which is generated based on sensor data acquired by the sensor 2 detecting an object present in the vehicle cabin. Specifically, the data processing unit 12 generates the detection data based on the sensor data acquired from the sensor 2, and the data acquisition unit 11 acquires the detection data generated by the data processing unit 12.

[0021] The data processing unit 12 performs pre-processing on the sensor data acquired from the sensor 2, which is performed by the abnormality detection device 1 in common when detecting whether an object has been left in the vehicle cabin and when detecting whether an object has entered the vehicle cabin, and generates detection data.

[0022] 2 is a diagram showing an example of the configuration of the data processing unit 12. As shown in Fig. 2, the data processing unit 12 includes, for example, a moving object extraction unit 121, a moving object analysis unit 122, and a three-dimensional spatial distribution generation unit 123.

[0023] The moving object extraction unit 121 extracts a moving object based on sensor data acquired from the sensor 2. In the first embodiment, the moving object extracted by the moving object extraction unit 121 is an intruder into the vehicle 100 or an occupant left in the vehicle cabin. The moving object extraction unit 121 extracts the moving object by detecting movement within the vehicle cabin based on the sensor data. The movement within the vehicle cabin includes movement due to a change in the position of the intruder into the vehicle 100 or an occupant left in the vehicle cabin, and body movement such as chest movement due to breathing. One possible method for the moving object extraction unit 121 to extract a moving object is to apply a known MTI (Moving Target Indicator) filter to the sensor data, in this case, the received signal acquired by the sensor 2, more specifically, to an intermediate frequency signal based on the received signal. By applying the MTI filter, signal components due to completely stationary objects such as seats in the vehicle cabin are removed, and signal components due to the moving object are extracted. The moving object extraction unit 121 outputs the sensor data from which the signal components due to the moving object have been extracted to the moving object analysis unit 122 .

[0024] The moving object analysis unit 122 extracts information such as the position, speed, and signal strength of the moving object (hereinafter referred to as "moving object information") by performing frequency analysis on the received signal indicating the moving object extracted by the moving object extraction unit 121 based on the sensor data output from the moving object extraction unit 121, and outputs the extracted moving object information to the three-dimensional spatial distribution generation unit 123. The moving object analysis unit 122 can extract the moving object information based on various known methods or procedures such as Fourier transform (FFT), integration processing, peak extraction, and beamforming.

[0025] The three-dimensional spatial distribution generation unit 123 generates a three-dimensional spatial distribution as detection data based on the moving object information output from the moving object analysis unit 122. The three-dimensional spatial distribution generated by the three-dimensional spatial distribution generation unit 123 represents the distribution of areas in the vehicle cabin where moving objects exist in three dimensions. Specifically, the three-dimensional spatial distribution represents the minute movements of moving objects, in other words, target objects that reflect radio waves emitted by the sensor 2, within the vehicle cabin using multiple grids associated with radio wave reflection points in three-dimensional space. The intermediate frequency signal is generated from reflected waves of radio waves reflected by the surface of the target object. The moving object analysis unit 122 can calculate the positions of the reflection points on the surface of the target object from the intermediate frequency signal and extract moving object information. Because there are multiple reflection points on the surface of the target object, the three-dimensional spatial distribution generation unit 123 can generate a three-dimensional spatial distribution for the target object based on the moving object information extracted by signal processing of the reflected waves from the multiple reflection points.

[0026] 3A and 3B are diagrams showing an example of a three-dimensional spatial distribution generated by the three-dimensional spatial distribution generation unit 123 in Embodiment 1. Fig. 3A shows an example of a top view of the three-dimensional spatial distribution, and Fig. 3B shows an example of a side view of the three-dimensional spatial distribution. The three-dimensional spatial distribution shown using Figs. 3A and 3B is a three-dimensional spatial distribution generated when there is an infant in a rearward-facing child car seat installed in the left rear seat inside the vehicle.

[0027] In the three-dimensional spatial distribution, each grid is assigned a numerical value corresponding to the velocity of the target object present at the grid position. Specifically, the greater the velocity of the target object present at the grid position, the greater the numerical value assigned to that grid. In other words, in the three-dimensional spatial distribution, grids included in the range where a moving object with slight motion is present are assigned a numerical value greater than grids where no moving object is present. In the three-dimensional spatial distribution shown in FIGS. 3A and 3B, grids with larger assigned numerical values ​​are shown darker. In other words, in the three-dimensional spatial distribution shown in FIGS. 3A and 3B, grids included in the range where a moving object is present are shown darker. Note that stationary objects, in other words, reflected wave components with a velocity of zero, do not appear in the three-dimensional spatial distribution.

[0028] When the three-dimensional spatial distribution generating unit 123 generates the three-dimensional spatial distribution, the data acquiring unit 11 acquires it and outputs the acquired three-dimensional spatial distribution to the anomaly detecting unit 13 .

[0029] The abnormality detection unit 13 selects whether to detect whether an object has been left in the vehicle cabin or whether an intrusion has occurred into the vehicle cabin, and executes the selected detection process. In the first embodiment, the mode in which the abnormality detection device 1 detects whether an object has been left in the vehicle cabin is referred to as the "abandonment detection mode," and the mode in which the abnormality detection device 1 detects whether an intrusion has occurred into the vehicle cabin is referred to as the "intrusion detection mode." The abnormality detection unit 13 selectively switches between the abandonment detection mode and the intrusion detection mode.

[0030] The selection unit 131 of the abnormality detection unit 13 selects whether to detect whether an object is left in the vehicle cabin or whether an intrusion into the vehicle cabin is to be performed. That is, the selection unit 131 selects whether to use the abandonment detection mode or the intrusion detection mode. When the selection unit 131 selects the abandonment detection mode, it outputs an instruction to the abandonment detection unit 132 to detect whether an object is left in the vehicle cabin (hereinafter referred to as an "abandonment detection start instruction"). When the selection unit 131 selects the intrusion detection mode, it outputs an instruction to the intrusion detection unit 133 to detect whether an intrusion into the vehicle cabin is to be performed (hereinafter referred to as an "intrusion detection start instruction").

[0031] Here, an example of the selection criteria used by the selection unit 131 to select either the abandonment detection mode or the intrusion detection mode will be described.

[0032] <Selection Example (1)> For example, when the operation of the anomaly detection device 1 is started, the selection unit 131 selects the abandonment detection mode until a preset time (hereinafter referred to as the "switching time") has elapsed, and selects the intrusion detection mode if no abandonment is detected even after the switching time has elapsed. That is, the selection unit 131 switches from the abandonment detection mode to the intrusion detection mode when the switching time has elapsed after the operation of the anomaly detection device 1 has started. The timing at which the operation of the anomaly detection device 1 is started is, for example, when the doors of the vehicle 100 are locked. That is, for example, after the doors are locked, the selection unit 131 selects the abandonment detection mode until the switching time has elapsed, and selects the intrusion detection mode if no abandonment is detected even after the switching time has elapsed. The timing at which the operation of the anomaly detection device 1 starts or ends will be described later with an example. Note that the selection unit 131 may, for example, acquire data indicating that the doors are locked or unlocked (hereinafter referred to as "door data") from a door sensor (not shown) mounted on the vehicle 100 that detects door locking or unlocking.

[0033] In this case, when the selection unit 131 determines that the door is locked based on the door data, it outputs an abandonment detection start instruction to the abandonment detection unit 132. The abandonment detection unit 132 detects whether or not an object has been left in the vehicle cabin based on the abandonment detection start instruction output from the selection unit 131. Then, the abandonment detection unit 132 outputs the detection result of whether or not an object has been left in the vehicle cabin (hereinafter referred to as the "abandonment detection result") to the selection unit 131 and the detection result output unit 14. Details of the abandonment detection unit 132 will be described later. If the abandonment detection unit 132 does not output an abandonment detection result indicating that an object has been detected in the vehicle cabin even after the switching time has elapsed, the selection unit 131 selects the intrusion detection mode. Then, the selection unit 131 outputs an intrusion detection start instruction to the intrusion detection unit 133. At this time, the selection unit 131 outputs an instruction to the abandonment detection unit 132 to end the detection of whether or not an object has been left in the vehicle cabin (hereinafter referred to as the "abandonment detection end instruction"). In this case, the abandonment detection unit 132 repeatedly detects whether or not an object is left in the vehicle compartment until the switching time has elapsed, for example.

[0034] <Selection Example (2)> For example, the selection unit 131 selects the abandonment detection mode when the operation of the anomaly detection device 1 starts, and selects the intrusion detection mode when the abandonment detection unit 132 completes detection of whether or not an object has been left in the vehicle cabin. That is, the selection unit 131 switches from the abandonment detection mode to the intrusion detection mode when the process of detecting whether or not an object has been left after the operation of the anomaly detection device 1 starts is completed. In <Selection Example (2)> as well, if the timing at which the operation of the anomaly detection device 1 starts is, for example, when the doors of the vehicle 100 are locked, the selection unit 131 outputs an abandonment detection start instruction to the abandonment detection unit 132 when it determines that the doors are locked based on the door data. The abandonment detection unit 132 detects whether or not an object has been left in the vehicle cabin based on the abandonment detection start instruction output from the selection unit 131. Then, when the abandonment detection unit 132 completes the process of detecting whether an object is left in the vehicle cabin, it outputs a notification indicating that the detection of whether an object is left in the vehicle cabin has been completed (hereinafter referred to as an "abandonment detection completion notification") to the selection unit 131. Note that the abandonment detection unit 132 may complete the detection of whether an object is left in the vehicle cabin when a detection result of whether an object is left in the vehicle cabin has been obtained once, or may complete the detection of whether an object is left in the vehicle cabin when a detection result of whether an object is left in the vehicle cabin has been repeatedly obtained a predetermined number of times. When the abandonment detection completion notification is output from the abandonment detection unit 132, the selection unit 131 selects the intrusion detection mode. Then, the selection unit 131 outputs an intrusion detection start instruction to the intrusion detection unit 133. At this time, the selection unit 131 outputs an abandonment detection end instruction to the abandonment detection unit 132.

[0035] The difference between the mode selection method by the selection unit 131 in the above-mentioned <Selection Example (2)> and the mode selection method by the selection unit 131 in <Selection Example (1)> is that the selection unit 131 does not set a predetermined period (switching time) and immediately switches to the intrusion detection mode if it detects that no object has been left in the vehicle cabin.

[0036] <Selection Example (3)> For example, when the operation of the anomaly detection device 1 is started, the selection unit 131 may select the abandonment detection mode while the owner of the vehicle 100 is present in the vicinity of the vehicle 100, and may select the intrusion detection mode when the owner of the vehicle 100 leaves the vicinity of the vehicle 100. That is, the selection unit 131 switches from the abandonment detection mode to the intrusion detection mode at the time when the owner of the vehicle 100 leaves the vicinity of the vehicle 100. The selection unit 131 determines whether the owner of the vehicle 100 is present in the vicinity of the vehicle 100, for example, based on whether the owner is present in a predetermined vehicle vicinity area. The selection unit 131 can determine whether the owner of the vehicle 100 is present in the vehicle vicinity area, for example, from the detection data acquired from the data acquisition unit 11, i.e., the three-dimensional spatial distribution in this case.

[0037] 4A and 4B are diagrams illustrating an example of a three-dimensional spatial distribution when the owner of vehicle 100 moves from within the vehicle periphery area to outside the vehicle periphery area in the first embodiment. FIG. 4A illustrates an example of a three-dimensional spatial distribution when the owner of vehicle 100 is present within the vehicle periphery area, and FIG. 4B illustrates an example of a three-dimensional spatial distribution when the owner of vehicle 100 moves outside the vehicle periphery area and is present outside the vehicle periphery area. Note that FIGS. 4A and 4B illustrate an example of a top view of the three-dimensional spatial distribution. In FIG. 4A , the owner of vehicle 100 is indicated by "D." The radio waves emitted from sensor 2 also propagate outside the vehicle cabin. Therefore, sensor 2 also detects people around vehicle 100, in other words, people outside the vehicle cabin. As a result, in the three-dimensional spatial distribution, the minute movements of people around vehicle 100 who reflect the radio waves emitted by sensor 2 are represented by a plurality of grids corresponding to the reflection points of the radio waves in the three-dimensional space outside the vehicle cabin.

[0038] The selection unit 131 stores, for example, the three-dimensional spatial distribution acquired from the data acquisition unit 11 for a predetermined time period in a storage unit (not shown). The storage unit is provided, for example, in a location accessible by the anomaly detection device 1. Based on the time-series three-dimensional spatial distribution stored in the storage unit, the selection unit 131 determines that the owner of the vehicle 100 has moved from within the vehicle periphery area to outside the vehicle periphery area, in other words, that the owner of the vehicle 100 has left the periphery of the vehicle 100, when the state in which a moving object is present inside the vehicle cabin changes to a state in which the moving object is present outside the vehicle cabin and within the vehicle periphery area (see FIG. 4A ), and then changes to a state in which the moving object is present outside the vehicle periphery area (see FIG. 4B ). Note that FIG. 4 does not illustrate an example of a three-dimensional spatial distribution indicating a state in which a moving object is present inside the vehicle cabin.

[0039] The above-described method of determining whether the owner of vehicle 100 has moved outside the area surrounding the vehicle is merely an example, and selection unit 131 may use other methods to determine whether the owner of vehicle 100 has moved outside the area surrounding the vehicle. For example, selection unit 131 may determine whether the owner of vehicle 100 has moved outside the area surrounding the vehicle by short-range wireless communication with a key sensor (not shown) provided on the key of vehicle 100.

[0040] The selection unit 131 selects the intrusion detection mode when it determines that the owner of the vehicle 100 has left the vicinity of the vehicle 100. Then, the selection unit 131 outputs an intrusion detection start instruction to the intrusion detection unit 133. At this time, the selection unit 131 outputs an abandonment detection end instruction to the abandonment detection unit 132. In this case, the abandonment detection unit 132 repeatedly detects whether or not an object has been left in the vehicle cabin until the abandonment detection end instruction is output from the selection unit 131, for example.

[0041] The selection unit 131 may determine whether to select the abandonment detection mode or the intrusion detection mode by combining, for example, the above-described <Selection Example (1)>, <Selection Example (2)>, or <Selection Example (3)>. For example, the selection unit 131 may combine <Selection Example (1)> and <Selection Example (3)> to select the intrusion detection mode and output an abandonment detection end instruction to the abandonment detection unit 132 when either the owner of the vehicle 100 leaves the vehicle 100 or the switching time has elapsed. Note that the above-described <Selection Example (1)>, <Selection Example (2)>, or <Selection Example (3)> is merely an example, and the selection unit 131 may determine whether to select the abandonment detection mode or the intrusion detection mode using selection criteria other than those described above. For example, the selection unit 131 may select the abandonment detection mode or the intrusion detection mode in accordance with the laws and regulations or assessments of the area in which the vehicle 100 is used.

[0042] When the selection unit 131 selects to detect whether an object is left inside the vehicle cabin, the abandonment detection unit 132 of the abnormality detection unit 13 detects whether an object is left inside the vehicle cabin based on the detection data (here, the three-dimensional spatial distribution) acquired by the data acquisition unit 11. Specifically, when an abandonment detection start instruction is output from the selection unit 131, the abandonment detection unit 132 detects whether an object is left inside the vehicle cabin based on the three-dimensional spatial distribution.

[0043] An example of a method by which the abandonment detection unit 132 detects whether an object is left behind in the vehicle cabin based on the three-dimensional spatial distribution will be described. For example, the abandonment detection unit 132 determines whether a moving object that is considered to be an infant is present in the vehicle cabin based on the three-dimensional spatial distribution. In the first embodiment, the determination by the abandonment detection unit 132 of "whether a moving object that is considered to be an infant is present in the vehicle cabin" more specifically means "whether a moving object that is considered to be an infant that has been left behind is present in the vehicle cabin." In other words, the abandonment detection unit 132 determines "whether only moving objects that are considered to be infants are present in the vehicle cabin."

[0044] The abandoned object detection unit 132 first determines whether a moving object deemed to be an occupant is present in the vehicle cabin in a portion of the three-dimensional spatial distribution corresponding to the area inside the vehicle cabin, and if so, in which seat the moving object deemed to be an occupant is located. Each grid of the three-dimensional spatial distribution is associated with coordinates in the real space inside the vehicle cabin, which are expressed in a three-dimensional coordinate system representing the real space inside the vehicle cabin. The abandoned object detection unit 132 can identify an area in the three-dimensional spatial distribution that corresponds to the area inside the vehicle cabin. Hereinafter, the three-dimensional spatial distribution of the portion corresponding to the area inside the vehicle cabin will be referred to as the "interior three-dimensional spatial distribution inside the vehicle cabin."

[0045] The abandonment detection unit 132 determines whether the spatial distribution included in the three-dimensional spatial distribution in the vehicle interior is a spatial distribution corresponding to an occupant (hereinafter referred to as an "occupant spatial distribution") based on the shape of the spatial distribution. Any method may be used for the determination based on the shape of the spatial distribution. For example, a model that learns the shape of the occupant spatial distribution may be used to determine whether the spatial distribution included in the three-dimensional spatial distribution in the vehicle interior is an occupant spatial distribution. Furthermore, the abandonment detection unit 132 may determine that the spatial distribution is an occupant spatial distribution if the relative velocity corresponding to the spatial distribution included in the three-dimensional spatial distribution in the vehicle interior is greater than a predetermined threshold, for example. Note that a relative velocity is associated with each grid of the three-dimensional spatial distribution. Even if the occupant is, for example, a small child sleeping in a child car seat, there is a certain degree of movement. If the abandonment detection unit 132 determines that the spatial distribution included in the three-dimensional spatial distribution in the vehicle interior is an occupant spatial distribution, it determines that a moving object considered to be an occupant is present in the vehicle interior.

[0046] Furthermore, the abandonment detection unit 132 determines whether the detected moving object regarded as an occupant is an adult or an infant based on the occupant space distribution included in the three-dimensional spatial distribution in the vehicle interior. Any method may be used for the determination based on the size of the occupant space distribution. For example, a conceivable mode is to determine whether the moving object regarded as an occupant is an adult or an infant based on the size of the occupant space distribution. Another conceivable mode is to determine whether the moving object regarded as an occupant is an adult or an infant using a model that has learned the size of the occupant space distribution for adults and infants. When there are multiple moving objects regarded as occupants, the abandonment detection unit 132 determines whether each of the moving objects is an adult or an infant.

[0047] The abandonment detection unit 132 then determines whether or not there is a moving object determined to be an infant among the moving objects determined to be occupants. If there is a moving object determined to be an infant among the moving objects determined to be occupants, the abandonment detection unit 132 determines whether or not there is any moving object determined to be an adult in addition to the moving object determined to be an infant. If there are no moving objects determined to be adults in addition to the moving objects determined to be infants, the abandonment detection unit 132 detects that only moving objects determined to be infants are present in the vehicle cabin. In other words, the abandonment detection unit 132 detects that an object has been abandoned in the vehicle cabin. On the other hand, if there are moving objects determined to be adults in addition to the moving objects determined to be infants, or if there are no moving objects determined to be infants among the moving objects determined to be occupants, that is, if all moving objects determined to be occupants are adults, the abandonment detection unit 132 detects that there is not a state in which only moving objects determined to be infants are present in the vehicle cabin. In other words, the abandonment detection unit 132 detects that no object has been abandoned in the vehicle cabin.

[0048] The abandoned object detection unit 132 executes the process of detecting whether an object has been left inside the vehicle cabin as described above at a preset cycle (hereinafter referred to as the "first cycle").

[0049] The abandonment detection unit 132 outputs an abandonment detection result indicating whether or not an object has been left inside the vehicle cabin to the detection result output unit 14. The abandonment detection unit 132 may output the abandonment detection result to the selection unit 131. Furthermore, when the abandonment detection unit 132 detects whether or not an object has been left inside the vehicle cabin a preset number of times, for example, which is one or more times, the abandonment detection unit 132 may output an abandonment detection completion notification to the selection unit 131 when the abandonment detection unit 132 has completed the detection of whether or not an object has been left inside the vehicle cabin.

[0050] When the abandoned vehicle detection is performed based solely on the presence or absence of a moving object in the vehicle cabin, it may be possible to detect that an occupant other than a person requiring assistance (in this case, an infant) has been abandoned, such as if the abandoned occupant is an adult. As a result, for example, an over-alarm problem may occur, in which an alarm indicating abandoned vehicle is output even when an alarm is not necessary. The detection result output unit 14 controls the output of the alarm. Details of the detection result output unit 14 will be described later. Therefore, the abandoned vehicle detection unit 132 not only detects the presence or absence of a moving object in the vehicle cabin, but also, if a moving object is present in the vehicle cabin, detects whether the moving object is a child for which an alarm is to be output. This prevents the abandoned vehicle detection unit 132 from accidentally detecting that an occupant who should not be considered abandoned, i.e., a person not requiring assistance, is abandoned in the vehicle cabin.

[0051] In the above example, the abandonment detection unit 132 determines whether the detected moving object considered to be an occupant is an adult or an infant by determining the physique based on the occupant spatial distribution. However, this is merely an example. The abandonment detection unit 132 may determine whether the detected moving object considered to be an occupant is an adult or an infant by other methods. For example, the abandonment detection unit 132 may determine whether the detected moving object considered to be an occupant is an adult or an infant by detecting the breathing rate based on the occupant spatial distribution. For example, the abandonment detection unit 132 may acquire an image of the interior of the vehicle cabin captured by a camera (not shown) connected to the abnormality detection device 1 and perform a known image recognition process on the image to determine whether the detected moving object considered to be an occupant is an adult or an infant. For example, the abandonment detection unit 132 may estimate the age of the occupant from the face of the occupant recognized in the captured image and determine whether the moving object considered to be an occupant is an adult or an infant. Furthermore, for example, the abandonment detection unit 132 may determine whether the moving object regarded as an occupant is an adult or an infant by detecting crying based on audio data acquired from a microphone (not shown) connected to the abnormality detection device 1. In this way, the abandonment detection unit 132 can determine whether the detected moving object regarded as an occupant is an adult or an infant using various known methods. The abandonment detection unit 132 may determine whether the detected moving object regarded as an occupant is an adult or an infant by combining various known methods.

[0052] When the selection unit 131 selects execution of detection of intrusion into the vehicle compartment, the intrusion detection unit 133 of the abnormality detection unit 13 detects the presence or absence of intrusion into the vehicle compartment based on the detection data (here, the three-dimensional spatial distribution) acquired by the data acquisition unit 11. Specifically, when an instruction to start intrusion detection is output from the selection unit 131, the abandonment detection unit 132 detects the presence or absence of intrusion into the vehicle compartment based on the three-dimensional spatial distribution.

[0053] An example of a method for detecting whether or not an intrusion has occurred in the vehicle cabin based on the three-dimensional spatial distribution by the intrusion detection unit 133 will be described below. For example, when the intrusion detection unit 133 detects a transition from a state in which no moving object is present in the vehicle cabin to a state in which a moving object is present in the vehicle cabin based on the time-series three-dimensional spatial distribution stored in the storage unit, the intrusion detection unit 133 considers the moved moving object to be a person, in other words, an intruder, and detects that an intrusion has occurred in the vehicle cabin.

[0054] In the above example, the intrusion detection unit 133 detects whether or not an intrusion has occurred in the vehicle cabin based on the spatial distribution of occupants, but this is merely an example. The intrusion detection unit 133 may detect whether or not an intrusion has occurred in the vehicle cabin using other methods. For example, the intrusion detection unit 133 may detect whether or not an intrusion has occurred in the vehicle cabin based on vibration data acquired from a vibration sensor (not shown) connected to the abnormality detection device 1. In this way, the intrusion detection unit 133 can detect whether or not an intrusion has occurred in the vehicle cabin using various known methods. The intrusion detection unit 133 may detect whether or not an intrusion has occurred in the vehicle cabin by combining various known methods.

[0055] The intrusion detection unit 133 executes the process of detecting whether or not an intrusion has occurred in the vehicle compartment as described above at a predetermined cycle (hereinafter referred to as the "second cycle"), where it is assumed that the length of the first cycle is the same as the length of the second cycle.

[0056] The intrusion detection unit 133 outputs a detection result of whether or not an intrusion has occurred in the vehicle interior (hereinafter referred to as an “intrusion detection result”) to the detection result output unit 14. The intrusion detection unit 133 may output the intrusion detection result to the selection unit 131.

[0057] The presence or absence of an intrusion into the vehicle interior can be realized as a function if it is possible to detect the presence or absence of a moving object present in the vehicle interior or around the vehicle 100. The intrusion detection unit 133 is not required to detect the attributes of the moving object (for example, whether the moving object is an adult or a child), as is the case with the abandonment detection unit 132 which detects the presence or absence of an abandoned object in the vehicle interior.

[0058] Here, Figure 5A is a diagram for explaining the concept of an example in which the abandonment detection unit 132 detects whether or not an object has been left behind in the vehicle cabin using a three-dimensional spatial distribution in embodiment 1, and Figure 5B is a diagram for explaining the concept of an example in which the intrusion detection unit 133 detects whether or not an object has been intruded into the vehicle cabin using a three-dimensional spatial distribution.

[0059] In FIG. 5A , the left side shows an example of a top view of the three-dimensional spatial distribution, and the right side shows an example of a side view of the three-dimensional spatial distribution. In the example of the top view of the three-dimensional spatial distribution shown on the left, an occupant is indicated by “D4,” and in the example of the side view of the three-dimensional spatial distribution shown on the right, an occupant is indicated by “D5.” Note that the occupant indicated by “D4” and the occupant indicated by “D5” are the same occupant. The abandonment detection unit 132 can detect the number and position of moving objects present in the vehicle cabin, in other words, occupants, based on the top view of the three-dimensional spatial distribution shown on the left. For example, in the example shown in FIG. 5A , the abandonment detection unit 132 can detect the presence of an occupant in the left rear seat in the vehicle cabin based on the top view of the three-dimensional spatial distribution on the left. However, the abandonment detection unit 132 cannot detect whether the occupant is an occupant who is considered to have been left behind, in other words, a child, based on only the top view of the three-dimensional spatial distribution. Therefore, the abandonment detection unit 132 also uses a side view of the three-dimensional spatial distribution, such as the one shown on the right side of FIG. 5A , to detect whether something has been left behind in the vehicle cabin, in other words, whether a small child is present in the vehicle cabin. By using the side view of the three-dimensional spatial distribution, the abandonment detection unit 132 can detect, for example, the occupant's sitting height and other physical characteristics. For example, in the example of FIG. 5A , the abandonment detection unit 132 detects from the top view of the three-dimensional spatial distribution that an occupant is present only in the left rear seat, and from the side view of the three-dimensional spatial distribution, detects that the occupant is a small child in a child seat. Looking at the side view of the three-dimensional spatial distribution, it can be seen that the occupant is positioned above the seat surface and that the occupant's physical characteristics are those of a small child. Therefore, the abandonment detection unit 132 detects that something has been left behind in the vehicle cabin.

[0060] In FIG. 5B , the three-dimensional spatial distribution on the left, the three-dimensional spatial distribution in the middle, and the three-dimensional spatial distribution on the right are top views of the time-series three-dimensional spatial distribution in this order. In the example of the top view of the three-dimensional spatial distribution shown on the left, the intruder is indicated by “D1.” In the example of the top view of the three-dimensional spatial distribution shown in the middle, the intruder is indicated by “D2.” In the example of the top view of the three-dimensional spatial distribution shown on the right, the intruder is indicated by “D3.” Note that the intruders indicated by “D1,” “D2,” and “D3” are the same person. Based on the top view of the time-series three-dimensional spatial distribution shown in FIG. 5B , the intrusion detection unit 133 detects that a moving object (intruder) has approached the vehicle 100 and entered the vehicle cabin when no occupant is present in the vehicle cabin. In other words, the intrusion detection unit 133 detects an intrusion into the vehicle cabin. In order to prevent an intrusion into the vehicle interior, for example, the intrusion detection unit 133 may detect an intrusion into the vehicle interior when a moving object is present for a certain period of time around the vehicle 100, assuming that there is a suspicious person attempting to intrude. In this case as well, the intrusion detection unit 133 can detect whether or not a moving object has been present for a certain period of time around the vehicle 100, based on a top view of the time-series three-dimensional spatial distribution.

[0061] In this way, the abandonment detection unit 132 and the intrusion detection unit 133 can each detect whether an object has been left in the vehicle cabin and whether an object has entered the vehicle cabin using the three-dimensional spatial distribution acquired from the common sensor 2, here a radio wave sensor. However, to detect whether an object has been left in the vehicle cabin, the abandonment detection unit 132 needs to detect attributes such as the occupant's physique and whether the occupant is an adult or a child. Therefore, the processing load when the abandonment detection unit 132 detects whether an object has been left in the vehicle cabin is higher than when the intrusion detection unit 133 detects whether an object has entered the vehicle cabin. On the other hand, the intrusion detection unit 133 can detect intrusion using only information from the three-dimensional spatial distribution viewed from above the vehicle 100. Furthermore, the intrusion detection unit 133 can detect whether an object has entered the vehicle cabin if it is known whether the moving object is inside or outside the vehicle cabin. For these reasons, the intrusion detection unit 133 can detect whether or not an intrusion has occurred inside the vehicle cabin by, for example, limiting the range in which it detects whether or not an intrusion has occurred inside the vehicle cabin to the vicinity of the boundary between the inside and outside of the vehicle cabin in the three-dimensional spatial distribution.

[0062] Returning to the description of the configuration example of the anomaly detection device 1 using FIG.

[0063] The detection result output unit 14 outputs to the output device 3 information (hereinafter referred to as "output control information") based on the detection result of whether or not an object has been left in the vehicle cabin by the abandonment detection unit 132, or output control information based on the detection result of whether or not an object has been intruded into the vehicle cabin by the intrusion detection unit 133.

[0064] FIG. 6 is a diagram illustrating a detailed configuration example of the detection result output unit 14 of the anomaly detection device 1 according to the first embodiment.

[0065] The detection result output unit 14 includes a detection result storage unit 141 , a notification content generation unit 142 , and an output control unit 143 .

[0066] The detection result storage unit 141 stores the abandonment detection result output from the abandonment detection unit 132 of the anomaly detection unit 13 or the intrusion detection result output from the intrusion detection unit 133 for a predetermined period of time. Note that, here, as shown in FIG. 6 , the detection result storage unit 141 is assumed to be provided in the anomaly detection device 1, but this is merely an example. The detection result storage unit 141 may be provided in a location outside the anomaly detection device 1 that can be referenced by the anomaly detection device 1. Furthermore, the detection result storage unit 141 may be common to the above-mentioned storage unit.

[0067] The notification content generation unit 142 generates output control information based on the abandonment detection result output from the abandonment detection unit 132 of the abnormality detection unit 13 or the intrusion detection result output from the intrusion detection unit 133. In the first embodiment, the output control information is information for causing the output device 3 to perform an operation in response to an abandonment in the vehicle cabin or an intrusion into the vehicle cabin. The notification content generation unit 142 determines an operation to be performed by the output device 3 based on, for example, the abandonment detection result or the intrusion detection result, and generates the output control information. For example, the operation to be performed by the output device 3 is, for example, a notification that an object has been abandoned in the vehicle cabin or that an intrusion has occurred into the vehicle cabin. The notification content generation unit 142 generates information to cause the output device 3 to perform a notification as the output control information. For example, when causing the output device 3 to output some information (e.g., an alarm), the notification content generation unit 142 may determine the output form and include information instructing the determined output form in the output control information. Note that the notification content generation unit 142 may obtain the abandonment detection result or the intrusion detection result directly from the abandonment detection unit 132 or the intrusion detection unit 133, or may obtain it from the detection result storage unit 141.

[0068] For example, it is assumed that the output device 3 is a display device. For example, when the notification content generation unit 142 acquires an abandonment detection result indicating that an object has been left inside the vehicle cabin, the notification content generation unit 142 generates output control information for causing the output device 3 to display a message notifying that an object has been left inside the vehicle cabin. Furthermore, for example, when the notification content generation unit 142 acquires an intrusion detection result indicating that an intrusion has occurred inside the vehicle cabin, the notification content generation unit 142 generates output control information for causing the output device 3 to display a message notifying that an intrusion has occurred inside the vehicle cabin or that there has been suspicious activity around the vehicle 100. The output device 3 performs display in accordance with the output control information.

[0069] For example, it is assumed that the output device 3 is an audio output device. For example, when the notification content generation unit 142 acquires an abandonment detection result indicating that an object has been left inside the vehicle cabin, the notification content generation unit 142 generates output control information for causing the output device 3 to output audio or an alarm sound notifying that the object has been left inside the vehicle cabin. Furthermore, for example, when the notification content generation unit 142 acquires an intrusion detection result indicating that an intrusion has occurred inside the vehicle cabin, the notification content generation unit 142 generates output control information for causing the output device 3 to output audio or an alarm sound notifying that an intrusion has occurred inside the vehicle cabin or that there has been suspicious activity around the vehicle 100. The output device 3 outputs audio or sound in accordance with the output control information.

[0070] For example, assume that the output device 3 is a driving control device. For example, when the notification content generation unit 142 acquires an abandonment detection result indicating that an infant has been left inside the vehicle cabin, it generates output control information to notify the output device 3 that abandonment has occurred. Furthermore, for example, when the notification content generation unit 142 acquires an intrusion detection result indicating that an intrusion has occurred inside the vehicle cabin, it generates output control information to notify the output device 3 that an intrusion has occurred inside the vehicle cabin or that there has been suspicious activity around the vehicle 100. The output device 3 performs vehicle control in accordance with the output control information. For example, when the output control information indicating that an infant has been left inside the vehicle cabin is output, the output device 3 performs control such as opening the windows of the vehicle 100 or turning on the air conditioning. If an infant has been left inside the vehicle cabin, it is possible that the left infant has suffered from heatstroke or the like. On the other hand, for example, when the output control information indicating that an intrusion has occurred inside the vehicle cabin is output, the output device 3 performs control to lock the doors of the vehicle 100.

[0071] In this way, the notification content generation unit 142 can generate output control information that causes the output device 3 to perform different control operations depending on whether an abandonment detection result or an intrusion detection result has been output.

[0072] The notification content generation unit 142 may also include supplemental information, such as an image of the interior of the vehicle cabin in which an intruder is captured, or information indicating the location of the abandoned child, in the output control information. The notification content generation unit 142 may, for example, acquire an image of the interior of the vehicle cabin from a camera (not shown) mounted on the vehicle 100. The notification content generation unit 142 may, for example, acquire the three-dimensional spatial distribution used to detect whether or not an infant has been left behind from the abandonment detection unit 132, and identify the location of the abandoned child from the acquired three-dimensional spatial distribution.

[0073] The alarm content generation unit 142 may also generate output control information that increases the level of the alarm depending on the amount of time that has elapsed since an object was detected to be left in the vehicle cabin or since an intrusion into the vehicle cabin was detected. For example, the alarm content generation unit 142 generates output control information that increases the level of the alarm as the amount of time that has elapsed since an object was detected to be left in the vehicle cabin or since an intrusion into the vehicle cabin increases. The alarm content generation unit 142 can acquire past abandonment detection results or intrusion detection results by referring to the detection result storage unit 141, and can determine the amount of time that has elapsed since an object was detected to be left in the vehicle cabin or since an intrusion into the vehicle cabin was detected. In this way, the alarm content generation unit 142 can also generate output control information that gradually changes the output mode.

[0074] For example, when it is detected that an object has been left inside the vehicle cabin, it may be that the owner of the vehicle 100 has intentionally temporarily left the vehicle 100. The notification content generation unit 142 generates output control information that raises the level of the alarm as the time that has elapsed since it was detected that an object has been left inside the vehicle cabin increases, so that the output device 3 issues a low-level alarm immediately after detecting that the object has been left inside the vehicle cabin, and raises the level of the alarm as the time that has elapsed since the object has been left inside the vehicle cabin increases.

[0075] The notification content generation unit 142 outputs the generated output control information to the output control unit 143 .

[0076] The output control unit 143 outputs the output control information output from the notification content generation unit 142 to the output device 3. At this time, the output control unit 143 controls the output destination of the output control information in accordance with the content of the output control information.

[0077] For example, if the output control information is output control information related to an intrusion into the vehicle interior, the output control unit 143 outputs the output control information only to the output device 3 provided outside the vehicle 100. An intrusion into the vehicle interior occurs when no occupants are inside the vehicle interior, so it is expected that no one will notice the alert even if it is issued inside the vehicle interior. By outputting the output control information only to the output device 3 provided outside the vehicle 100, the output control unit 143 can prevent the output of unnecessary output control information.

[0078] For example, if the output control information is output control information regarding abandonment in the vehicle cabin, the output control unit 143 outputs the output control information not only to the output device 3 provided outside the vehicle 100 but also to the output device 3 provided inside the vehicle cabin. By outputting the output control information to the output device 3 provided inside the vehicle cabin, the output control unit 143 can notify the abandoned infant that they have been abandoned.

[0079] The output control unit 143 can combine various output devices 3 as the output destination of the output control information. The output control unit 143 outputs the output control information output from the notification content generation unit 142 to the output device 3, and causes the output device 3 to perform control in accordance with the output control information, so that the anomaly detection device 1 can allow the owner of the vehicle 100 or the like to understand the cause of the anomaly occurring in the vehicle 100, more specifically, whether the vehicle has been abandoned or whether an intrusion has occurred.

[0080] Furthermore, the output control unit 143 may output information for stopping an alarm or the like based on the output control information (hereinafter referred to as "stop control information") to the output device 3. For example, when an input device (not shown) that accepts an instruction to stop an alarm accepts an instruction to stop the alarm, the output control unit 143 acquires the stop instruction and outputs the stop control information to the output device 3. When the stop control information is output, the output device 3 stops the control that has been performed based on the output control information, such as an alarm.

[0081] The output control unit 143 may change the control content instructed by the stop control information depending on whether the output device 3 is being caused to output based on the output control information generated based on the abandonment detection result or the intrusion detection result. For example, if the output control unit 143 receives a stop instruction after causing the output device 3 to output an alarm based on the output control information generated based on the abandonment detection result, the output control unit 143 unconditionally stops the output of the alarm that the output device 3 has been caused to output based on the output control information. On the other hand, if the output control unit 143 receives a stop instruction after causing the output device 3 to output an alarm based on the output control information generated based on the intrusion detection result, the output control unit 143 prevents the output device 3 from stopping the output of the alarm that the output device 3 is performing based on the output control information unless the key for the vehicle 100 is present near the vehicle 100. In this way, the output control unit 143 can, for example, prevent an intruder from stopping the alarm.

[0082] The input device that accepts the stop instruction may be mounted on the vehicle 100 or provided outside the vehicle. The input device may be, for example, a communication device. For example, even if the stop instruction is input by remote operation, the output control unit 143 can accept the stop instruction and stop the control that the output device 3 has been performing based on the output control information, thereby further improving the convenience of the anomaly detection device 1.

[0083] The above-described content of the output control information generated by the detection result output unit 14 is merely an example, and the detection result output unit 14 may generate output control information with content other than the above-described content and control the output device 3 accordingly. For example, the detection result output unit 14 may generate output control information with content that complies with the laws and regulations or assessments of the region in which the vehicle 100 is used and control the output device 3 accordingly.

[0084] The following describes the operation of the anomaly detection device 1 according to embodiment 1. Fig. 7 is a flowchart for explaining an example of the operation of the anomaly detection device 1 according to embodiment 1.

[0085] First, some examples of the timing at which the anomaly detection device 1 starts and ends the operation shown in the flowchart of FIG. 7 will be described.

[0086] <Timing Example (1)> For example, the abnormality detection device 1 starts operation when the doors of the vehicle 100 are locked, and then repeats the operation shown in the flowchart of Fig. 7 until the doors are unlocked. For example, the control unit (not shown) of the abnormality detection device 1 acquires door data from a door sensor, and when it detects that the doors are locked, it instructs the data acquisition unit 11, the abnormality detection unit 13, and the detection result output unit 14 of the abnormality detection device 1 to start operation, and when it detects that the doors are unlocked, it instructs the data acquisition unit 11, the abnormality detection unit 13, and the detection result output unit 14 to end operation.

[0087] <Timing Example (2)> Because it is assumed that an occupant of the vehicle 100 may leave the vehicle 100 without locking the doors, for example, the abnormality detection device 1 may start its operation in response to a trigger such as the opening or closing of a door of the vehicle 100 after the vehicle 100 has stopped, and may repeat the operation shown in the flowchart of Fig. 7 until the door is opened again. For example, the control unit of the abnormality detection device 1 detects that the vehicle 100 has stopped from a vehicle speed sensor (not shown) or a shift position sensor (not shown) mounted on the vehicle 100, and detects the opening or closing of the door from the door data.

[0088] <Timing Example (3)> For example, the abnormality detection device 1 may start operation when triggered by the owner of the vehicle 100 leaving the vicinity of the vehicle 100, and may repeat the operation shown in the flowchart of FIG. 7 until the owner of the vehicle 100 approaches (returns) to the vicinity of the vehicle 100. The control unit may detect whether the owner of the vehicle 100 is present in the vicinity of the vehicle 100 using various known methods, such as a method using a signal from the key of the vehicle 100. The control unit may detect whether the owner of the vehicle 100 is present in the vicinity of the vehicle 100, for example, using a three-dimensional spatial distribution acquired by the data acquisition unit 11. The control unit can determine that the owner of the vehicle 100 has left the vehicle 100 by detecting whether an occupant of the vehicle 100 is present in the vicinity of the vehicle 100 immediately after getting off the vehicle 100 and using information such as the disappearance of a biological reaction after getting off the vehicle 100.

[0089] <Timing Example (4)> For example, the abnormality detection device 1 may start operation when triggered by the vehicle 100 entering a traveling state or an occupant getting into the vehicle 100, and may repeat the operation shown in the flowchart of Fig. 7 until the vehicle 100 enters a traveling state again after the door of the vehicle 100 is opened or closed, or until an occupant gets into the vehicle 100. The control unit can detect that the vehicle 100 has entered a traveling state, for example, from a vehicle speed sensor. Furthermore, the control unit can detect that an occupant has gotten into the vehicle 100, for example, from a seat occupancy sensor (not shown) mounted on the vehicle 100.

[0090] For example, the abnormality detection device 1 may combine the timings described in the above examples with respect to the timing for starting or ending the operation shown in the flowchart of Fig. 7. For example, the abnormality detection device 1 may combine <Timing example (1)> and <Timing example (2)>, and start its operation when triggered by the vehicle 100 coming to a stop and the door of the vehicle 100 being opened or closed, and repeat the operation shown in the flowchart of Fig. 7 until the door is next unlocked.

[0091] Furthermore, the above-described example is merely an example, and the abnormality detection device 1 may start or end the operation shown in the flowchart of Fig. 7 at timings other than those described above. For example, the abnormality detection device 1 may set the timings for starting or ending the operation shown in the flowchart of Fig. 7 in accordance with the laws and regulations or assessments of the area in which the vehicle 100 is used.

[0092] When the anomaly detection device 1 starts operating at the timing described in the example above, the anomaly detection unit 13 in the anomaly detection device 1 selectively switches between the abandonment detection mode and the intrusion detection mode until the anomaly detection device 1 finishes operating.

[0093] For example, the abnormality detection device 1 starts operation when the doors of the vehicle 100 are locked (see <Timing Example (1)>). The selection unit 131 selects the abandonment detection mode until the switching time has elapsed, and the abandonment detection unit 132 executes detection of whether or not an object has been left inside the vehicle cabin. If abandonment is not detected even after the switching time has elapsed, the selection unit 131 switches from the abandonment detection mode to the intrusion detection mode. The abandonment detection unit 132 ends detection of whether or not an object has been left inside the vehicle cabin, and the intrusion detection unit 133 executes detection of whether or not an intrusion has occurred into the vehicle cabin (see <Selection Example (1)> above). Next, until the doors of the vehicle 100 are unlocked, the selection unit 131 continues to select the intrusion detection mode because the switching time has elapsed. When the doors of the vehicle 100 are unlocked, the abnormality detection device 1 ends operation. That is, both the process of detecting whether or not an object has been left inside the vehicle cabin and the process of detecting whether or not an intrusion has occurred into the vehicle cabin end.

[0094] Also, for example, the abnormality detection device 1 starts operating when the vehicle 100 is in a traveling state (see <Timing Example (4)>). Here, the selection unit 131 may cause the abandonment detection unit 132 to detect whether or not a small child, who is the target of abandoned detection, is present in the vehicle cabin. If the abandonment detection unit 132 detects that no small child is present in the vehicle cabin, the selection unit 131 may thereafter not select the abandonment detection mode, but may select the intrusion detection mode at the timing of switching the selection (for example, after the switching time has elapsed).

[0095] An example of the operation of the abnormality detection device 1 shown in the flowchart of FIG. 7 will now be described in detail.

[0096] The data processing unit 12 performs pre-processing on the sensor data acquired from the sensor 2, which is performed in common by the abnormality detection device 1 when detecting whether an object has been left in the vehicle cabin and when detecting whether an object has entered the vehicle cabin, and generates detection data, in this case a three-dimensional spatial distribution (step ST1).

[0097] The data acquiring unit 11 acquires the detection data (here, the three-dimensional spatial distribution) generated by the data processing unit 12 in step ST1 (step ST2). The data acquiring unit 11 outputs the acquired detection data (here, the three-dimensional spatial distribution) to the anomaly detecting unit 13.

[0098] The selection unit 131 of the abnormality detection unit 13 selects whether to perform abandonment detection in the vehicle interior or intrusion detection into the vehicle interior (step ST3). That is, the selection unit 131 selects whether to use the abandonment detection mode or the intrusion detection mode. If the selection unit 131 selects the abandonment detection mode, it outputs an abandonment detection start instruction to the abandonment detection unit 132. If the selection unit 131 selects the intrusion detection mode, it outputs an intrusion detection start instruction to the intrusion detection unit 133.

[0099] If the selection unit 131 selects execution of vehicle interior abandonment detection in step ST3, the abandonment detection unit 132 of the abnormality detection unit 13 detects whether or not an object is abandoned in the vehicle interior based on the detection data (here, the three-dimensional spatial distribution) acquired by the data acquisition unit 11 in step ST2. If the selection unit 131 selects execution of vehicle interior intrusion detection in step ST3, the intrusion detection unit 133 of the abnormality detection unit 13 detects whether or not an object is abandoned in the vehicle interior based on the detection data (here, the three-dimensional spatial distribution) acquired by the data acquisition unit 11 in step ST2 (step ST4). If the abandonment detection unit 132 operates, the abandonment detection unit 132 outputs the abandonment detection result to the detection result output unit 14. If the intrusion detection unit 133 operates, the intrusion detection unit 133 outputs the intrusion detection result to the detection result output unit 14.

[0100] The detection result output unit 14 outputs output control information based on the detection result of whether or not an object has been left in the vehicle compartment by the abandonment detection unit 132 in step ST4, or output control information based on the detection result of whether or not an object has been intruded into the vehicle compartment by the intrusion detection unit 133, to the output device 3 (step ST5). The output device 3 performs control in accordance with the output control information.

[0101] 7 may be repeatedly executed at the operation timing described above until the operation ends, or, alternatively, if the anomaly detection device 1 detects an abnormality (i.e., if an object is left in the vehicle compartment or an intrusion into the vehicle compartment is detected and output control information is output to the output device 3), the processing of the anomaly detection device 1 may be terminated even if the operation end timing has not yet arrived. By continuing the repeated processing even after detecting an abnormality, the anomaly detection device 1 can perform step-by-step control such as issuing an alarm to the output device 3.

[0102] In this way, the anomaly detection device 1 acquires detection data related to a moving object present in the vehicle cabin, which is generated based on sensor data acquired by the sensor 2 when the sensor 2 detects an object present in the vehicle cabin, and selects whether to perform abandonment detection in the vehicle cabin or intrusion detection into the vehicle cabin. If the anomaly detection device 1 selects to perform abandonment detection in the vehicle cabin, it detects whether an object has been abandoned in the vehicle cabin based on the acquired detection data, and if the anomaly detection device 1 selects to perform intrusion detection into the vehicle cabin, it detects whether an intrusion into the vehicle cabin has occurred based on the acquired detection data.

[0103] As described above, when detecting abandonment in the vehicle cabin, the abnormality detection device 1 needs to detect not only the presence or absence of a moving object in the vehicle cabin but also whether the moving object in the vehicle cabin is a child for which an alarm should be output, so as not to detect the presence of an adult other than a child as abandoned. If the abnormality detection device 1 does not determine whether the abandoned occupant is a child for which an alarm should be output, the abnormality detection device 1 may mistakenly detect, for example, an adult resting in the vehicle cabin as abandoned. As a result, the abnormality detection device 1 may cause the output device 3 to output unnecessary alarms, resulting in an over-alarm. On the other hand, when detecting intrusion into the vehicle cabin, the abnormality detection device 1 is only required to detect the presence or absence of a moving object in the vehicle cabin or around the vehicle 100; it is not necessary to detect whether the moving object is an adult or a child. In other words, the abandonment detection process is more complex than the intrusion detection process. Furthermore, because abandonment detection can potentially be life-threatening, higher accuracy is required in terms of detection performance compared to intrusion detection. Therefore, in general, the detection of an object being left in a vehicle interior tends to require a higher processing load than the detection of an intrusion into the vehicle interior, and as a result, the detection of an object being left in a vehicle interior tends to require a higher power consumption for processing than the detection of an intrusion into the vehicle interior.

[0104] In contrast, the anomaly detection device 1 selectively switches between detecting whether an object has been left in the vehicle cabin and detecting whether an intrusion has occurred into the vehicle cabin, and can therefore simultaneously detect whether an intrusion has occurred into the target area and whether an infant or other such object has been left in the target area while suppressing the execution of unnecessary processes. As a result, the anomaly detection device 1 can achieve lower power consumption than the conventional technology described above, which continuously detects whether an object has been left in the vehicle cabin and whether an intrusion has occurred into the vehicle cabin.

[0105] Furthermore, for example, in the case of the conventional technology described above, where both abandonment detection and intrusion detection functions are simultaneously realized based solely on the detection of a moving object deemed to be a living body in the vehicle cabin, even if an alarm is output, the user may not be able to determine whether the alarm is due to abandonment or intrusion. Thus, there is a problem of low usability, such as the user not understanding the cause of a notification even when receiving such a notification. In response to this problem, the abnormality detection device 1 can selectively switch between abandonment detection in the vehicle cabin and intrusion detection into the vehicle cabin, and, if abandonment in the vehicle cabin or intrusion into the vehicle cabin is detected, the output device 3 can be operated in a manner that allows the user to understand this. Therefore, the abnormality detection device 1 can allow the user to understand whether a notification is a notification of abandonment in the vehicle cabin or a notification of intrusion into the vehicle cabin.

[0106] As described above, detecting whether an object has been left behind requires more complex processing and higher accuracy than detecting whether an object has been intruded. From the perspective of the sensor 2, detecting whether an object has been intruded does not require as detailed sensing as detecting whether an object has been left behind. Therefore, in the above-described first embodiment, the selection unit 131 may control the operation of the sensor 2 depending on whether to perform detection of whether an object has been left inside the vehicle compartment or detection of an intrusion into the vehicle compartment.

[0107] For example, the selection unit 131 may cause the sensor 2 to change the number of transmitting antennas or the transmission waveform between the abandonment detection mode and the intrusion detection mode. When detecting whether an object is abandoned in the vehicle cabin, the abandonment detection unit 132 needs to detect whether the detected moving object, which is considered to be an occupant, is an adult or an infant based on the occupant spatial distribution included in the three-dimensional spatial distribution in the vehicle cabin, as described above. For the abandonment detection unit 132 to accurately detect the difference in physique between an adult and an infant, the difference in distance and angle from the sensor 2 (more specifically, the radio wave sensor) to the head of an adult and an infant is important, and the sensor 2 is required to have high distance resolution and angle resolution. On the other hand, when the intrusion detection unit 133 detects whether an object is abandoned in the vehicle cabin, the sensor 2 does not need to have the same resolution as when the abandonment detection unit 132 detects whether an object is abandoned in the vehicle cabin. It is sufficient for the sensor 2 to roughly detect whether a biological reaction is inside or outside the vehicle cabin. Therefore, for example, the selection unit 131 changes the number of transmitting antennas or the transmission waveform of the sensor 2 between the abandoned detection mode and the intrusion detection mode.

[0108] 8A, 8B, and 8C are diagrams for explaining an example of control content when the selection unit 131 is able to control the operation of the sensor 2 in embodiment 1. Figures 8A, 8B, and 8C are diagrams for explaining an example of control content when the selection unit 131 is able to control the operation of the sensor 2 in embodiment 1.

[0109] For example, when the abandonment detection mode is selected, the selection unit 131 controls the sensor 2 to use two transmitting antennas Tx1 and Tx2 and transmit radio waves using frequency bandwidths Freq. 1 to Freq. 2 (see FIG. 8A). On the other hand, when the intrusion detection mode is selected, the selection unit 131 controls the sensor 2 to reduce the number of transmitting antennas (see FIG. 8B). This shortens the radio wave transmission time in the intrusion detection mode and eliminates the need for antenna operation control processing for the reduced number of transmitting antennas, enabling the anomaly detection device 1 to reduce power consumption of the sensor 2.

[0110] Furthermore, for example, the selector 131 may control the sensor 2 to limit the frequency bandwidth between Freq. 1 and Freq. 3 (see FIG. 8C ). This shortens the radio wave transmission time, allowing the anomaly detection device 1 to reduce the power consumption of the sensor 2.

[0111] The selector 131 may also combine the control of changing the transmission waveform for the sensor 2 as described above.

[0112] The selection unit 131 may change the anomaly detection cycle depending on, for example, the selection of the abandonment detection mode or the intrusion detection mode. For example, in the above-described first embodiment, the length of the cycle (first cycle) in which the abandonment detection unit 132 detects whether an object is abandoned in the vehicle cabin and the length of the cycle (second cycle) in which the intrusion detection unit 133 detects whether an object is intruded into the vehicle cabin are assumed to be the same. However, the selection unit 131 may make the length of the first cycle different from the length of the second cycle. Furthermore, for example, the selection unit 131 may change the anomaly detection cycle by changing the cycle in which the data processing unit 12 acquires sensor data from the sensor 2 depending on the selection of the abandonment detection mode or the intrusion detection mode. Note that the arrow from the selection unit 131 to the data processing unit 12 is omitted in FIG. 1 .

[0113] 9A, 9B, and 9C are diagrams illustrating the concept of the selector 131 changing the anomaly detection cycle in embodiment 1. FIG. 9B is a diagram illustrating an example of a three-dimensional spatial distribution generated based on sensor data obtained by a sensor 2 with a smaller number of antennas detecting an object present in the vehicle cabin. FIG. 9A is a diagram illustrating an example of a three-dimensional spatial distribution generated based on sensor data obtained by a sensor 2 with a larger number of antennas than that of FIG. 9B detecting an object present in the vehicle cabin. FIG. 9C is a diagram illustrating a three-dimensional spatial distribution in which the distribution of moving objects is superimposed on the three-dimensional spatial distribution illustrated in FIG. 9B. Note that FIGS. 9A, 9B, and 9C are all side views of the three-dimensional spatial distribution.

[0114] For example, suppose sensor 2 has a small number of antennas. In this case, as shown in FIG. 9B , the number of biological responses detectable in a single sensing operation is smaller than when sensor 2 has a large number of antennas, as shown in FIG. 9A . Therefore, it is difficult for the abnormality detection device 1 to obtain a three-dimensional spatial distribution capable of detecting the physique of an occupant left in the vehicle cabin in a single sensing operation using sensor 2. For example, as shown in FIG. 9C , sensor 2 continuously acquires sensor data at short intervals, and the abnormality detection device 1 generates a three-dimensional spatial distribution by overlaying the biological responses obtained at each interval, thereby clearly showing the silhouette of the occupant, thereby detecting whether or not an occupant has been left in the vehicle cabin. Therefore, when the selection unit 131 selects the abandoned vehicle detection mode, it instructs the data processing unit 12 to shorten the period for acquiring sensor data.

[0115] On the other hand, in the intrusion detection mode, it is not necessary to detect the physique of the occupant, but it is sufficient to know whether there is a biological reaction around or inside the vehicle cabin of the vehicle 100. Therefore, for example, in the intrusion detection mode, the selection unit 131 may lengthen the cycle for acquiring sensor data compared to the abandonment detection mode.

[0116] As described above, the selection unit 131 may change the period of the anomaly detection by varying the length of the period during which the data processing unit 12 acquires sensor data. For example, the selection unit 131 may change the length of a first period during which the abandonment detection unit 132 detects whether an object has been abandoned in the vehicle cabin and the length of a second period during which the intrusion detection unit 133 detects whether an intrusion has been detected into the vehicle cabin, without changing the period during which the data processing unit 12 acquires sensor data. This allows the anomaly detection device 1 to perform the abandonment detection unit 132's detection of whether an object has been abandoned in the vehicle cabin and the intrusion detection unit 133's detection of whether an intrusion has been detected into the vehicle cabin at the required period, thereby further reducing the execution of unnecessary processes. As a result, the anomaly detection device 1 can achieve lower power consumption. However, varying the period during which the data processing unit 12 acquires sensor data results in a greater effect of reducing power consumption when the detection period is lengthened, because this would also change, for example, the period during which the sensor data is acquired by the sensor 2 or the period during which a three-dimensional spatial distribution is generated.

[0117] Furthermore, in the above-described first embodiment, the functions of the data processing unit 12 may be provided by the sensor 2. In this case, the data acquisition unit 11 acquires detection data, in this case, a three-dimensional spatial distribution, from the sensor 2. The anomaly detection device 1 does not necessarily have to include the data processing unit 12. Furthermore, with regard to the operation of the anomaly detection device 1 described using the flowchart in FIG. 7 , the anomaly detection device 1 can omit the processing of step ST1.

[0118] In the first embodiment described above, the abandonment detection unit 132 and the intrusion detection unit 133 each detect whether an object is abandoned in the vehicle cabin or whether an object has entered the vehicle cabin based on common detection data, specifically a three-dimensional spatial distribution, based on sensor data acquired from a common sensor 2, specifically a radio wave sensor. However, this is merely an example. For example, the abandonment detection unit 132 may detect whether an object is abandoned in the vehicle cabin using detection data generated based on sensor data acquired by another type of sensor 2, in addition to the same three-dimensional spatial distribution as the detection data used by the intrusion detection unit 133 to detect whether an object has entered the vehicle cabin. For example, the abandonment detection unit 132 may acquire, as detection data, an image generated from a camera based on an image of the vehicle cabin captured by the camera and to which information about a moving object in the vehicle cabin is added (hereinafter referred to as a "post-moving-object-detection captured image"), in addition to the three-dimensional spatial distribution, and detect whether an object is abandoned in the vehicle cabin by combining the three-dimensional spatial distribution and the post-moving-object-detection captured image. The intrusion detection unit 133 detects whether or not an intrusion has occurred in the vehicle cabin using only the three-dimensional spatial distribution. When the selection unit 131 selects the intrusion detection mode, it may stop the operation of the camera. This allows the anomaly detection device 1 to detect whether or not an object has been left in the vehicle cabin with high accuracy, and to detect whether or not an intrusion has occurred in the vehicle cabin while omitting unnecessary data acquisition or processing and reducing power consumption.

[0119] In the above example, the abandonment detection unit 132 detects whether an object has been left in the vehicle cabin by using the same three-dimensional spatial distribution as the detection data used by the intrusion detection unit 133 to detect whether an object has been left in the vehicle cabin, as well as detection data generated based on sensor data acquired by another type of sensor 2. However, this is merely an example. For example, the intrusion detection unit 133 may detect whether an object has been left in the vehicle cabin by using the same three-dimensional spatial distribution as the detection data used by the intrusion detection unit 132 to detect whether an object has been left in the vehicle cabin, as well as detection data generated based on sensor data acquired by another type of sensor 2. For example, the intrusion detection unit 133 may acquire a post-motion detection captured image from a camera as detection data in addition to the three-dimensional spatial distribution, and detect whether an object has been left in the vehicle cabin by combining the three-dimensional spatial distribution and the post-motion detection captured image. The abandonment detection unit 132 detects whether an object has been left in the vehicle cabin by using only the three-dimensional spatial distribution. When the intrusion detection mode is selected, the selection unit 131 may cause the camera to start operating. This allows the anomaly detection device 1 to check in more detail the reaction outside the vehicle detected by the radio wave sensor from the image captured after motion detection when detecting whether or not an intrusion has occurred in the vehicle cabin. To give a specific example, even if a tree near the vehicle 100 sways and causes a reaction in the radio wave sensor, the anomaly detection device 1 can confirm from the image captured after motion detection that the reaction is not caused by an intruder. This allows the anomaly detection device 1 to detect whether or not an intrusion has occurred in the vehicle cabin with high accuracy and prevent overdetection of the intrusion, while also detecting whether or not an object has been left behind in the vehicle cabin while omitting unnecessary data acquisition or processing and reducing power consumption.

[0120] Furthermore, in the above-described first embodiment, the abandonment detection unit 132 and the intrusion detection unit 133 may each detect whether an object is abandoned in the vehicle cabin or whether an intrusion has occurred into the vehicle cabin using detection data generated based on sensor data acquired by different types of sensors 2. In this case, the abandonment detection unit 132 and the intrusion detection unit 133 do not use common detection data, and therefore the abandonment detection unit 132 and the intrusion detection unit 133 individually generate detection data when detecting whether an object is abandoned in the vehicle cabin or when detecting whether an intrusion has occurred into the vehicle cabin. The abnormality detection device 1 may be configured without the data processing unit 12, and the processing of step ST1 may be omitted in the operation of the abnormality detection device 1 described using the flowchart shown in FIG. 7. In this case, in the operation of the abnormality detection device 1 shown in the flowchart of FIG. 7, the abnormality detection device 1 may perform the processing of step ST3 before the processing of step ST2. As a result, in the abnormality detection device 1, the data acquisition unit 11 only needs to acquire the detection data necessary for the mode selected by the selection unit 131 (abandonment detection mode or intrusion detection mode), and the acquisition of unnecessary detection data can be omitted.

[0121] However, since the overall system configuration is simpler if the abandonment detection unit 132 and the intrusion detection unit 133 each detect whether an object is abandoned in the vehicle cabin or whether an intrusion has occurred into the vehicle cabin based on detection data that is based on sensor data acquired by a common sensor 2, it is desirable for the abandonment detection unit 132 and the intrusion detection unit 133 to use detection data that is generated based on sensor data acquired by the same sensor 2 as much as possible.

[0122] For example, as described in the first embodiment above, when a radio wave sensor is used as the sensor 2, the radio wave sensor can penetrate cloth decorations such as seat surfaces or blankets other than metal parts, making it possible to observe the situation inside the entire vehicle cabin. Furthermore, because radio waves from the radio wave sensor can also penetrate outside the vehicle through window frames and the like, the radio wave sensor can also observe the situation around the vehicle 100. Therefore, the abnormality detection device 1 can perform both the functions of detecting whether an object has been left inside the vehicle cabin and whether an intrusion has occurred inside the vehicle cabin, using detection data generated based on sensor data acquired by a common radio wave sensor.

[0123] Alternatively, for example, the common sensor 2 may be a camera, and the anomaly detection device 1 may perform both the functions of detecting whether an object has been left in the vehicle cabin and the function of detecting whether an intrusion has occurred into the vehicle cabin, using detection data (i.e., a post-motion-object-detection captured image) generated based on sensor data (i.e., a captured image) acquired by the common camera. The camera is installed in the vehicle cabin so as to capture an area where an occupant may be present. The camera can capture images of the interior of the vehicle cabin as well as the surroundings of the vehicle 100 through windows, etc. Note that, for example, when the common sensor 2 is a camera, the detailed configuration example of the data processing unit 12 does not need to be the configuration example shown in FIG. 2 . The data processing unit 12 may be configured to acquire captured images from the camera, detect a moving object in the captured images using known image recognition technology, and generate a post-motion-object-detection captured image.

[0124] In the first embodiment, the sensor 2 from which the sensor data is acquired remains unchanged from the start of operation to the end of operation of the anomaly detection device 1. However, this is merely an example, and the anomaly detection device 1 may be configured to add or change the sensor 2 from which the sensor data is acquired from the start of operation to the end of operation. For example, in the anomaly detection device 1, when the intrusion detection unit 133 detects an intrusion into the vehicle cabin based on detection data, i.e., three-dimensional spatial distribution, based on sensor data acquired by a radio wave sensor, the intrusion detection unit 133 considers this as a tentative detection and switches the sensor 2 from which the detection data is acquired from the radio wave sensor to a camera. That is, the intrusion detection unit 133 checks whether an intrusion into the vehicle cabin can be detected based on post-motion-detection captured images captured by the camera. When the intrusion detection unit 133 detects an intrusion of a moving object into the vehicle cabin based on the post-motion-detection captured images, the intrusion detection unit 133 determines that an intrusion into the vehicle cabin has been detected and outputs an intrusion detection result indicating an intrusion into the vehicle cabin to the detection result output unit 14. For example, when the intrusion detection unit 133 detects an intrusion into the vehicle cabin based on the three-dimensional spatial distribution, it notifies the control unit of the anomaly detection device 1 of this. The control unit activates the camera and instructs the data acquisition unit 11 to acquire a captured image after the detection of a moving object. When the data acquisition unit 11 acquires the captured image after the detection of a moving object, it outputs the image to the intrusion detection unit 133 as detection data. The intrusion detection unit 133 performs the above-mentioned confirmation based on the captured image after the detection of a moving object acquired from the data acquisition unit 11. Also, for example, in the anomaly detection device 1, when the intrusion detection unit 133 detects an intrusion into the vehicle cabin based on detection data based on sensor data acquired by a radio wave sensor, i.e., the three-dimensional spatial distribution, it notifies the control unit of this, and the control unit activates the camera in addition to the radio wave sensor, and the alarm content generation unit 142 acquires an image of the vehicle cabin from the camera and includes the image as supplemental information in the output control information. In this way, the anomaly detection device 1 may be able to add or change the sensor 2 from which sensor data is acquired during the period from when the operation starts to when the operation ends.This allows the anomaly detection device 1 to further improve the accuracy of detecting whether an intrusion has occurred in the vehicle interior or whether an object has been left inside the vehicle interior. Furthermore, the anomaly detection device 1 can provide information that an intrusion has occurred in the vehicle interior or that an object has been left inside the vehicle interior in a more easily understandable manner.

[0125] In addition, in the specific example given above in which the anomaly detection cycle may be changed depending on whether the abandonment detection mode or the intrusion detection mode is selected, the selection unit 131 may, for example, lengthen the cycle for acquiring sensor data in the intrusion detection mode compared to the abandonment detection mode. For example, if the common sensor 2 is a camera, the selection unit 131 may lengthen the cycle for acquiring sensor data in the abandonment detection mode compared to the intrusion detection mode. For example, when performing abandonment detection, the anomaly detection device 1 can detect whether an object is abandoned in the vehicle cabin from a post-motion detection captured image based on an image of the vehicle cabin captured at least once. Therefore, in the anomaly detection device 1, there is no problem if the selection unit 131 lengthens the camera's image capture cycle or the first cycle in the abandonment detection mode. On the other hand, for example, when performing intrusion detection, the anomaly detection device 1 should shorten the camera's image capture cycle or the second cycle so as not to miss any quick movement of an intruder, such as theft of luggage through a window of the vehicle 100. In this way, when the selection unit 131 changes the abnormality detection period depending on the selection of the abandonment detection mode or the intrusion detection mode, the selection unit 131 can determine which period to make longer or shorter, the abandonment detection mode or the intrusion detection mode, depending on the type of sensor 2.

[0126] Furthermore, in the first embodiment described above, the selection unit 131 selects either the abandonment detection mode or the intrusion detection mode. However, this is not limiting. The selection unit 131 may be capable of selecting the abandonment detection mode and the intrusion detection mode in parallel. That is, the selection unit 131 may be capable of selecting the execution of abandonment detection in the vehicle interior and the execution of intrusion detection into the vehicle interior in parallel. For example, the selection unit 131 can allocate the processing capacity of the abnormality detection unit 13 to a process for detecting whether an object has been abandoned in the vehicle interior and a process for detecting whether an intrusion has occurred into the vehicle interior. This allows the abnormality detection device 1 to perform both the detection of whether an object has been abandoned in the vehicle interior and the detection of whether an intrusion has occurred into the vehicle interior in parallel, rather than performing only one of the detection of whether an object has been abandoned in the vehicle interior and the detection of whether an intrusion has occurred into the vehicle interior. However, performance or detection constraints may arise compared to when the processes are performed individually. For example, when switching from the abandonment detection mode to the intrusion detection mode, the selection unit 131 can detect whether an object has been left in the vehicle cabin and whether an intrusion has occurred into the vehicle cabin in parallel before and after the mode switch, thereby enabling gradual switching between the abandonment detection mode and the intrusion detection mode. Furthermore, the abnormality detection device 1 can also handle cases where the conditions, such as the detection time, defined in laws and regulations or assessments for detecting whether an object has been left in the vehicle cabin and detecting whether an intrusion has occurred into the vehicle cabin overlap.

[0127] In the first embodiment described above, the anomaly detection device 1 is an on-board device mounted on the vehicle 100, but this is merely an example. For example, some or all of the data acquisition unit 11, data processing unit 12, anomaly detection unit 13, or detection result output unit 14 provided in the anomaly detection device 1 may be provided in a server (not shown), and the on-board device and the server may constitute an anomaly detection system.

[0128] Furthermore, in the above-described first embodiment, the target area is the interior of a regular passenger car, but this is merely an example. The target area may also be the interior of a commercial vehicle such as a bus. Furthermore, the target area is not limited to the interior of a vehicle, but may be, for example, the interior of a moving object other than a vehicle, such as the interior of an airplane or a train. Furthermore, the target area is not limited to the interior of a moving object, but may be, for example, the interior of a room whose entrance to the target area can be opened and closed.

[0129] 10A and 10B are diagrams illustrating an example of the hardware configuration of the anomaly detection device 1 according to the first embodiment. In the first embodiment, the functions of the data acquisition unit 11, the data processing unit 12, the anomaly detection unit 13, the detection result output unit 14, and a control unit (not shown) are realized by a processing circuit 1001. That is, the anomaly detection device 1 includes the processing circuit 1001 for performing control to detect whether an object has been left in the vehicle cabin or whether an object has entered the vehicle cabin. The processing circuit 1001 may be dedicated hardware as shown in FIG. 10A , or may be a processor 1004 that executes a program stored in a memory as shown in FIG. 10B .

[0130] If the processing circuit 1001 is dedicated hardware, the processing circuit 1001 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof.

[0131] When the processing circuit is a processor 1004, the functions of the data acquisition unit 11, data processing unit 12, anomaly detection unit 13, detection result output unit 14, and a control unit (not shown) are realized by software, firmware, or a combination of software and firmware. The software or firmware is written as a program and stored in memory 1005. The processor 1004 reads and executes the program stored in memory 1005 to execute the functions of the data acquisition unit 11, data processing unit 12, anomaly detection unit 13, detection result output unit 14, and a control unit (not shown). In other words, the anomaly detection device 1 includes memory 1005 for storing a program that, when executed by the processor 1004, results in the execution of steps ST1 to ST5 of FIG. 7 described above. It can also be said that the program stored in memory 1005 causes a computer to execute the procedures or methods of the data acquisition unit 11, data processing unit 12, anomaly detection unit 13, detection result output unit 14, and a control unit (not shown). Here, the memory 1005 may be, for example, a non-volatile or volatile semiconductor memory such as a RAM, a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), or an EEPROM (Electrically Erasable Programmable Read-Only Memory), a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, or a DVD (Digital Versatile Disc).

[0132] The functions of the data acquisition unit 11, data processing unit 12, anomaly detection unit 13, detection result output unit 14, and a control unit (not shown) may be partially implemented by dedicated hardware and partially implemented by software or firmware. For example, the functions of the data acquisition unit 11 and data processing unit 12 may be implemented by a processing circuit 1001 as dedicated hardware, while the functions of the anomaly detection unit 13, detection result output unit 14, and a control unit (not shown) may be implemented by a processor 1004 reading and executing a program stored in a memory 1005. The detection result storage unit 141 and a storage unit (not shown) may be configured, for example, with a memory. The anomaly detection device 1 also includes an input interface device 1002 and an output interface device 1003 that communicate with devices such as the sensor 2 or the output device 3 via wired or wireless communication.

[0133] As described above, according to the first embodiment, the anomaly detection device 1 is configured to include: a data acquisition unit 11 that acquires detection data regarding a moving object present in the target area, which is generated based on sensor data acquired by the sensor 2 when the sensor 2 detects an object present in the target area; a selection unit 131 that selects whether to perform abandonment detection in the target area or intrusion detection into the target area; an abandonment detection unit 132 that, if the selection unit 131 selects execution of abandonment detection in the target area, detects whether an object has been abandoned in the target area based on the detection data acquired by the data acquisition unit 11; and an intrusion detection unit 133 that, if the selection unit 131 selects execution of intrusion detection into the target area, detects whether an object has been abandoned in the target area based on the detection data acquired by the data acquisition unit 11. Therefore, the anomaly detection device 1 can simultaneously detect whether an intrusion into the target area has occurred and whether an infant or other person has been left behind in the target area while suppressing the execution of unnecessary processes. As a result, the anomaly detection device 1 can suppress power consumption due to the execution of unnecessary processes.

[0134] Any of the components of the embodiments may be modified or omitted.

[0135] The anomaly detection device according to the present disclosure can simultaneously detect whether or not an intrusion into a target area has occurred and whether or not an infant or other person has been left behind in the target area, while suppressing the execution of unnecessary processes.

[0136] 1 Anomaly detection device, 11 Data acquisition unit, 12 Data processing unit, 121 Moving object extraction unit, 122 Moving object analysis unit, 123 Three-dimensional spatial distribution generation unit, 13 Anomaly detection unit, 131 Selection unit, 132 Abandonment detection unit, 133 Intrusion detection unit, 14 Detection result output unit, 141 Detection result storage unit, 142 Notification content generation unit, 143 Output control unit, 100 Vehicle, 1001 Processing circuit, 1002 Input interface device, 1003 Output interface device, 1004 Processor, 1005 Memory.

Claims

1. a data acquisition unit that acquires detection data related to a moving object present in a target area, the detection data being generated based on sensor data acquired by a sensor detecting an object present in the target area; a selection unit that selects whether to perform abandonment detection in the target area or intrusion detection into the target area; an abandonment detection unit that detects whether or not an object is left in the target area based on the detection data acquired by the data acquisition unit when the selection unit selects to perform abandonment detection in the target area; an intrusion detection unit that detects whether or not an intrusion has occurred in the target area based on the detection data acquired by the data acquisition unit when the selection unit selects execution of intrusion detection into the target area; An abnormality detection device comprising:

2. The abandonment detection unit detects whether the moving object regarded as a person requiring assistance is present in the target area based on the detection data, thereby detecting whether the moving object is abandoned in the target area.

2. The anomaly detection device according to claim 1.

3. a data processing unit that detects the moving object present in the target area based on the sensor data and generates the detection data; The data acquisition unit acquires the detection data generated by the data processing unit.

3. The abnormality detection device according to claim 1 or 2.

4. The abandonment detection unit and the intrusion detection unit each detect whether an object is abandoned in the target area or whether an object is intruded into the target area, based on the detection data based on the sensor data acquired by the common sensor.

3. The abnormality detection device according to claim 1 or 2.

5. The selection unit controls an operation of the sensor in accordance with a selection of whether to perform abandonment detection in the target area or intrusion detection into the target area.

3. The abnormality detection device according to claim 1 or 2.

6. The abandonment detection unit detects the presence or absence of abandonment in the target area by using the detection data based on the sensor data acquired by the common sensor, as well as the detection data based on the sensor data acquired by the sensor other than the common sensor.

5. The abnormality detection device according to claim 4.

7. The selection unit selects execution of abandonment detection in the target area until a switching time has elapsed after a door used for entering and exiting the target area has been locked, and selects execution of intrusion detection into the target area if abandonment in the target area has not been detected even after the switching time has elapsed.

3. The abnormality detection device according to claim 1 or 2.

8. The selection unit selects execution of abandonment detection in the target area when a door used for entering and exiting the target area is locked, and selects execution of intrusion detection into the target area when the abandonment detection unit completes detection of the presence or absence of abandonment in the target area based on the selection by the selection unit.

3. The abnormality detection device according to claim 1 or 2.

9. the target area is a vehicle interior; The selection unit selects execution of abandoned vehicle detection in the target area when the owner of the vehicle is present in the vicinity of the vehicle, and selects execution of intrusion detection into the target area when the owner of the vehicle leaves the vicinity of the vehicle.

3. The abnormality detection device according to claim 1 or 2.

10. The selection unit is capable of selecting in parallel execution of abandonment detection in the target area and execution of intrusion detection into the target area.

3. The abnormality detection device according to claim 1 or 2.

11. When a door used for entering and exiting the target area is locked, the selection unit repeatedly selects whether to perform abandonment detection in the target area or intrusion detection into the target area until the door is unlocked.

3. The abnormality detection device according to claim 1 or 2.

12. The abandonment detection unit detects the presence or absence of abandonment in the target area in a first period, the intrusion detection unit detects the presence or absence of an intrusion into the target area in a second period; The first period and the second period have different lengths.

3. The abnormality detection device according to claim 1 or 2.

13. a detection result output unit that outputs output control information based on a detection result of whether or not the object is left in the target area by the abandonment detection unit, or the output control information based on a detection result of whether or not the object is intruded into the target area by the intrusion detection unit. The anomaly detection device according to claim 1 or 2, further comprising:

14. The output control information is information for outputting an alarm.

14. The anomaly detection device according to claim 13.

15. The sensor includes a radio wave sensor that radiates radio waves toward the target area and acquires reflected waves that are reflected by objects within the target area as the sensor data.

3. The abnormality detection device according to claim 1 or 2.

16. The sensor includes a camera that images the region of interest.

3. The abnormality detection device according to claim 1 or 2.

17. The target area is the interior of a vehicle.

3. The abnormality detection device according to claim 1 or 2.

18. The sensor from which the sensor data is obtained can be added or changed between the start and end of the operation.

3. The abnormality detection device according to claim 1 or 2.

19. the target area is a vehicle interior; The output control information is information for controlling the operation of the vehicle.

14. The anomaly detection device according to claim 13.

20. A data acquisition unit acquires detection data related to a moving object existing in a target area, the detection data being generated based on sensor data acquired by a sensor detecting an object existing in the target area; A selection unit selects whether to perform abandonment detection in the target area or intrusion detection into the target area; a step of detecting whether or not an object is left behind in the target area based on the detection data acquired by the data acquisition unit, when the selection unit selects to execute the abandonment detection in the target area; an intrusion detection unit, when the selection unit selects execution of intrusion detection into the target area, detecting the presence or absence of an intrusion into the target area based on the detection data acquired by the data acquisition unit; The anomaly detection method includes: