Patent Grant
12123939
The present disclosure relates to an object tracking apparatus.
An object tracking apparatus that tracks an object is known. In the object tracking apparatus, an object may be tracked based on observation by a radar apparatus. For example.
a relative velocity of the object may be determined from phase rotation of frequency components that are continuously detected for the same object.
One aspect of the present disclosure is an object tracking apparatus that acquires detection information from a sensor that is mounted to a moving body, detects an object that is present in a vicinity of the moving body from the acquired detection information, and calculates a distance of the detected object, a relative velocity of the detected object, and an orientation of the detected object. For the object that is initially detected, the object tracking apparatus calculates a plurality of aliased velocities of which aliasing (folding) from the calculated relative velocity is assumed, and generates a plurality of target candidates that correspond to the plurality of aliased velocities. For each target candidate that is included in the generated plurality of target candidates, the object tracking apparatus estimates a current state of the target candidate from a past state of the target candidate and observation information that includes the calculated distance, the calculated relative velocity, and the calculated orientation. The object tracking apparatus selects the target candidate that is estimated to be a true target from the generated plurality of target candidates for the same object. The object tracking apparatus calculates, as a reference velocity, an aliased velocity of the calculated relative velocity that is equal to or greater than a velocity lower-limit value and is most negative or smallest in magnitude. The velocity lower-limit value is set based on a movement velocity of the moving body. The object tracking apparatus calculates the plurality of aliased velocities that are the calculated reference velocity that is aliased (folded) 0 to n times, n being a natural number, in a positive direction.
In the accompanying drawings:
The following embodiments relate to an object tracking apparatus that tracks an object.
When an object is tracked based on observation by a radar apparatus, a relative velocity of the observed object may have ambiguity. For example, when a method in which the relative velocity may be determined from phase rotation of frequency components that are continuously detected regarding the same object is used, an actual phase may be φ+2π×m (m being an integer) in relation to a detected phase φ. The relative velocity cannot be identified.
In K. L I et al, ‘Multitarget Tracking with Doppler Ambiguity,’ IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS VOL. 49, NO. 4 Oct. 2013, a technology in which a plurality of targets are generated by a plurality of relative velocities being assumed, and a true relative velocity is identified through tracking of the plurality of generated targets is proposed. Specifically, a likelihood is calculated for each of the plurality of assumptions, and the assumption of which the likelihood is higher than that of others based on tracking is identified as the true relative velocity.
In the technology for identifying the true relative velocity described above, estimation accuracy regarding the true relative velocity increases with increase in the number of assumptions of the relative velocity. However, processing load may increase in correspondence to the increase in the number of assumptions. However, as a result of detailed examination by the inventors, an issue has been found in that, when the number of assumptions is decreased, the relative velocities of the number of assumptions may not include the true relative velocity. Furthermore, an issue has been found in that, when the number of assumptions is decreased, tracking of the object may not be able to be performed, and lack of recognition of the object may occur.
It is thus desired to preferably enable suppression of lack of recognition of an object while suppressing processing load.
A first exemplary embodiment of the present disclosure provides an object tracking apparatus that includes an acquiring unit, an object detecting unit, a distance calculating unit, a velocity calculating unit, an orientation calculating unit, a candidate generating unit, a state estimating unit, and a candidate selecting unit. The acquiring unit is configured to acquire detection information from a sensor that is mounted to a moving body. The object detecting unit is configured to detect an object that is present in a vicinity of the moving body from the detection information that is acquired by the acquiring unit. The distance calculating unit is configured to calculate a distance of the object that is detected by the object detecting unit. The velocity calculating unit is configured to calculate a relative velocity of the object that is detected by the object detecting unit. The orientation calculating unit is configured to calculate an orientation of the object that is detected by the object detecting unit.
For the object that is initially detected by the object detecting unit, the candidate generating unit is configured to calculate a plurality of aliased velocities of which aliasing (folding) from the relative velocity that is calculated by the velocity calculating unit is assumed, and is configured to generate a plurality of target candidates that correspond to the plurality of aliased velocities.
For each target candidate that is included in the plurality of target candidates that are generated by the candidate generating unit, the state estimating unit is configured to estimate a current state of the target candidate from a past state of the target candidate and observation information. The observation information includes the distance that is calculated by the distance calculating unit, the relative velocity that is calculated by the velocity calculating unit, and the orientation that is calculated by the orientation calculating unit. The candidate selecting unit is configured to select the target candidate that is estimated to be a true target from the plurality of target candidates that are generated for the same object by the candidate generating unit.
The candidate generating unit includes a reference velocity calculating unit and an aliased velocity calculating unit. The reference velocity calculating unit is configured to calculate, as a reference velocity, an aliased velocity of the relative velocity that is calculated by the velocity calculating unit that is equal to or greater than a velocity lower-limit value and is most negative or smallest in magnitude. The velocity lower-limit value is set based on a movement velocity of the moving body. The aliased velocity calculating unit is configured to calculate the plurality of aliased velocities that are the reference velocity that is calculated by the reference velocity calculating unit that is aliased (folded) 0 to n times, n being a natural number, in a positive direction.
The relative velocity of an object that is observed by a sensor changes based on the velocity of the object and the velocity of the moving body. An object that requires attention of the moving body is an object that is present in an advancing direction of the moving body. The relative velocity of the object that is present in the advancing direction of the moving body has a higher likelihood of increasing towards a negative side as the velocity of the moving body increases. That is, the relative velocity to be assumed changes based on the velocity of the moving body. In addition, an object that has a high risk of colliding with the moving body is required to be preferentially recognized. Therefore, when the number of velocity aliasings (foldings) is assumed, a relative velocity that is greatest towards the negative side among the relative velocities that may be observed is required to be included.
A second exemplary embodiment of the present disclosure provides an object tracking apparatus that includes a processor, a non-transitory computer-readable storage medium, and a set of computer-executable instructions stored on the computer-readable storage medium. When read and executed by the processor, the set of computer-executable instructions cause the processor to implement: acquiring detection information from a sensor that is mounted to a moving body; detecting an object that is present in a vicinity of the moving body from the acquired detection information; calculating a distance of the detected object, a relative velocity of the detected object, and an orientation of the detected object; calculating, for the object that is initially detected by the object detecting unit, a plurality of aliased velocities of which aliasing from the calculated relative velocity is assumed, and generating a plurality of target candidates that correspond to the plurality of aliased velocities; estimating, for each target candidate that is included in the generated plurality of target candidates, a current state of the target candidate from a past state of the target candidate and observation information that includes the calculated distance, the calculated relative velocity, and the calculated orientation; selecting the target candidate that is estimated to be a true target from the generated plurality of target candidates for a same object; calculating, as a reference velocity, an aliased velocity of the calculated relative velocity that is equal to or greater than a velocity lower-limit value and is most negative or smallest in magnitude, the velocity lower-limit value being set based on a movement velocity of the moving body; and calculating the plurality of aliased velocities that are the calculated reference velocity that is aliased 0 to n times, n being a natural number, in a positive direction.
A third exemplary embodiment of the present disclosure provides an object tracking method that includes: acquiring detection information from a sensor that is mounted to a moving body; detecting an object that is present in a vicinity of the moving body from the acquired detection information; calculating a distance of the detected object, a relative velocity of the detected object, and an orientation of the detected object; calculating, for the object that is initially detected by the object detecting unit, a plurality of aliased velocities of which aliasing from the calculated relative velocity is assumed, and generating a plurality of target candidates that correspond to the plurality of aliased velocities; estimating, for each target candidate that is included in the generated plurality of target candidates, a current state of the target candidate from a past state of the target candidate and observation information that includes the calculated distance, the calculated relative velocity, and the calculated orientation; selecting the target candidate that is estimated to be a true target from the generated plurality of target candidates for a same object; calculating, as a reference velocity, an aliased velocity of the calculated relative velocity that is equal to or greater than a velocity lower-limit value and is most negative or smallest in magnitude, the velocity lower-limit value being set based on a movement velocity of the moving body; and calculating the plurality of aliased velocities that are the calculated reference velocity that is aliased 0 to n times, n being a natural number, in a positive direction.
Therefore, in the object tracking apparatus according to the first to third exemplary embodiments of the present disclosure, an aliased velocity that is equal to or less than the velocity lower-limit value that is set based on the velocity of the moving body and is most negative or smallest in magnitude is calculated as the reference velocity. Moreover, the plurality of aliased velocities that are the reference velocity that is aliased in the positive direction are calculated, and the plurality of target candidates that correspond to the plurality of aliased velocities are generated. As a result, the plurality of aliased velocities over an appropriate range based on the velocity of the vehicle can be calculated. Consequently, estimation accuracy regarding a true relative velocity can be improved while processing load is suppressed. Furthermore, lack of recognition of an object can be suppressed while processing load is suppressed.
An embodiment for implementing the present disclosure will hereinafter be described with reference to the drawings.
<1. Configuration>
First, a configuration of a driving assistance system 100 according to the present embodiment will be described with reference to
The radar apparatus 10 may be mounted in a front center (such as a center of a front bumper) of a vehicle 80, and have an area that is centered and in front of the vehicle 80 as a detection area. In addition, the radar apparatus 10 may be mounted in each of a front left side and a front right side (such as a left end and a right end of the front bumper) of the vehicle 80, and have areas on the front left side and the front right side of the vehicle 80 as the detection areas.
Furthermore, the radar apparatus 10 may be mounted in each of a rear left side and a rear right side (such as a left end and a right end of a rear bumper) of the vehicle 80, and have areas on the rear left side and the rear right side of the vehicle 80 as the detection areas. Not all five of these radar apparatuses 10 need be mounted to the vehicle 80. Only one of the five radar apparatuses 10 may be mounted to the vehicle 80. Alternatively, two or more of the five radar apparatuses 10 may be mounted to the vehicle 80. Here, the radar apparatus 10 corresponds to a sensor of the present disclosure. The vehicle 80 corresponds to a moving body of the present disclosure.
The radar apparatus 10 is a millimeter-wave radar. The radar apparatus 10 includes a transmission array antenna that includes a plurality of antenna elements and a reception array antenna that includes a plurality of antenna elements. The radar apparatus 10 repeatedly transmits transmission waves at a predetermined cycle and received reflected waves that are generated by the transmission waves being reflected by an object. Furthermore, the radar apparatus 10 combines the transmission wave and the reflected wave and generates a beat signal, and outputs sampled beat signals (that is, detection information) to the object tracking apparatus 20. The radar apparatus 10 is a modulation-type radar in which observation velocity has ambiguity, such as a Fast Chirp Modulation (FCM)-type radar.
The object tracking apparatus 20 includes a microcomputer that has a central processing unit (CPU) and a semiconductor memory such as a read-only memory (ROM) or a random access memory (RAM). The object tracking apparatus 20 actualizes various functions by the CPU running various programs that are stored in the ROM. Specifically, as shown in
The driving assistance apparatus 50 controls the vehicle 80 and actualizes driving assistance using the target information that is generated by the object tracking apparatus 20, and state information and behavior information of the vehicle 80 that are acquired from various sensors that are mounted to the vehicle 80.
<2. Processes>
Next, the object tracking process performed by the object tracking apparatus 20 according to the first embodiment will be described with reference to a flowchart in
First, at S10, the object detecting unit 22 detects at least one object that is present in a vicinity of the vehicle 80 from the detection information that is acquired by the acquiring unit 21. Furthermore, for each detected object, the distance calculating unit 23 calculates a distance of the detected object, and the velocity calculating unit 24 calculates a relative velocity Vobs of the detected object. In addition, for each detected object, the orientation calculating unit 25 calculates an orientation of the detected object. The distance of the object is a distance from the vehicle 80 to the object and the relative velocity Vobs of the object is a velocity of the object in relation to the vehicle 80. In addition, the orientation of the object is an orientation of the object in relation to the vehicle 80. Here, as shown in
Next, at S20, the state estimating unit 29 determines whether unprocessed target information is present. Specifically, the state estimating unit 29 determines whether a target candidate for which subsequent processes at S30 to S60 that follow have not been performed is present among registered target candidates. When determined that an unprocessed target candidate is present at S20, the state estimating unit 29 proceeds to S30.
At S30, for a single target candidate among the unprocessed target candidates, the state estimating unit 29 estimates a current state of the target candidate from a past state of the target candidate and observation values acquired at S10. Specifically, the state estimating unit 29 calculates a prediction value of a state quantity of the target candidate in a present processing cycle from an estimation value (that is, the past state) of the state quantity of the target candidate that is calculated in a previous processing cycle. In a manner similar to the observation values, the state quantity of the target may have the distance, the relative velocity, and the orientation of the target candidate as elements, or may have an X-axis coordinate value, a Y-axis coordinate value, an X-direction relative velocity, and a Y-direction relative velocity as the elements. The X-axis is an axis that runs along a width direction of the vehicle 80. The Y-axis is an axis that is orthogonal to the X-axis and runs along a length direction of the vehicle 80. In addition, the state quantity of the target may be elements other than the foregoing.
Furthermore, the state estimating unit 29 determines the observation value that is associated with the calculated prediction value from the observation values acquired at S10. Specifically, the state estimating unit 29 sets a predetermined range that is centered on the calculated prediction value, and determines an observation value that is within the predetermined range and closest to the prediction value as the observation value to be associated with the prediction value. The predetermined range is the range of observation values that are estimated to be acquired from the same object as the prediction value.
Then, the state estimating unit 29 calculates an estimation value (that is, a current state) in the present processing cycle using a Kalman filter or the like, from the calculated prediction value and the observation value that is determined to be associated.
In addition, the state estimating unit 29 calculates a likelihood of the calculated estimation value. When the estimation value is calculated from the prediction value and the observation value that is determined to be associated with the prediction value, the likelihood of the estimation value increases. At this time, the likelihood of the estimation value may increase as a difference between the prediction value and the observation value decreases. Moreover, when the observation value that is determined to be associated is not present and the estimation value is calculated from only the prediction value, the likelihood of the estimation value decreases.
Next, at S40, the target selecting unit 30 determines whether processing of all target candidates that are generated from the same object is completed. That is, the target selecting unit 30 determines whether the estimation value of the state quantity is calculated for all target candidates that are generated from the same object. As shown in
As shown in
Next, at S40, when determined that processing of all object candidates that are generated from the same object is not completed, the target selecting unit 30 returns to the process at S20. When determined that processing of all object candidates is completed, the target selecting unit 30 proceeds to S50.
At S50, the target selecting unit 30 determines whether any of the target candidates that are generated from the same object satisfies a selection condition. The selection condition is a condition for determining that the target candidate is a true target and, for example, may be the likelihood of the estimation value being equal to or greater than a selection threshold. When none of the target candidates satisfies the selection condition, the target selecting unit 30 returns to the process at S20. When any of the target candidates satisfies the selection condition, the target selecting unit 30 proceeds to S60.
As shown in
In addition, when determined that an unprocessed target candidate is not present at S20, the target selecting unit 30 proceeds to step S70.
At S70, the target candidate generating unit 26 determines whether an unused observation value is present among the observation values acquired at S10. That is, the target candidate generating unit 26 determines whether an observation value that is not associated with any of the prediction values is present among the observation values detected at S10. The unused observation value corresponds to an observation value of an object that is initially detected in the current processing cycle. When determined that an unused observation value is not present at S70, the target candidate generating unit 26 ends the present process. Meanwhile, when determined that the unused observation value is present at S70, the target candidate generating unit 26 proceeds to S80.
Next, at S80, a target candidate is generated from one of the observation values among the unused observation values. That is, a plurality of target candidates of which a plurality of aliased velocities are assumed are generated for the object that is initially detected in the current processing cycle. Specifically, a flowchart shown in
First, at S100, the reference velocity calculating unit 27 calculates a reference velocity for generating the plurality of target candidates. Specifically, as shown in
Here, at least an object that is approaching the vehicle 80 is preferably recognizable. Therefore, as shown in
Furthermore, a stationary object is preferably recognizable. The relative velocity of a stationary object is a negative value of the vehicle velocity of the vehicle 80. Here, as shown in
In addition, among the objects that are present in the vicinity of the vehicle 80, in particular, an object that may collide with the vehicle 80 is preferably recognizable. Here, as shown in
For example, as shown in
In addition, as shown in
Furthermore, as shown in
In addition, a plurality of corresponding relationships between the vehicle velocity of the vehicle 80 and the velocity lower-limit value may be provided. In this case, the reference velocity calculating unit 27 may calculate the reference velocity using the corresponding relationship that is appropriate based on a traveling state of the vehicle 80. For example, the traveling state may be a state in which the vehicle 80 is traveling through an urban area, a state in which the vehicle 80 is traveling on an expressway, or the like.
Next, at S110, the aliased velocity calculating unit 28 calculates (n+1) aliased velocities that are the reference velocity calculated by the reference velocity calculating unit 27 aliased (folded) 0 to n times, n being a natural number, in a positive direction. According to the present embodiment, as shown in
At this time, the aliased velocity calculating unit 28 excludes the target candidate that corresponds to the aliased velocity that is equal to or greater than a velocity upper-limit value set in advance, from the generated three target candidates. The velocity upper-limit value is set to a positive value and is set to eliminate a target that has an unlikely relative velocity.
For example, as shown in
<3. Effects>
According to the present embodiment described above, following effects are achieved.
An embodiment for carrying out the present disclosure is described above. However, the present disclosure is not limited to the above-described embodiment and can be modified in various ways.
The object tracking apparatus 20 and the method thereof described in the present disclosure may be actualized by a dedicated computer that is provided so as to be configured by a processor and a memory, the processor being programmed to provide one or a plurality of functions that are realized by a computer program. Alternatively, the object tracking apparatus 20 and the method thereof described in the present disclosure may be actualized by a dedicated computer that is provided by a processor being configured by a single dedicated hardware logic circuit or more. As another alternative, the object tracking apparatus 20 and the method thereof described in the present disclosure may be actualized by a single dedicated computer or more. The dedicated computer may be configured by a combination of a processor that is programmed to provide one or a plurality of functions, a memory, and a processor that is configured by a single hardware logic circuit or more. In addition, the computer program may be stored in a non-transitory computer-readable storage medium that can be read by a computer as instructions performed by the computer. A method for actualizing functions of sections that are included in the object tracking apparatus 20 is not necessarily required to include software. All of the functions may be actualized using a single or a plurality of pieces of hardware.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2019-188668 | Oct 2019 | JP | national |
The present application is a continuation application of International Application No. PCT/JP2020/038021, filed on Oct. 7, 2020, which claims priority to Japanese Patent Application No. 2019-188668, filed on Oct. 15, 2019. The contents of these applications are incorporated herein by reference in their entirety.
| Number | Name | Date | Kind |
|---|---|---|---|
| 5757308 | Lissel | May 1998 | A |
| 8686894 | Fukuda | Apr 2014 | B2 |
| 10048366 | Hong | Aug 2018 | B1 |
| 10054680 | Owirka | Aug 2018 | B2 |
| 10062181 | Longhurst | Aug 2018 | B1 |
| 10634766 | Stachnik | Apr 2020 | B2 |
| 11209284 | Hwang | Dec 2021 | B2 |
| 11262448 | Davis | Mar 2022 | B2 |
| 11448746 | Stachnik | Sep 2022 | B2 |
| 11941985 | Sakai | Mar 2024 | B2 |
| 20190120953 | Takayama | Apr 2019 | A1 |
| 20190315351 | Smith | Oct 2019 | A1 |
| 20190317219 | Smith | Oct 2019 | A1 |
| 20190318206 | Smith | Oct 2019 | A1 |
| 20200057141 | Kim | Feb 2020 | A1 |
| 20210003695 | Akamine | Jan 2021 | A1 |
| 20210063561 | Yamada | Mar 2021 | A1 |
| 20210146932 | Smith | May 2021 | A1 |
| 20230133611 | Smith | May 2023 | A1 |
| 20230271615 | Smith | Aug 2023 | A1 |
| 20240115810 | Schabbach | Apr 2024 | A1 |
| Number | Date | Country |
|---|---|---|
| 1372000 | Dec 2003 | EP |
| 2000065921 | Mar 2000 | JP |
| 2017090143 | May 2017 | JP |
| Entry |
|---|
| Li, et al, “Multitarget Tracking with Doppler Ambiguity”, IEEE Transactions on Aerospace and Electronic Systems, vol. 49, No. 4, Oct. 2013, pp. 2640-2656. |
| Number | Date | Country | |
|---|---|---|---|
| 20220236398 A1 | Jul 2022 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/JP2020/038021 | Oct 2020 | WO |
| Child | 17659154 | US |
