Tracking device, tracking method, and tracking program

Abstract

To provide a tracking device enabling improvement of tracking accuracy for a mobile object.SOLUTION: A processor of a tracking device is configured to perform: obtaining an observed value of a mobile object observed at an observation time; obtaining an estimated state value by estimating a state value of the mobile object at the observation time; and estimating a real value of the state value at the observation time by means of nonlinear filtering using the observation value and the estimated state value at the observation time as variables. Estimating the real value includes: setting, according to a visibility degree for each vertex m from an external sensor system, a weighting coefficient s for each of a plurality of vertices m in a rectangular model in which the mobile object is modeled; obtaining an observation error at each vertex m, based on the observation value and the estimated state value at the observation time; and obtaining a covariance of the observation error, based on the weighting coefficient s for each vertex m.SELECTED DRAWING: Figure 7

Description

[Technical field]

[0001] The present disclosure relates to a tracking technology for tracking a moving object. [Background technology]

[0002] A tracking technique for tracking a moving object by estimating the state value of the moving object in a time series based on the observation value by an external sensor system is widely known. As one type of tracking technique, Non-Patent Document 1 proposes a method of repeating the estimation of the state value of the moving object in a time series by filtering using a Kalman filter. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] 3D Multi-Object Tracking: A Baseline and New Evaluation Metrics(URL:https: / / arxiv.org / pdf / 1907.03961.pdf) Summary of the Invention [Problem to be solved by the invention]

[0004] However, the method proposed in Non-Patent Document 1 is premised on the assumption that the entire moving object is sufficiently observed from the external sensor system. Therefore, when it is difficult to observe a part of the moving object from the external sensor system, the estimated state value may deviate from the true value, and the tracking accuracy may decrease.

[0005] An object of the present disclosure is to provide a tracking device that improves tracking accuracy for a moving object. Another object of the present disclosure is to provide a tracking method that improves tracking accuracy for a moving object. Yet another object of the present disclosure is to provide a tracking program that improves tracking accuracy for a moving object. [Means for solving the problem]

[0006] The technical means of the present disclosure for solving the problems will be described below. Note that the claims and the reference characters in parentheses in this section indicate the corresponding relationship with the specific means described in the embodiments described later in detail, and do not limit the technical scope of the present disclosure.

[0007] A first aspect of the present disclosure is A tracking device having a processor (12) for tracking a moving object by estimating state values ​​of the moving object (3) in a time series based on observed values ​​by an external sensor system (2), The processor Obtaining an observation value of a moving object observed at an observation time; Obtaining a predicted state value by predicting a state value of the moving object at an observation time; and estimating a true value of the state value at the observation time by nonlinear filtering using the observation value at the observation time and the predicted state value as variables. To estimate the true value, setting weighting coefficients for each of a plurality of vertices in a rectangular model that models a moving object according to the visibility of each of the vertices from an external sensor system; Obtaining an observation error at each vertex based on an observation value at an observation time and a predicted state value; and obtaining a covariance of the observation errors based on the weighting coefficients for each vertex.

[0008] A second aspect of the present disclosure is A tracking method for tracking a moving object by estimating state values ​​of the moving object in a time series manner based on observation values ​​of an external sensor system, the method comprising: Obtaining an observation value of a moving object observed at an observation time; Obtaining a predicted state value by predicting a state value of the moving object at an observation time; Estimating a true value of the state value at the observation time by nonlinear filtering using the observed value at the observation time and the predicted state value as variables, To estimate the true value, setting weighting coefficients for each of a plurality of vertices in a rectangular model that models a moving object according to the visibility of each of the vertices from an external sensor system; Obtaining an observation error at each vertex based on an observation value at an observation time and a predicted state value; and obtaining a covariance of the observation errors based on the weighting coefficients for each vertex.

[0009] A third aspect of the present disclosure is A tracking program stored in a storage medium (10), the tracking program including instructions to be executed by a processor (12) for tracking a moving object by estimating state values ​​of the moving object (3) in a time series based on observed values ​​by an external sensor system (2), The command is, Obtaining an observation value of a moving object observed at an observation time; Obtaining a predicted state value by predicting a state value of the moving object at an observation time; and estimating a true value of the state value at the observation time by nonlinear filtering using the observed value at the observation time and the predicted state value as variables, To estimate the true value, setting weight coefficients for each of a plurality of vertices in a rectangular model that models a moving object according to the visibility of each of the vertices from an external sensor system; Obtaining an observation error at each vertex based on an observation value at an observation time and a predicted state value; Obtaining a covariance of the observation errors based on a weighting coefficient for each vertex.

[0010] According to these first to third aspects, the true value of the state value at the observation time is estimated by nonlinear filtering using the observed value at the observation time and the predicted state value as variables. At this time, the observation error at each vertex of a rectangular model that models a moving body is obtained based on the observed value at the observation time and the predicted state value, and the covariance of the observation error is obtained based on the weighting coefficient for each vertex. Therefore, according to the first to third aspects in which a weighting coefficient is set for each vertex according to the visibility from an external sensor system, the visibility can be reflected in the estimation of the true value of the state value. Therefore, it is possible to accurately estimate the true value of the state value and improve the tracking accuracy of the moving body. [Brief description of the drawings]

[0011] [Figure 1] 1 is a block diagram showing an overall configuration of a tracking device according to an embodiment; [Diagram 2] FIG. 1 is a schematic diagram for explaining observed values ​​and a rectangular model according to an embodiment. [Diagram 3] FIG. 2 is a block diagram showing a functional configuration of a tracking device according to an embodiment. [Figure 4] 1 is a flow chart illustrating a tracking method according to one embodiment. [Diagram 5] FIG. 1 is a schematic diagram illustrating a predicted state value and a rectangle model according to an embodiment. [Figure 6] 1 is a flowchart illustrating an estimation process according to an embodiment. [Figure 7] FIG. 13 is a schematic diagram illustrating an example of weight setting according to an embodiment. [Figure 8] 11 is a flowchart illustrating a weight setting subroutine according to one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment will be described with reference to the drawings.

[0013] 1, a tracking device 1 according to an embodiment tracks a moving object 3 by estimating state values ​​of the moving object 3 in time series based on observation values ​​by an external sensor system 2. To this end, the tracking device 1 is mounted on a vehicle 4 together with the external sensor system 2.

[0014] The vehicle 4 is provided with an autonomous driving mode temporarily by switching between a manual driving mode and a manual driving mode, or steadily without the switching being substantially performed. The autonomous driving mode may be realized by autonomous driving control, such as conditional driving automation, highly automated driving, or fully automated driving, in which a system performs all driving tasks when activated. The autonomous driving mode may also be realized by advanced driving assistance control, such as driving assistance or partial driving automation, in which an occupant performs some or all driving tasks. The autonomous driving mode may be realized by either one of the autonomous driving control and the advanced driving assistance control, a combination of the two, or switching between them.

[0015] The external sensor system 2 outputs the observation value in the sensing area AS by observing the inside of the sensing area AS set in the external world of the vehicle 4. The external sensor system 2 is composed of, for example, LiDAR (Light Detection and Ranging / Laser Imaging Detection and Ranging), radar, a camera, or a fusion of at least two types of these sensing devices.

[0016] The external sensor system 2 is controlled to repeat observation at a predetermined tracking period. When a moving object 3 is present in the sensing area AS, the external sensor system 2 outputs an observation value z k is output for each such tracking period. Here, the observed value z kis defined by Equation 1 using physical quantities shown diagrammatically in Fig. 2. In Equation 1, x and y are the horizontal and vertical center positions, respectively, of the moving body 3 in the Cartesian coordinate system defined in the observation space. In Equation 1, θ is the azimuth angle with respect to the horizontal direction of the moving body 3 in the Cartesian coordinate system defined in the observation space. In Equation 1, l and w are the front-to-rear length and left-to-right width, respectively, of the moving body 3 in the Cartesian coordinate system defined in the observation space.

number

[0017] The tracking device 1 shown in FIG. 1 is connected to an external sensor system 2 via at least one of a LAN (Local Area Network), a wire harness, an internal bus, and a wireless communication line. The tracking device 1 includes at least one dedicated computer. The dedicated computer constituting the tracking device 1 may be a driving control ECU (Electronic Control Unit) that controls driving including the autonomous driving mode of the vehicle 4. The dedicated computer constituting the tracking device 1 may be a locator ECU that estimates the self-state quantity of the vehicle 4. The dedicated computer constituting the tracking device 1 may be a navigation ECU that navigates the driving route of the vehicle 4. The dedicated computer constituting the tracking device 1 may be at least one external computer that constitutes, for example, an external center or a mobile terminal capable of communicating with the vehicle 4.

[0018] The dedicated computer constituting the tracking device 1 has at least one memory 10 and one processor 12. The memory 10 is at least one type of non-transitory tangible storage medium, such as a semiconductor memory, a magnetic medium, or an optical medium, that non-temporarily stores computer-readable programs and data. The processor 12 includes at least one type of core, such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a RISC (Reduced Instruction Set Computer)-CPU.

[0019] The processor 12 executes a plurality of instructions included in the tracking program stored in the memory 10. In this way, the tracking device 1 constructs a plurality of functional blocks for tracking the moving object 3. In this manner, in the tracking device 1, the tracking program stored in the memory 10 causes the processor 12 to execute a plurality of instructions to track the moving object 3, thereby constructing a plurality of functional blocks. As shown in FIG. 3, the plurality of functional blocks constructed by the tracking device 1 include a prediction block 100, an observation block 110, and an estimation block 120.

[0020] The flow of the tracking method in which the tracking device 1 tracks the moving object 3 by cooperation of the prediction block 100, the observation block 110, and the estimation block 120 will be described below with reference to Fig. 4. This flow is executed for each tracking period. In this flow, "S" means each of multiple steps executed by multiple commands included in the tracking program.

[0021] In S100 of the tracking method, the prediction block 100 predicts the state value of the moving object 3 at the observation time k, thereby obtaining the predicted state value Z k|k-1 and its error covariance P k|k-1 Here, we obtain the predicted state value Z k|k-1is defined by Equation 2 using physical quantities shown diagrammatically in FIG. 5. In Equation 2, X and Y are the horizontal and vertical center positions, respectively, of the moving body 3 in the Cartesian coordinate system defined in the observation space. In Equation 2, Θ is the azimuth angle with respect to the horizontal direction of the moving body 3 in the Cartesian coordinate system defined in the observation space. In Equation 2, L and W are the front-to-rear length and left-to-right width, respectively, of the moving body 3 in the Cartesian coordinate system defined in the observation space. In Equation 2, Vx and Vy (not shown in FIG. 5) are the horizontal and vertical central speeds, respectively, of the moving body 3 in the Cartesian coordinate system defined in the observation space.

number

[0022] The prediction block 100 in S100 calculates an estimated state value Z k-1|k-1 , and its error covariance P k-1|k-1 By performing a time update operation after time conversion on the predicted state value Z k|k-1 is predicted. At this time, the predicted state value Z k|k-1 and error covariance P k|k-1 is the estimated state value Z at time k-1 in the past. k-1|k-1 and error covariance P k-1|k-1 Here, Q in Equation 4 is the covariance matrix of the system noise (process noise).

number

number

number

[0023] In S110 of the tracking method shown in FIG. 4, the observation block 110 calculates the observed value z k is obtained from the external sensor system 2. The observed value zk The acquisition of z is performed by the external sensor system 2. k may be executed as k may be temporarily buffered in the memory 10 and then executed. k The acquisition of the predicted state value Z by S100 k|k-1 The above steps may be performed in parallel (i.e., simultaneously) with respect to the acquisition of the above steps, or may be performed before or after the acquisition of the above steps.

[0024] In S120 of the tracking method shown in FIG. 4, the estimation block 120 estimates the state value of the moving object 3 at the observation time k, and obtains an estimated state value Z k|k and its error covariance P k|k At this time, the estimated state value Z k|k and error covariance P k|k is the observed value z at observation time k k and the predicted state value Z k|k-1 It is obtained by nonlinear filtering that performs observation update calculations with variables.

[0025] Specifically, the estimation block 120 in S120 executes the estimation process shown in Fig. 6. In S200 of the estimation process, the estimation block 120 estimates a plurality of vertices m in a rectangular model M in which the moving object 3 is modeled as a rectangle as shown in Figs. fl ,m bl ,m fr ,m br For each, the weighting factor s fl ,s bl ,s fr ,s br Set.

[0026] At this time, the estimation block 120 in S200 estimates each vertex m fl ,m bl ,m fr ,m br Visibility of each ω fl ,ω bl ,ω fr ,ω brDepending on the weighting factor s fl ,s bl ,s fr ,s br The weight setting subroutine shown in FIG. 8 is executed to set the weights of the vertices m fl ,m bl ,m fr ,m br The process is repeated a number of times corresponding to the number of vertices m fl ,m bl ,m fr ,m br Among them, the vertex that is the target of execution each time in this subroutine is referred to as the target vertex m. In the explanation of this subroutine, except for some examples based on Fig. 7, the notations fl, bl, fr, and br, which indicate the left front, left rear, right front, and right rear of the moving body 3, respectively, as suffixes indicated by subscripts for various variables, are omitted, for example, to target vertex m.

[0027] In S300 of the weight setting subroutine shown in Fig. 8, the estimation block 120 judges whether or not an obstructing target ST as shown in Fig. 7 exists between the observation origin O of the external sensor system 2 and the target vertex m. Here, the obstructing target ST is the right rear vertex m br As shown in the example of FIG. 7, the observation origin O and the target vertex m (m br The target ST may be an object that exists separately from the moving object 3 and is located on the line segment connecting the right front vertex m fr As shown in the example of FIG. 7, the observation origin O and the target vertex m (m fr ) and may be a component part of the moving object 3 corresponding to one side of the rectangular model M. The observation origin O may be a sensing origin set in a single sensor device constituting the external sensor system 2, or may be a spatial origin assumed in the observation space by the fusion of multiple sensor devices constituting the external sensor system 2.

[0028] As shown in Fig. 8, when it is determined in S300 that a shielding target ST exists, the weight setting subroutine proceeds to S310, whereby the estimation block 120 sets the visibility ω to the minimum value. In the example shown in Fig. 7, the right rear vertex m hidden by the shielding target ST br and right front vertex m fr Visibility ω br ,ω fr is set to the minimum value of "0".

[0029] 8, when it is determined in S300 that the obstructing target ST does not exist, the weight setting subroutine proceeds to S320, whereby the estimation block 120 determines whether or not the target vertex m exists outside the sensing area AS of the external sensor system 2. When it is determined that the target vertex m exists outside the sensing area AS, the weight setting subroutine proceeds to S310, whereby the estimation block 120 sets the visibility ω to the minimum value. The sensing area AS may be a viewing angle set for a single sensor device constituting the external sensor system 2, or may be an overlapping area between the viewing angles of multiple sensor devices constituting the external sensor system 2.

[0030] If it is determined in S320 that the target vertex m does not exist outside the sensing area AS, the weight setting subroutine proceeds to S330, whereby the estimation block 120 acquires the visibility judgment angle φ. As shown in FIG. 7, the visibility judgment angle φ is defined as the minimum angle among the angles between the two sides of the target vertex m connected to the target vertex m and the line segment connecting the observation origin O and the target vertex m. Here, the left rear vertex m bl In the example of FIG. 7, the minimum angle φ bl , and the left front vertex m fl In the example of FIG. 7, the minimum angle φ fl respectively correspond to the visual recognition judgment angle φ.

[0031] As shown in Fig. 8, in S340 of the weight setting subroutine following S330, the estimation block 120 determines whether the visibility determination angle φ exceeds 90 degrees. If the visibility determination angle φ exceeds 90 degrees, the weight setting subroutine proceeds to S350, where the estimation block 120 sets the visibility ω to the maximum value. On the other hand, if the visibility determination angle φ is 90 degrees or less, the weight setting subroutine proceeds to S360, where the estimation block 120 sets the visibility ω to be equal to or greater than the minimum value and equal to or less than the maximum value according to equation 6. Here, the left rear vertex m bl In the example of FIG. 7, the visual recognition angle φ exceeds 90 degrees. bl Visibility ω bl is set to "1" as the maximum value, while the left front vertex m fl In the example of FIG. 7, the visual recognition angle φ is 90 degrees or less. fl Visibility ω fl is the calculated value in equation 6 (sinφ fl )

number

[0032] 8, in S370 of the weight setting subroutine following S310, S350, and S360, the estimation block 120 determines whether the visibility ω is the minimum value of "0". If the visibility ω is 0, the weight setting subroutine proceeds to S380, where the estimation block 120 sets the weight coefficient s to the maximum value s max On the other hand, if the visibility ω is not 0, the weight setting subroutine proceeds to S390, whereby the reciprocal of the visibility ω (1 / ω) and the maximum value s max The estimation block 120 sets the weighting coefficient s to the smaller value of these two values. In particular, when the weighting coefficient s is set to the inverse of the visibility ω, the weighting coefficient s is set to be smaller for a vertex m with a higher visibility ω.

[0033] Here, the right rear vertex m hidden by the shielding target ST br and right front vertex m fr In the example of FIG. 7, the weighting factor s is a maximum value s maxMeanwhile, the left rear vertex m bl and the left front vertex m fl In the example of FIG. 7, the maximum value of the visibility ω or the reciprocal (1 or 1 / sinφ fl ) and the maximum value s max The weighting coefficient s is set to the smaller value of s. max is the observation error e new Covariance S new is defined as a guard value greater than "1" that can prevent from becoming infinite by nonlinear filtering.

[0034] Such a weight setting subroutine by S200 is executed for all vertices m fl ,m bl ,m fr ,m br , the estimation process proceeds to S210 as shown in FIG. 6. In S210, the estimation block 120 fl ,m bl ,m fr ,m br The observation error e new Let z be the observed value at observation time k. k and the predicted state value Z k|k-1 Obtained based on the above.

[0035] At this time, the estimation block 120 in S210 calculates the nonlinear function h new And the observed value z is calculated by the matrix transformation function of Eqs. 8 to 11. k The physical quantities x, y, θ, l, and w of the rectangular model M are fl ,m bl ,m fr ,m br Expanded observation z new Here, in numbers 8 to 11, x fl ,x bl ,x fr ,x br is the observed value z as shown in Figure 2. k The horizontal position x of each vertex m fl ,m bl ,m fr ,m br Expanded to the expanded observation znew In numbers 8 to 11, y fl ,y bl ,y fr ,y br is the observed value z as shown in Figure 2. k The vertical position y of each vertex m fl ,m bl ,m fr ,m br Expanded to the expanded observation z new These are the position coordinates that make up the

number

number

number

number

number

[0036] The estimation block 120 in S210 estimates the nonlinear function h new And the predicted state value Z k|k-1 The physical quantities X, Y, Θ, L, and W of the rectangular model M are fl ,m bl ,m fr ,m br The expanded state value Z new Here, in numbers 13 to 16, X fl ,X bl ,X fr ,X br is the predicted state value Z k|k-1 The horizontal position X of each vertex m fl ,m bl ,m fr ,m br Expanded to Z, the expansion state value new In numbers 13 to 16, Y fl ,Y bl,Y fr ,Y br is the predicted state value Z k|k-1 The vertical position Y of each vertex m fl ,m bl ,m fr ,m br Expanded to Z, the expansion state value new These are the position coordinates that make up the

number

number

number

number

number

[0037] In the estimation block 120 in S210, the expanded observation value z new and the deployment state value Z new Using Equation 17, the observation error e new will be obtained.

number

[0038] In S220 of the estimation process following S210, the estimation block 120 calculates the observation error e new Covariance S new For each vertex m fl ,m bl ,m fr ,m br Weighting coefficient for each fl ,s bl ,s fr ,s br At this time, the estimation block 120 obtains the weight coefficient s fl ,s bl ,s fr ,s brUsing equation 18, each vertex m fl ,m bl ,m fr ,m br The observation error e new The 8×8 covariance matrix R new Here, R′ in Equation 18 is a covariance matrix for the horizontal and vertical positions, and is an adjustment parameter that can be adjusted by presetting. Furthermore, the estimation block 120 obtains a partial differential matrix (Jacobian) H new is obtained by Equation 19. Thus, in the estimation block 120, the observation error e new Covariance matrix R new and the partial differential matrix H new Along with the predicted state value Z k|k-1 The error covariance of P k|k-1 Using equation 20, the observation error e new Covariance S new is obtained.

number

number

number

[0039] In S230 of the estimation process following S220, the estimation block 120 performs nonlinear filtering using an extended Kalman filter to obtain a predicted state value Z k|k-1 The estimated state value Z as the updated true value k|k Let us take the error covariance P k|k At this time, the estimation block 120 obtains the observation error e new Covariance S new and the partial differential matrix H new Along with the predicted state value Z k|k-1 The error covariance of P k|k-1 Using Equation 21, the Kalman gain K new This allows the estimation block 120 to obtain the Kalman gain K new and the observation error enew Along with the predicted state value Z k|k-1 Using Equation 22, the estimated state value Z k|k Further, in the estimation block 120, the Kalman gain K new and the partial differential matrix H new Along with the predicted state value Z k|k-1 The error covariance of P k|k-1 Using Equation 23, the estimated state value Z k|k The error covariance of P k|k Here, I in Equation 23 is the identity matrix.

number

number

number

[0040] As shown in Fig. 4, the estimated state value Z k|k is output to the driving control ECU and is used for driving control including the autonomous driving mode of the vehicle 4. In addition, the estimated state value Z k|k In the tracking method for the next tracking cycle, the estimated state value Z k-1|k-1 is used for prediction in S100 by the prediction block 100.

[0041] (Action and effect) The effects of the present embodiment described above will be described below. In the description of the effects, the subscript suffixes fl, bl, fr, and fr indicating the left front, left rear, right front, and right rear of the moving body 3, respectively, are omitted.

[0042] According to this embodiment, the observed value z at the observation time k k and the predicted state value Z k|k-1The true state value at the observation time k is estimated by nonlinear filtering with the variables. At this time, the observation error e at each vertex m of the rectangular model M that models the moving object 3 is new is the observed value z at observation time k k and the predicted state value Z k|k-1 The measurement error e is obtained based on the weighting coefficient s for each vertex m. new Covariance S new Therefore, according to the present embodiment in which the weighting coefficient s is set for each vertex according to the visibility ω from the external sensor system 2, the visibility ω can be reflected in the estimation of the true value of the state value. Therefore, the estimated state value Z k|k It is possible to accurately estimate the position of the moving object 3 and improve the tracking accuracy.

[0043] According to this embodiment, the weighting coefficient s is set smaller for a vertex m with a higher visibility ω. According to this, for a vertex m with a high visibility ω from the external sensor system 2, the covariance S new As the matrix components of the matrix become smaller, the contribution of the state value to the true value estimation becomes higher. Therefore, the estimated state value Z k|k It is possible to improve the tracking accuracy of the moving object 3 by estimating the above.

[0044] According to this embodiment, for a vertex m where a shielding target ST exists between the external sensor system 2 and the vertex m, the weight coefficient s is set to the maximum value s max According to this, for a vertex m where the visibility ω from the external sensor system 2 is assumed to be substantially zero because it is hidden by a shielding target ST, the covariance S new As the matrix components of the matrix become larger, the contribution of the state value to the true value estimation becomes lower. Therefore, the estimated state value Z k|k It is possible to improve the tracking accuracy of the moving object 3 by estimating the above.

[0045] According to this embodiment, for a vertex m that exists outside the sensing area AS of the external sensor system 2, the weight coefficient s is set to the maximum value s maxAccording to this, for a vertex m where the visibility ω from the external sensor system 2 is assumed to be substantially zero because it is outside the sensing area AS, the covariance S new As the matrix components of the matrix become larger, the contribution of the state value to the true value estimation becomes lower. Therefore, the estimated state value Z k|k It is possible to improve the tracking accuracy of the moving object 3 by estimating the above.

[0046] (Other embodiments) Although one embodiment has been described above, the present disclosure should not be construed as being limited to this embodiment, and can be applied to various embodiments within the scope not departing from the gist of the present disclosure.

[0047] In a modified example, the dedicated computer constituting the tracking device 1 may include at least one of a digital circuit and an analog circuit as a processor. Here, the digital circuit is at least one of ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array), SOC (System on a Chip), PGA (Programmable Gate Array), CPLD (Complex Programmable Logic Device), etc. In addition, such a digital circuit may have a memory that stores a program.

[0048] The tracking device, the tracking method, and the tracking program according to the modified example may be applied to a device other than a vehicle. In this case, the tracking device applied to a device other than a vehicle may be mounted or installed on the same target device as the external sensor system 2, or may be mounted or installed on a target device different from the external sensor system 2. [Explanation of symbols]

[0049] 2: External sensor system, 3: Moving object, 10: Memory, 12: Processor

Claims

1. A tracking device having a processor (12) for tracking a moving object (3) by estimating state values ​​of the moving object (3) in a time series based on observed values ​​by an external sensor system (2), comprising: The processor, acquiring the observation value of the moving object observed at an observation time; obtaining a predicted state value by predicting the state value of the moving object at the observation time; and estimating a true value of the state value at the observation time by nonlinear filtering using the observation value and the predicted state value at the observation time as variables. The estimating of the true value includes: setting weighting coefficients for each of a plurality of vertices in a rectangular model that models the moving object in accordance with the visibility of each of the vertices from the external sensor system; Obtaining an observation error at each of the vertices based on the observation value and the predicted state value at the observation time; obtaining a covariance of the observation errors based on the weighting coefficient for each of the vertices.

2. Setting the weighting coefficients includes: The tracking device according to claim 1 , further comprising: setting the weighting coefficient to be smaller for a vertex having a higher degree of visibility.

3. Setting the weighting coefficients includes:

3. The tracking device according to claim 1, further comprising setting the weight coefficient to a maximum value for the vertex where an obstructing target exists between the vertex and the external sensor system.

4. Setting the weighting coefficients includes:

4. The tracking device according to claim 1, further comprising setting the weighting coefficient to a maximum value for the vertices that exist outside a sensing area of ​​the external sensor system.

5. Obtaining the predicted state value includes: The tracking device according to any one of claims 1 to 4, further comprising: obtaining the predicted state value at the observation time based on the true value estimated at a past time prior to the observation time.

6. The estimating of the true value includes: The tracking device according to any one of claims 1 to 5, further comprising: acquiring the true value by updating the predicted state value through the nonlinear filtering using an extended Kalman filter.

7. A tracking method for tracking a moving object by estimating state values ​​of the moving object in a time series manner based on observation values ​​by an external sensor system, the method comprising: acquiring the observation value of the moving object observed at an observation time; obtaining a predicted state value by predicting the state value of the moving object at the observation time; estimating a true value of the state value at the observation time by nonlinear filtering using the observation value and the predicted state value at the observation time as variables; The estimating of the true value includes: setting weighting coefficients for each of a plurality of vertices in a rectangular model that models the moving object in accordance with the visibility of each of the vertices from the external sensor system; Obtaining an observation error at each of the vertices based on the observation value and the predicted state value at the observation time; and obtaining a covariance of the observation errors based on the weighting coefficients for each of the vertices.

8. A tracking program stored in a storage medium (10), the tracking program including instructions to be executed by a processor (12) for estimating state values ​​of a moving object (3) in a time series based on observation values ​​by an external sensor system (2) and tracking the moving object, The instruction: acquiring the observation value of the moving object observed at an observation time; predicting the state value of the moving object at the observation time to obtain a predicted state value; and estimating a true value of the state value at the observation time by nonlinear filtering using the observation value and the predicted state value at the observation time as variables, The estimation of the true value includes: setting weight coefficients for each of a plurality of vertices in a rectangular model that models the moving object in accordance with the visibility of each of the vertices from the external sensor system; Obtaining an observation error at each of the vertices based on the observation value and the predicted state value at the observation time; and obtaining a covariance of the observation errors based on the weighting coefficient for each of the vertices.

Images