Object detecting device, object detecting method, and control program

Abstract

The present invention realizes an object detecting device with which the overall reliability of a detection result is obtained by appropriately taking into account mutually different characteristics of a plurality of sensors. A sensor module (1) comprises: a first reliability determining unit (15) that determines a first reliability of a detection result in accordance with a detection parameter of a first sensor (2); a second reliability determining unit (16) that determines a second reliability of the detection result in accordance with a detection parameter of a second sensor (3); and a first combining unit (17) that determines a first combined reliability obtained by combining the first reliability and the second reliability.

Description

Object detection device, object detection method, and control program

[0001] The present invention relates to an object detection device and an object detection method.

[0002] Patent Literature 1 discloses a vehicle driving control processing system. The driving control processing system includes a roadside unit having a plurality of sensors for detecting obstacles. The driving control processing system determines the reliability of obstacle information and corrects the braking distance of the vehicle based on the reliability.

[0003] Patent Literature 2 discloses an information processing device that determines the probability of an object existing around a moving body. The information processing device determines a first probability of an object existing around the moving body for each of a plurality of sensors with different characteristics. The information processing device determines a second probability of the object's existence based on the first probability of existence and unmeasured information that specifies the number of times position information could not be obtained.

[0004] Japanese Patent No. 7308900 Japanese Patent No. 6696697

[0005] When detecting an object using multiple sensors with different characteristics, the reliability of the detection results of each sensor varies depending on the object being detected. For example, optical sensors have difficulty detecting transparent objects such as glass. On the other hand, ultrasonic sensors can detect transparent objects, but have difficulty detecting objects that attenuate ultrasonic waves, such as rubber and foam.

[0006] The technology of Patent Document 1 does not determine the reliability of the detection results by taking into consideration the different characteristics of a plurality of sensors.

[0007] The first existence probability and second existence probability described in Patent Document 2 are merely numerical values ​​indicating the probability that an object exists, and are not numerical values ​​indicating the reliability of the detection result (the degree to which the detection result that an object exists / does not exist is reliable). For example, a state in which the existence probability is 1 is considered to be a state in which the detection result that an object exists is reliable, and a state in which the existence probability is 0 is considered to be a state in which the detection result that an object does not exist is reliable. The existence probability and reliability are not the same concept.

[0008] An object of one embodiment of the present invention is to provide an object detection device that obtains overall reliability of detection results by appropriately considering different characteristics of a plurality of sensors.

[0009] An object detection device according to one aspect of the present invention is configured to include a first detection unit that detects an object from the output of a first sensor for object detection; a second detection unit that has characteristics different from those of the first sensor and detects the object from the output of a second sensor for object detection; a first reliability determination unit that determines a first reliability of a detection result of the first detection unit according to detection parameters of the first sensor; a second reliability determination unit that determines a second reliability of a detection result of the second detection unit according to detection parameters of the second sensor; and a first synthesis unit that determines a first synthesized reliability by synthesizing at least the first reliability and the second reliability as the reliability of a detection result of the object detected by a combination of multiple sensors including the first sensor and the second sensor.

[0010] An object detection method according to one aspect of the present invention includes a first detection step of detecting an object from the output of a first sensor for object detection; a second detection step of detecting the object from the output of a second sensor for object detection, the second sensor having characteristics different from those of the first sensor; a first reliability determination step of determining a first reliability of the detection result in the first detection step according to detection parameters of the first sensor; a second reliability determination step of determining a second reliability of the detection result in the second detection step according to detection parameters of the second sensor; and a combination step of determining a first combined reliability by combining at least the first reliability and the second reliability as the reliability of the detection result of the object detected by a combination of multiple sensors including the first sensor and the second sensor.

[0011] According to one aspect of the present invention, it is possible to obtain the overall reliability of the detection results by appropriately considering the different characteristics of a plurality of sensors.

[0012] Fig. 1 is a block diagram showing the configuration of a sensor module of one embodiment of the present invention; Fig. 2 is a diagram showing the processing flow of a first control unit; Fig. 3 is a diagram showing weighting coefficients when N = 2; Fig. 4 is a diagram showing weighting coefficients when N = 3; Fig. 4 is a block diagram showing the configuration of an autonomous traveling transport robot of one embodiment of the present invention; Fig. 5 is a diagram showing the processing flow of a second control unit;

[0013] [First Embodiment] Hereinafter, an embodiment according to one aspect of the present invention (hereinafter also referred to as "the present embodiment") will be described with reference to the drawings.

[0014] §1 Application Examples Depending on the characteristics of the sensor, there are different objects that are easy to detect and objects that are difficult to detect. For example, no matter how many times a sensor tries to detect a transparent object that is difficult to detect with multiple optical sensors, there is a high possibility that the transparent object will not be detected in the end.

[0015] Therefore, it becomes necessary to detect objects using multiple sensors with different characteristics. However, in order to operate a machine based on the detection results, reliability information, which is an index that indicates how reliable the detection results from multiple sensors with different characteristics are, is required. The sensor module of this embodiment combines the reliability determined for each sensor according to the characteristics of each sensor, and determines a composite reliability, which is the reliability of the object detection result obtained by combining the multiple sensors.

[0016] §2 Configuration Example Figure 1 is a block diagram showing the configuration of a sensor module 1 according to this embodiment. The sensor module 1 is an object detection device that detects surrounding objects using multiple sensors and outputs the detection results and the reliability of the detection results. The sensor module 1 includes multiple sensors, namely, a first sensor 2 and a second sensor 3, and a first control unit 10. The first sensor 2 and the second sensor 3 are sensors for object detection and have different characteristics from each other.

[0017] Here, the first sensor 2 is an optical sensor that detects an object using TOF (Time of Flight). The first sensor 2 has a plurality of detection pixels and detects the distance to the object in a plurality of directions. The first sensor 2 outputs information indicating the distance to the object in a plurality of directions and information indicating the intensity of the detected signal (detected reflected light) in a plurality of directions. The information indicating the distance to the object in a plurality of directions can also be said to indicate, for example, in which direction and at what distance the object is located.

[0018] The second sensor 3 is an ultrasonic sensor that detects the distance to an object in multiple directions by emitting ultrasonic waves and detecting ultrasonic waves reflected by the object. The second sensor 3 outputs information indicating the distance to the object in multiple directions and information indicating the intensity of the detected signals (detected reflected waves) in multiple directions. The second sensor 3 has different characteristics from the first sensor 2, and the type (material) of objects that it is suitable for detecting is also different.

[0019] The first control unit 10 includes a first acquisition unit 11, a second acquisition unit 12, a first detection unit 13, a second detection unit 14, a first reliability determination unit 15, a second reliability determination unit 16, a first synthesis unit 17, and a first output unit 18.

[0020] The first acquisition unit 11 acquires the output of the first sensor 2. Specifically, the first acquisition unit 11 acquires information indicating the distance to an object in multiple directions and information indicating the strength of the detected signal in multiple directions as the output of the first sensor 2. The first acquisition unit 11 outputs the output of the first sensor 2 to the first detection unit 13.

[0021] The first detection unit 13 detects an object from the output of the first sensor 2. For example, the first detection unit 13 identifies the presence or absence of an object in each direction, the distance to the object, and the size of the object. The first detection unit 13 outputs information on the presence or absence of an object in the detection range, the direction (position) of the detected object, the distance to the object, and the size of the object as an object detection result to the first reliability determination unit 15. The first detection unit 13 also outputs the output of the first sensor 2 to the first reliability determination unit 15.

[0022] The first reliability determination unit 15 determines the first reliability of the detection result by the first detection unit 13 according to the detection parameters of the first sensor 2. The detection parameters of the first sensor 2 are, for example, the detection signal strength corresponding to the detected object, the time of the detection signal corresponding to the detected object, and / or the number of consecutive pixels corresponding to the detected object. The time of the detection signal corresponding to the detected object corresponds to the distance to the object. The number of pixels corresponding to the detected object corresponds to the size of the object. The first reliability determination unit 15 outputs the detection result by the first detection unit 13 and the first reliability to the first combination unit 17.

[0023] The second acquisition unit 12 acquires the output of the second sensor 3. Specifically, the second acquisition unit 12 acquires information indicating the distance to an object in multiple directions and information indicating the strength of the detected signal in multiple directions as the output of the second sensor 3. The second acquisition unit 12 outputs the output of the second sensor 3 to the second detection unit 14.

[0024] The second detection unit 14 detects an object from the output of the second sensor 3. For example, the second detection unit 14 identifies the presence or absence of an object in each direction, the distance to the object, and the size of the object. The second detection unit 14 outputs information on the presence or absence of an object in the detection range, the direction (position) of the detected object, the distance to the object, and the size of the object as an object detection result to the second reliability determination unit 16. The second detection unit 14 also outputs the output of the second sensor 3 to the second reliability determination unit 16.

[0025] The second reliability determination unit 16 determines the second reliability of the detection result by the second detection unit 14 in accordance with the detection parameters of the second sensor 3. The detection parameters of the second sensor 3 are, for example, the detection signal strength corresponding to the detected object or the time of the detection signal corresponding to the detected object. The second reliability determination unit 16 outputs the detection result by the second detection unit 14 and the second reliability to the first combination unit 17.

[0026] The first combiner 17 determines a first combined reliability by combining multiple reliabilities of the multiple sensors. The multiple reliabilities of the multiple sensors include at least a first reliability and a second reliability. The first combined reliability is the reliability of the object detection result obtained by combining the first sensor 2 and the second sensor 3. The first combiner 17 outputs information on the presence or absence of an object, the direction (position) of the detected object, the distance to the object, the size of the object, and the first combined reliability to the first output unit 18.

[0027] The first output unit 18 outputs information on the presence or absence of an object, the direction (position) of the detected object, the distance to the object, the size of the object, and the first composite reliability to an external device.

[0028] (Operation of First Control Unit 10) Fig. 2 is a diagram showing a processing flow of the first control unit 10. The first acquisition unit 11 acquires the output of the first sensor 2 (S1).

[0029] The first detection unit 13 detects an object from the output of the first sensor 2 (S2). The first detection unit 13 identifies the presence or absence of an object in each direction, the distance to the object, and the size of the object.

[0030] The first reliability determiner 15 determines the first reliability of the detection result by the first detector 13 according to the detection parameters of the first sensor 2 (S3). For example, if the strength of the detection signal by the first sensor 2 corresponding to the detected object is large, the detection result is likely to be reliable. The first reliability determiner 15 may set the first reliability higher as the strength of the detection signal corresponding to the detected object is large.

[0031] For example, the larger the detected object, the lower the possibility of erroneous detection. The first reliability determiner 15 may set the first reliability higher as the number of consecutive pixels corresponding to the detected object increases.

[0032] For example, it is considered that the detection accuracy is higher in a range closer to the first sensor 2. The first reliability determiner 15 may set the first reliability higher as the time of the detection signal corresponding to the detected object (i.e., the distance to the object) becomes shorter.

[0033] The second acquisition unit 12 acquires the output of the second sensor 3 (S4).

[0034] The second detection unit 14 detects an object from the output of the second sensor 3 (S5). The second detection unit 14 identifies the presence or absence of an object in each direction, the distance to the object, and the size of the object.

[0035] The second reliability determiner 16 determines the second reliability of the detection result by the second detector 14 according to the detection parameters of the second sensor 3 (S6). For example, the second reliability determiner 16 may set the second reliability higher as the detection signal strength corresponding to the detected object increases. Also, for example, the second reliability determiner 16 may set the second reliability higher as the time of the detection signal corresponding to the detected object (i.e., the distance to the object) decreases. The first reliability and the second reliability are multi-valued numerical values, not binary values, that represent the degree to which the detection result is reliable.

[0036] Consider a case where an object is detected in the same direction by the first detection unit 13 and the second detection unit 14. For objects detected in the same direction, the first synthesis unit 17 determines the shorter of the first distance to the object detected by the first detection unit 13 and the second distance to the object detected by the second detection unit 14 as the distance to the object.

[0037] The first combiner 17 determines a first combined reliability by combining the first reliability and the second reliability (S7). The first combiner 17 determines the weighted sum of the first reliability and the second reliability as the first combined reliability. The first combiner 17 changes the weighting coefficients depending on the detection result of the first detector 13 or the detection result of the second detector 14. The weighting coefficient of the first reliability is set to α1, and the weighting coefficient of the second reliability is set to α2. The first combined reliability is calculated as follows: First combined reliability = α1 × first reliability + α2 × second reliability, where α1 + α2 = 1.

[0038] For example, the first synthesis unit 17 changes the weighting coefficients α1 and α2 depending on the first distance to the object detected by the first detection unit 13 and the second distance to the object detected by the second detection unit 14. The first synthesis unit 17 may decrease the weighting coefficient α1 of the first reliability as the value obtained by subtracting the second distance from the first distance (first distance - second distance) increases. Alternatively, the first synthesis unit 17 may decrease the weighting coefficient α1 of the first reliability as the value obtained by dividing the first distance by the second distance (first distance / second distance) increases.

[0039] For example, the first synthesis unit 17 may determine the weighting coefficients α1 and α2 for the following cases: (1) when the first distance is greater than the second distance + A; (2) when the first distance is approximately equal to the second distance, i.e., when the second distance + A ≥ the first distance ≥ the second distance − B; and (3) when the second distance − B is greater than the first distance, where A and B ≥ 0. In other words, the first synthesis unit 17 changes the weighting coefficients α1 and α2 depending on whether the first distance is within a predetermined range including the second distance (from the second distance + A to the second distance − B), is greater than the predetermined range, or is smaller than the predetermined range.

[0040] (1) When the first distance is greater than the second distance + A, the detection result of the second sensor 3 is prioritized. Therefore, the first synthesis unit 17 increases the weighting coefficient α2 of the second reliability and decreases the weighting coefficient α1 of the first reliability compared to the case (2).

[0041] (2) When the second distance + A > the first distance > the second distance - B, the first synthesis unit 17 determines α1 and α2 according to the first distance - the second distance. For example, when the first distance = the second distance, α1 = α2 may be set.

[0042] (3) When the second distance - B>the first distance, the first synthesis unit 17 increases the weighting coefficient α1 of the first reliability and decreases the weighting coefficient α2 of the second reliability compared to the case (2).

[0043] As an example, if there are N sensors, the first combiner 17 may determine the weighting coefficients according to the following formula: where N≧2. αp: Weighting coefficient for the reliability of the prioritized sensor. αi: Weighting coefficient for the reliability of the other i-th sensor. fi(N): Adjustment function for the i-th sensor based on the number of sensors N. ai(): Function representing the base coefficient of the i-th sensor. bi(Di): Attenuation function that monotonically decreases based on the distance characteristics of the i-th sensor. Di: Detection distance by the i-th sensor. Dp: Detection distance by the prioritized sensor. ΔDi: |Dp-Di|.

[0044] The priority sensor is the sensor with the shortest detection distance. ai() is a function whose value decreases exponentially as Di and ΔDi increase, for example. In the above formula, the larger the ΔDi of the sensor, the smaller the weighting coefficient.

[0045] 3 is a diagram showing the weighting coefficient α2 when the priority sensor is the first sensor and N=2. The vertical axis represents the weighting coefficient value, and the horizontal axis represents ΔD. When ΔD=0, α1=α2=0.5. As ΔD increases, α2 decreases and approaches 0.

[0046] FIG. 4 is a diagram showing weighting coefficients α2 and α3 when the priority sensor is the first sensor and N=3. The vertical axis represents the weighting coefficient value. The horizontal axis represents ΔD. When ΔD=0, α1=0.5 and α2=α3=0.25. As ΔD increases, α2 and α3 decrease and approach 0. As shown here, the attenuation function of the weighting coefficients may differ for each sensor depending on the characteristics of the sensor. If ΔD is too large, for example, some of the weighting coefficients may be set to 0.

[0047] For example, if a transparent object such as glass is present, the ultrasonic second sensor 3 may detect the transparent object, while the optical first sensor 2 may not detect the transparent object but may instead detect another object behind the transparent object. Even if the first reliability of the first sensor 2 and the second reliability of the second sensor 3 are the same, the sensor module 1 determines the first combined reliability by assigning a larger weighting coefficient to the second reliability that detected the object closer to the first sensor. This allows for detection results that compensate for the strengths and weaknesses of multiple sensors with different characteristics and a first combined reliability that places greater emphasis on the reliability of the strong sensors. Therefore, the sensor module 1 can obtain the first combined reliability of the object detection results obtained by combining the multiple sensors, appropriately taking into account the different characteristics of the multiple sensors.

[0048] (Modification) The first synthesis unit 17 may decrease the weighting coefficient depending on the magnitude of the detected noise. For example, when the first sensor 2 detects noise equal to or greater than the first standard, the first synthesis unit 17 may decrease the weighting coefficient α1 of the first reliability compared to when the first sensor 2 detects noise below the first standard. However, this is under the condition that the detection result of the second sensor 3 is the same. Since α1 + α2 = 1, the first synthesis unit 17 increases the weighting coefficient α2 of the second reliability accordingly. For example, when the second sensor 3 detects noise equal to or greater than the second standard, the first synthesis unit 17 may decrease the weighting coefficient α2 of the second reliability compared to when the second sensor 3 detects noise below the second standard. When both noises are large, α1 = α2 may be used to strike a balance.

[0049] The first synthesis unit 17 may increase the weighting coefficient when the relative position of the detected object to the device (sensor module 1) changes over time compared to when it does not change. When the object's position is changing (approaching or moving away), it is more likely that a relatively moving object has been detected, rather than noise. For example, when the object is approaching, the detected distance monotonically decreases, and when the object is moving away, the detected distance monotonically increases. For example, when the relative position of the object detected by the first detection unit 13 is approaching over time, the first synthesis unit 17 may increase the weighting coefficient α1 of the first reliability compared to when the relative position of the object detected by the first detection unit 13 does not change. However, this is under the condition that the detection result of the second detection unit 14 remains the same.

[0050] The first synthesis unit 17 may reduce the weighting coefficient when there is a large variation in the position of an object detected over a recent predetermined period. When the variation is large, it is highly likely that the sensor is not good at detecting certain objects and that the object detection is not being performed accurately. For example, when the variation in the position of an object detected by the first detection unit 13 over a recent predetermined period is equal to or greater than a second standard, the weighting coefficient of the first reliability may be reduced compared to when the variation is less than the second standard. However, this is under the condition that the detection result of the second detection unit 14 is the same.

[0051] The above-mentioned weighting coefficient change also applies to the second sensor 3.

[0052] The optical first sensor 2 may make a false detection due to disturbance light (light in the surrounding environment). When the frequency at which the detection signal strength of the first sensor 2 is saturated (reaching its upper limit) in the most recent predetermined period is equal to or greater than a third criterion, the first synthesis unit 17 may reduce the weighting coefficient of the first reliability compared to when the frequency is less than the third criterion, provided that the detection result of the second detection unit 14 remains the same.

[0053] The above-described methods for determining the weighting coefficients may be used in combination. For example, the first combiner 17 may increase or decrease the weighting coefficients based on the detected distance, noise, a change in the distance over time, a variation in the distance, and / or a saturation of the detected signal strength.

[0054] The weighting coefficients may be fixed.

[0055] In addition, the first reliability determination unit 15 or the second reliability determination unit 16 may determine the first reliability or the second reliability based on the detection signal strength, the detected distance, and / or the number of pixels corresponding to the detected object.

[0056] The sensor module may include three or more sensors, determine the reliability of the detection result of each sensor, and determine a first combined reliability by combining the multiple reliabilities.

[0057] [Embodiment 2] Another embodiment of the present invention will be described below. For ease of explanation, the same reference numerals will be used to designate components having the same functions as those described in the above embodiment, and the description thereof will not be repeated.

[0058] 5 is a block diagram showing the configuration of the autonomous mobile transport robot 4 of this embodiment. The autonomous mobile transport robot 4 (AMR) is a transport device that detects surrounding objects and autonomously determines a travel route based on the detection results. The autonomous mobile transport robot 4 functions as an object detection device. The autonomous mobile transport robot 4 includes a sensor module 1, a third sensor 5, and a second control unit 20. The sensor module 1 is the same as that described in the first embodiment.

[0059] The third sensor 5 is a sensor for object detection and has characteristics different from those of the first sensor 2 and the second sensor 3. The third sensor 5 may be, for example, a two-dimensional imaging device or LiDAR. Here, the third sensor 5 is a LiDAR having multiple detection pixels and detecting distances to objects in multiple directions. The third sensor 5 may be a sensor that takes longer to process information for object detection (has a larger amount of information) than the first sensor 2 and the second sensor 3. Therefore, it is preferable that the object detection results by the third sensor 5 are more accurate than the object detection results by the first sensor 2 and the second sensor 3. The third sensor 5 outputs information indicating the distances to objects in multiple directions and information indicating the intensities of detected signals (detected reflected light) in multiple directions.

[0060] The second control unit 20 includes a fourth acquisition unit 21 , a third acquisition unit 22 , a third detection unit 24 , a third reliability determination unit 26 , a second synthesis unit 27 , an operation control unit 28 , and a second output unit 29 .

[0061] The fourth acquisition unit 21 acquires output from the sensor module 1. The fourth acquisition unit 21 acquires information on the presence or absence of an object, the direction (position) of the detected object, the distance to the object, the size of the object, and the first synthesis reliability. The fourth acquisition unit 21 outputs this information to the second synthesis unit 27.

[0062] The third acquisition unit 22 acquires the output of the third sensor 5. Specifically, the third acquisition unit 22 acquires information indicating the distance to an object in multiple directions and information indicating the strength of the detected signal in multiple directions, as the output of the third sensor 5. The third acquisition unit 22 outputs the output of the third sensor to the third detection unit 24.

[0063] The third detection unit 24 detects an object from the output of the third sensor 5. For example, the third detection unit 24 identifies the presence or absence of an object in each direction, the distance to the object, and the size of the object. The third detection unit 24 outputs information on the presence or absence of an object in the detection range, the direction (position) of the detected object, the distance to the object, and the size of the object as an object detection result to the third reliability determination unit 26. The third detection unit 24 also outputs the output of the third sensor 5 to the third reliability determination unit 26.

[0064] The third reliability determination unit 26 determines the third reliability of the detection result by the third detection unit 24 according to the detection parameters of the third sensor 5. The detection parameters of the third sensor 5 are, for example, the detection signal strength corresponding to the detected object, the detection signal time corresponding to the detected object, or the number of consecutive pixels corresponding to the detected object. The third reliability determination unit 26 outputs the detection result by the third detection unit 24 and the third reliability to the second combination unit 27.

[0065] The second combining unit 27 instructs the third acquiring unit 22 to acquire the output of the third sensor 5 according to the first combined reliability. The second combining unit 27 also determines a second combined reliability by combining the first combined reliability and the third reliability. The second combined reliability is the reliability of the object detection result obtained by combining the first sensor 2, the second sensor 3, and the third sensor 5. The second combining unit 27 outputs the first combined reliability or the second combined reliability to the operation control unit 28 as the overall reliability of the object detection result obtained by the multiple sensors (sensor module 1 and third sensor 5). The second combining unit 27 also outputs the presence or absence of an object, the direction (position) of the detected object, the distance to the object, and the size of the object to the operation control unit 28.

[0066] The operation control unit 28 controls the operation (travel) of the autonomous mobile transfer robot 4 based on the presence or absence of an object, the direction (position) of the detected object, the distance to the object, the size of the object, and the overall reliability. The operation control unit 28 outputs the overall reliability to the second output unit 29.

[0067] The second output unit 29 outputs a notification according to the overall reliability of the detected object to a higher-level management device or a user terminal.

[0068] (Operation of the second control unit 20) Fig. 6 is a diagram showing a processing flow of the second control unit 20. The fourth acquisition unit 21 acquires the output of the sensor module 1 (S11).

[0069] If the first combined reliability is equal to or greater than the first reliability threshold (Yes in S12), the second combining unit 27 sets the first combined reliability as the overall reliability (S13). In this case, the third acquiring unit 22 omits acquiring the output of the third sensor 5, and also omits the processing of the third detecting unit 24 and the third reliability determining unit 26. Note that the processing of the third acquiring unit 22, the third detecting unit 24, and the third reliability determining unit 26 may be performed every time without being omitted.

[0070] If the first combined reliability is less than the first reliability threshold (No in S12), the second combining unit 27 instructs the third acquiring unit 22 to acquire the output of the third sensor 5. The second combining unit 27 also instructs the operation control unit 28 to reduce the traveling speed. The operation control unit 28 controls the traveling drive unit of the autonomous traveling transfer robot 4 to reduce the traveling speed (S14). The third acquiring unit 22 acquires the output of the third sensor 5 (S15). The third detecting unit 24 detects an object from the output of the third sensor 5 (S16). The third detecting unit 24 detects a third distance to the object.

[0071] The third reliability determination unit 26 determines the third reliability of the detection result by the third detection unit 24 according to the detection parameters of the third sensor 5 (S17). The third reliability can be calculated in the same manner as the first reliability or the second reliability. The third reliability is a multi-valued numerical value, not a binary value, that indicates the degree to which the detection result is reliable.

[0072] After S17, the second synthesis unit 27 determines, for objects detected in the same direction, the shorter of the distance to the object measured by the sensor module 1 and the third distance to the object measured by the third detection unit 24 as the distance to the object.

[0073] The second combining unit 27 determines a second combined reliability by combining the first combined reliability and the third reliability (S18). The second combining unit 27 determines the second combined reliability as a weighted sum of the first combined reliability and the third reliability. The weighting coefficient can be determined using the method described in the first embodiment. The second combining unit 27 determines the second combined reliability as an overall reliability (S19).

[0074] After S13 or S19, the second synthesis unit 27 outputs the presence or absence of an object, the direction (position) of the detected object, the distance to the object, the size of the object, and the overall reliability to the operation control unit 28 (S20). The operation control unit 28 controls the operation of the autonomous mobile transfer robot 4 based on the overall reliability and the object detection results (direction, distance, size, etc.) (S21).

[0075] For example, if the overall reliability is equal to or greater than the second reliability threshold, the object detection result is reliable. In this case, the operation control unit 28 continues traveling at a normal traveling speed based on the object detection result.

[0076] If the overall reliability is less than the second reliability threshold and greater than or equal to the third reliability threshold, there is a slight possibility that the object detection result is incorrect. The second reliability threshold is greater than the third reliability threshold. In this case, the operation control unit 28 reduces the traveling speed to a slower speed than normal and continues traveling. While traveling at a slow speed, the autonomous mobile transport robot 4 can perform object detection again and reconfirm the safety of the route.

[0077] If the overall reliability is less than the third reliability threshold, the object detection result is likely to be erroneous. In this case, for example, the operation control unit 28 stops traveling and instructs the second output unit 29 to notify the outside of the overall reliability. The second output unit 29 outputs an error notification to a higher-level management device or a user terminal, indicating that the surrounding objects cannot be accurately detected. Upon receiving the error notification, the user can check the situation and directly issue operation instructions to the autonomous mobile transport robot 4. Note that each operation according to the overall reliability is an example, and a different operation may be performed based on the overall reliability.

[0078] In this way, the autonomous mobile transport robot 4 can operate based on the first composite reliability determined by the sensor module 1 and the third reliability determined from other sensors that the autonomous mobile transport robot 4 is equipped with. Therefore, the autonomous mobile transport robot 4 can integrate information from sensors for detecting more objects and make more appropriate decisions by taking into account the different characteristics of the multiple sensors.

[0079] Furthermore, when the first combined reliability is sufficiently high, the autonomous mobile transport robot 4 omits calculating the third distance, the third reliability, and the second combined reliability, thereby enabling object detection, determination of the overall reliability, and subsequent decision on an action to be taken more quickly. When the first combined reliability is low, the autonomous mobile transport robot 4 performs object detection using the third sensor 5 and calculates the third reliability and the second combined reliability. This allows the autonomous mobile transport robot 4 to use the third sensor 5, which can obtain more accurate object detection results, only when necessary.

[0080] The autonomously traveling transfer robot 4 may be provided with other sensors in addition to the third sensor 5, and the reliabilities determined for the other sensors may also be used in determining the second combined reliability.

[0081] (Modification) The first synthesis unit 17 and the second synthesis unit 27 may determine the synthesis reliability by using a learning model that receives a plurality of detection parameters from a plurality of sensors as input and outputs the synthesis reliability.

[0082] The autonomous transport robot may be equipped with multiple sensors including a third sensor in addition to the sensor module, determine the reliability of each sensor, and determine a second composite reliability by combining the first composite reliability and the reliability of the multiple sensors.

[0083] [Example of implementation using software] The functions of the sensor module 1 and the autonomous mobile transport robot 4 (hereinafter referred to as the "device") can be realized by a program that causes a computer to function as the device, and a program that causes a computer to function as each control block of the device (particularly each part included in the first control unit 10 and the second control unit 20).

[0084] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The functions described in each of the above embodiments are realized by executing the program using the control device and storage device.

[0085] The program may be non-transitory and may be recorded on one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

[0086] Furthermore, some or all of the functions of the control blocks can be realized by logic circuits. For example, an integrated circuit in which a logic circuit that functions as each of the control blocks is formed is also included in the scope of the present invention. In addition, the functions of the control blocks can also be realized by, for example, a quantum computer.

[0087] Furthermore, each process described in each of the above embodiments may be executed by AI (Artificial Intelligence). In this case, the AI ​​may run on the control device or on another device (for example, an edge computer or a cloud server).

[0088] [Summary] An object detection device according to aspect 1 of the present invention includes a first detection unit that detects an object from the output of a first sensor for object detection; a second detection unit that has characteristics different from those of the first sensor and detects the object from the output of a second sensor for object detection; a first reliability determination unit that determines a first reliability of a detection result of the first detection unit according to a detection parameter of the first sensor; a second reliability determination unit that determines a second reliability of a detection result of the second detection unit according to a detection parameter of the second sensor; and a first synthesis unit that determines a first synthesized reliability by synthesizing at least the first reliability and the second reliability as the reliability of a detection result of the object detected by a combination of multiple sensors including the first sensor and the second sensor.

[0089] According to the above configuration, the reliability determined for each of the plurality of sensors according to the characteristics of each sensor can be combined to obtain a first combined reliability of the object detection result obtained by combining the first and second sensors. Therefore, it is possible to obtain an overall reliability of the detection result by appropriately considering the different characteristics of the plurality of sensors.

[0090] The object detection device according to Aspect 2 of the present invention is based on Aspect 1 described above, and may be configured such that the first synthesis unit synthesizes at least the first reliability and the second reliability by a weighted sum.

[0091] An object detection device according to aspect 3 of the present invention may be configured in the above-described aspect 2 such that the first synthesis unit changes the weighting coefficient depending on the detection result of the first detection unit or the detection result of the second detection unit.

[0092] According to the above configuration, by weighting the detection results of each sensor, it is possible to obtain a first composite reliability that appropriately takes into account the different characteristics of the plurality of sensors.

[0093] An object detection device according to aspect 4 of the present invention may be configured such that, in the above-mentioned aspects 2 or 3, the first synthesis unit changes a weighting coefficient depending on a first distance to the object detected by the first detection unit and a second distance to the object detected by the second detection unit.

[0094] According to the above configuration, by placing more importance on the reliability of sensors that are good at detecting the target object, it is possible to obtain a first composite reliability that compensates for the detection results of sensors that are not good at detecting the target object. Therefore, it is possible to obtain a first composite reliability of the object detection results obtained by combining the multiple sensors, taking into appropriate consideration the different characteristics of the multiple sensors.

[0095] An object detection device according to aspect 5 of the present invention may be configured in the above-mentioned aspect 4 such that the first synthesis unit increases the weighting coefficient of the first reliability when the first distance is smaller than the predetermined range including the second distance compared to the weighting coefficient of the first reliability when the first distance is larger than the predetermined range including the second distance.

[0096] According to the above configuration, it is possible to place importance on the reliability of a sensor that detects an object at a shorter distance, that is, a sensor that detects an object at a closer distance.

[0097] An object detection device according to aspect 6 of the present invention may be configured in the above-mentioned aspect 4 such that the first synthesis unit reduces the weighting coefficient of the first reliability as the value obtained by subtracting the second distance from the first distance or the value obtained by dividing the first distance by the second distance increases.

[0098] According to the above configuration, the degree of emphasis on the reliability can be changed according to the difference or ratio of the distances detected by the multiple sensors, thereby obtaining a first composite reliability that more appropriately takes into account the characteristics of the multiple sensors.

[0099] An object detection device according to aspect 7 of the present invention may be configured such that, in any of aspects 2 to 6 above, the first synthesis unit reduces the weighting coefficient of the first reliability when noise equal to or greater than a first standard is detected by the first sensor compared to when noise less than the first standard is detected by the first sensor.

[0100] According to the above configuration, it is possible to obtain a first composite reliability that takes into account the influence of noise and places more importance on the reliability of sensors with less noise.

[0101] An object detection device according to aspect 8 of the present invention may be configured such that, in any of aspects 2 to 7 above, the first synthesis unit increases the weighting coefficient of the first reliability when the relative position of the object detected by the first detection unit changes over time compared to when the relative position of the object detected by the first detection unit does not change over time.

[0102] According to the above configuration, it is possible to obtain a first composite reliability that places more emphasis on the reliability of the sensor that is likely to accurately detect the object, based on the time-series changes in the position of the detected object.

[0103] An object detection device according to aspect 9 of the present invention may be configured such that, in any of aspects 2 to 8 above, the first synthesis unit reduces the weighting coefficient of the first reliability when the variation in the position of the object detected by the first detection unit during the most recent specified period is equal to or greater than a second standard, compared to when the variation in the position of the object detected by the first detection unit during the most recent specified period is less than the second standard.

[0104] According to the above configuration, it is possible to obtain a first composite reliability that places more importance on the reliability of the sensor that is likely to accurately detect the object, based on the variation in the position of the detected object.

[0105] An object detection device according to aspect 10 of the present invention may be configured in any one of aspects 1 to 9 above, including: a third detection unit that has characteristics different from those of the first sensor and the second sensor and detects the object from the output of a third sensor for object detection; a third reliability determination unit that determines a third reliability of the detection result of the third detection unit according to detection parameters of the third sensor; and a second synthesis unit that determines a second synthesized reliability by combining at least the first synthesized reliability and the third reliability as the reliability of the detection result of the object detected by a combination of multiple sensors including the first sensor, the second sensor, and the third sensor.

[0106] According to the above configuration, information from sensors for detecting more objects can be integrated, and different characteristics of the plurality of sensors can be taken into consideration to obtain a more appropriate second combined reliability.

[0107] An object detection device according to aspect 11 of the present invention may be configured such that, in the above-described aspect 10, the second synthesis unit outputs the first synthetic reliability as the reliability of the overall object detection result when the first synthetic reliability is equal to or greater than a threshold, and outputs the second synthetic reliability as the reliability of the overall object detection result when the first synthetic reliability is less than the threshold.

[0108] According to the above configuration, if the first composite reliability is sufficiently high, the first composite reliability based on the first and second sensors is used, and if the first composite reliability is low, the second composite reliability that also takes into account the detection result of the third sensor is used. This allows the reliability to be determined quickly, and the detection result of the third sensor can also be taken into account when necessary. Therefore, for example, the object detection device can omit detection processing using the third sensor when it is not necessary.

[0109] An object detection device according to aspect 12 of the present invention may be configured in any one of aspects 1 to 11 above, wherein the first reliability determination unit determines the first reliability according to the detection signal strength of the first sensor.

[0110] An object detection device according to aspect 13 of the present invention may be configured in any one of aspects 1 to 12 above, wherein the first sensor is an optical sensor having a plurality of pixels, and the first reliability determination unit determines the first reliability according to the number of pixels corresponding to the object detected by the first sensor.

[0111] An object detection method according to aspect 14 of the present invention includes a first detection step of detecting an object from the output of a first sensor for object detection; a second detection step of detecting the object from the output of a second sensor for object detection, the second sensor having characteristics different from those of the first sensor; a first reliability determination step of determining a first reliability of the detection result in the first detection step according to detection parameters of the first sensor; a second reliability determination step of determining a second reliability of the detection result in the second detection step according to detection parameters of the second sensor; and a combination step of determining a first combined reliability by combining at least the first reliability and the second reliability as the reliability of the detection result of the object detected by a combination of multiple sensors including the first sensor and the second sensor.

[0112] A control program according to aspect 15 of the present invention is a control program for causing a computer to execute the object detection method according to aspect 14 above, and may be configured to cause a computer to execute the first detection step, the second detection step, the first reliability determination step, the second reliability determination step, and the synthesis step.

[0113] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention.

[0114] REFERENCE SIGNS LIST 1 Sensor module (object detection device) 2 First sensor 3 Second sensor 4 Autonomous traveling transport robot (object detection device) 5 Third sensor 10 First control unit 11 First acquisition unit 12 Second acquisition unit 13 First detection unit 14 Second detection unit 15 First reliability determination unit 16 Second reliability determination unit 17 First synthesis unit 18 First output unit 20 Second control unit 21 Fourth acquisition unit 22 Third acquisition unit 24 Third detection unit 26 Third reliability determination unit 27 Second synthesis unit 28 Operation control unit 29 Second output unit

Claims

1. An object detection device comprising: a first detection unit that detects an object from the output of a first sensor for object detection; a second detection unit that has characteristics different from those of the first sensor and detects the object from the output of a second sensor for object detection; a first reliability determination unit that determines a first reliability of the detection result of the first detection unit according to detection parameters of the first sensor; a second reliability determination unit that determines a second reliability of the detection result of the second detection unit according to detection parameters of the second sensor; and a first combination unit that determines a first combined reliability by combining at least the first reliability and the second reliability as the reliability of the detection result of the object detected by a combination of multiple sensors including the first sensor and the second sensor.

2. The object detection device according to claim 1, wherein the first synthesis unit synthesizes at least the first reliability and the second reliability by a weighted sum.

3. The object detection device according to claim 2, wherein the first synthesis unit changes the weighting coefficient depending on the detection result of the first detection unit or the detection result of the second detection unit.

4. An object detection device as described in claim 2, wherein the first synthesis unit changes a weighting coefficient depending on a first distance to the object detected by the first detection unit and a second distance to the object detected by the second detection unit.

5. The object detection device described in claim 4, wherein the first synthesis unit increases the weighting coefficient of the first reliability when the first distance is smaller than the predetermined range including the second distance compared to the weighting coefficient of the first reliability when the first distance is larger than the predetermined range including the second distance.

6. The object detection device according to claim 4, wherein the first synthesis unit reduces the weighting coefficient of the first reliability as the value obtained by subtracting the second distance from the first distance or the value obtained by dividing the first distance by the second distance increases.

7. The object detection device of claim 2, wherein the first synthesis unit reduces the weighting coefficient of the first reliability when noise equal to or greater than a first standard is detected by the first sensor compared to when noise less than the first standard is detected by the first sensor.

8. An object detection device as described in claim 2, wherein the first synthesis unit increases the weighting coefficient of the first reliability when the relative position of the object detected by the first detection unit changes over time compared to when the relative position of the object detected by the first detection unit does not change over time.

9. The object detection device described in claim 2, wherein the first synthesis unit reduces the weighting coefficient of the first reliability when the variation in the position of the object detected by the first detection unit during the most recent specified period is equal to or greater than a second standard, compared to when the variation in the position of the object detected by the first detection unit during the most recent specified period is less than the second standard.

10. An object detection device as described in any one of claims 1 to 9, comprising: a third detection unit that has characteristics different from those of the first sensor and the second sensor and detects the object from the output of a third sensor for object detection; a third reliability determination unit that determines a third reliability of the detection result of the third detection unit according to detection parameters of the third sensor; and a second synthesis unit that determines a second synthesized reliability by synthesizing at least the first synthesized reliability and the third reliability as the reliability of the detection result of the object detected by a combination of multiple sensors including the first sensor, the second sensor, and the third sensor.

11. The object detection device described in claim 10, wherein the second synthesis unit outputs the first synthetic reliability as the overall reliability of the object detection result when the first synthetic reliability is equal to or greater than a threshold, and outputs the second synthetic reliability as the overall reliability of the object detection result when the first synthetic reliability is less than the threshold.

12. The object detection device according to claim 1, wherein the first reliability determination unit determines the first reliability according to the strength of the detection signal of the first sensor.

13. The object detection device according to claim 1, wherein the first sensor is an optical sensor having a plurality of pixels, and the first reliability determination unit determines the first reliability according to the number of pixels corresponding to the object detected by the first sensor.

14. An object detection method comprising: a first detection step of detecting an object from the output of a first sensor for object detection; a second detection step of detecting the object from the output of a second sensor for object detection, the second sensor having characteristics different from those of the first sensor; a first reliability determination step of determining a first reliability of the detection result in the first detection step according to detection parameters of the first sensor; a second reliability determination step of determining a second reliability of the detection result in the second detection step according to detection parameters of the second sensor; and a combination step of determining a first combined reliability by combining at least the first reliability and the second reliability as the reliability of the detection result of the object detected by a combination of multiple sensors including the first sensor and the second sensor.

15. A control program for causing a computer to execute the object detection method of claim 14, the control program causing the computer to execute the first detection step, the second detection step, the first reliability determination step, the second reliability determination step, and the synthesis step.

Images