Driving assistance device, vehicle, and driving assistance method

Abstract

A driving assist device of an embodiment of the present disclosure includes: an information acquisition unit configured to acquire observation information including one or more first phenomenon information; a hypothesis inference unit configured to perform hypothesis inference by applying background knowledge information including a plurality of evidence information to the observation information, thereby inferring a plurality of hypotheses each including one or more second phenomenon information; a derivation unit configured to derive a cost of each of the first phenomenon information included in the observation information; and a determination unit configured to determine a hypothesis having a lowest total cost among the plurality of hypotheses, i.e., a maximum likelihood hypothesis, based on the cost of each of the first phenomenon information included in the observation information and the cost of each of the second phenomenon information included in each of the hypotheses. The derivation unit is configured to derive the cost of each of the first phenomenon information based on information of a reliability of each of the first phenomenon information in the observation information when the information of the reliability is obtained, and derive the cost of each of the first phenomenon information based on a predetermined parameter related to recognition of an object to be recognized by the information acquisition unit when the information of the reliability of each of the first phenomenon information in the observation information is not obtained.

Description

Technical Field

[0001] This disclosure relates to driving assistance devices, vehicles, and driving assistance methods. Background Technology

[0002] Various technologies have been disclosed as driving assistance devices and driving assistance methods that use hypothetical inference to assist the driving of a target vehicle (for example, see Patent Document 1).

[0003] Existing technical documents Patent documents Patent Document 1: Japanese Patent Application Publication No. 2016-91039 Summary of the Invention

[0004] One embodiment of the driving assistance device disclosed herein is a device for assisting the driving of a target vehicle using hypothesis inference. The driving assistance device includes: an information acquisition unit configured to acquire observation information including one or more first phenomenon information; a hypothesis inference unit configured to perform hypothesis inference by applying background knowledge information, including multiple pieces of evidence information, to the observation information, thereby inferring multiple hypotheses, each including one or more second phenomenon information; a derivation unit configured to derive the cost of each first phenomenon information included in the observation information; and a decision unit configured to determine the hypothesis with the lowest total cost among the multiple hypotheses, i.e., the maximum likelihood hypothesis, based on the cost of each first phenomenon information included in the observation information and the cost of each second phenomenon information included in each hypothesis. The derivation unit is configured to derive the cost of each first phenomenon information based on the reliability information when the reliability information of each first phenomenon information in the observation information is obtained, and to derive the cost of each first phenomenon information based on predetermined parameters related to the identification of the target object by the information acquisition unit when the reliability information of each first phenomenon information in the observation information is not obtained.

[0005] A vehicle according to one embodiment of the present disclosure is equipped with the driving assistance device described in one embodiment of the present disclosure.

[0006] One embodiment of the driving assistance method disclosed herein is a method for assisting a vehicle using hypothesis inference. The driving assistance method includes: acquiring observation information including one or more first phenomenon information; applying background knowledge information, including multiple pieces of evidence information, to the observation information to perform hypothesis inference, thereby inferring multiple hypotheses, each including one or more second phenomenon information; deriving the cost of each first phenomenon information included in the observation information; and determining the hypothesis with the lowest total cost among the multiple hypotheses, i.e., the maximum likelihood hypothesis, based on the cost of each first phenomenon information included in the observation information and the cost of each second phenomenon information included in each hypothesis. The cases of deriving the cost of each first phenomenon information include: deriving the cost of each first phenomenon information based on the reliability information when the reliability information of each first phenomenon information in the observation information is obtained; and deriving the cost of each first phenomenon information based on predetermined parameters related to the identification of the target object when the reliability information of each first phenomenon information in the observation information is not obtained. Attached Figure Description

[0007] The accompanying drawings are provided to further understand this disclosure and are incorporated in and constitute a part of this specification. The drawings serve to illustrate one embodiment and, together with the description, explain the principles of this disclosure.

[0008] Figure 1 This is a schematic diagram illustrating a general configuration example of a vehicle equipped with a driving assistance device according to an embodiment of the present disclosure.

[0009] Figure 2 It means Figure 1 A block diagram illustrating a detailed configuration example of a vehicle.

[0010] Figure 3 This is a flowchart illustrating an example of a driving assistance process implemented in this way.

[0011] Figure 4 It is used for explanation Figure 3 A schematic diagram illustrating the outline of the hypothetical inferences.

[0012] Figure 5 It is used for explanation Figure 3 Another schematic diagram illustrating the outline of the hypothetical inferences.

[0013] Figure 6 It is shown Figure 3 The flowchart shows a detailed example of the cost derivation process for the first phenomenon information.

[0014] Figure 7 It is shown Figure 6 A diagram showing a specific example of the predetermined parameters.

[0015] Figure 8 It is used for explanation Figure 7 The diagram shows the occlusion rate.

[0016] Figure 9 It is used for explanation Figure 7 Another schematic diagram showing the occlusion rate.

[0017] Figure 10 It is used for explanation Figure 7 A schematic diagram showing the relative distances.

[0018] Figure 11 It is used for explanation Figure 7 Another schematic diagram showing the relative distance.

[0019] Figure 12 It is used for explanation Figure 7 The diagram shown illustrates the determination of duration.

[0020] Figure 13 It is used for explanation Figure 7 Another schematic diagram showing the determination of duration.

[0021] Figure 14 It is used for explanation Figure 7 Another schematic diagram showing the determination of duration. Detailed Implementation

[0022] In driver assistance devices and similar applications that use hypothesis reasoning to assist driving a vehicle, there is a need to improve the ease of determining the most likely hypothesis. There is a desire to provide driver assistance devices, vehicles, and driving assistance methods that can improve the ease of determining the most likely hypothesis.

[0023] Hereinafter, several exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. It should be noted that the following description illustrates a specific example of the present disclosure and should not be construed as limiting the present disclosure. For example, elements including numerical values, shapes, materials, components, the positions of the components, and the connection methods of the components are merely examples and should not be construed as limiting the present disclosure. Furthermore, in the following exemplary embodiments, constituent elements not described in the independent claims based on the highest concept of the present disclosure are arbitrary and can be provided as needed. The drawings are schematic and are not intended to be illustrated at their original dimensions. Throughout this specification and the drawings, constituent elements having substantially the same function and substantially the same structure are labeled with the same reference numerals, and repeated descriptions are omitted. Additionally, constituent elements not directly related to an embodiment of the present disclosure are not illustrated in the drawings.

[0024] <1. Implementation Method> [constitute] Figure 1 A schematic example of the configuration of a vehicle 1 equipped with a driving assistance device (driving assistance device 111 described below) according to an embodiment of the present disclosure is shown. Figure 2 A block diagram is shown. Figure 1 The detailed configuration example of vehicle 1 shown.

[0025] like Figure 1 , Figure 2 As shown, the vehicle 1 includes a vehicle control unit 11, a battery 12, a communicator 13, and a camera 14. It should be noted that, as Figure 1 As shown, occupant 9 (driver, etc.) is riding in vehicle 1.

[0026] (A. Vehicle Control Unit 11) The vehicle control unit 11 is a component (control unit) that controls various actions or performs various calculations in the vehicle 1. Specifically, the vehicle control unit 11 is configured, for example, to include one or more processors (CPU: Central Processing Unit) that execute programs, and one or more memories that are communicatively connected to these processors. In addition, such memories are configured, for example, to temporarily store processed data in RAM (Random Access Memory) and to store programs in ROM (Read Only Memory).

[0027] exist Figure 2 In the example shown, the vehicle control unit 11 includes a driving assistance device 111, a battery control unit 112, a communication control unit 113, and a camera control unit 114.

[0028] (A-1. Driving assistance device 111) Driving assistance device 111 is a device that provides driving assistance to the target vehicle (vehicle 1) using the hypothetical inference described later. Figure 2 In the example shown, the driving assistance device 111 includes an information acquisition unit 21, a hypothesis inference unit 22, a derivation unit 23, a decision unit 24, a notification unit 25, and a driving control unit 26.

[0029] It should be noted that the notification unit 25 and the driving control unit 26 each correspond to a specific example of the "control unit" in one embodiment of this disclosure.

[0030] The information acquisition unit 21 is a unit that acquires various information related to the vehicle 1. Specifically, in this embodiment, the information acquisition unit 21 acquires observation information Im, including one or more first phenomenon information Ie1 described later, using the camera 14, various sensors, etc., as one of these various types of information. It should be noted that details regarding such observation information Im will be described later.

[0031] Hypothesis inference unit 22 infers hypotheses by applying background knowledge information Ibk, including multiple pieces of evidence information Ir (described later), to the aforementioned observation information Im, thereby inferring units of multiple hypotheses H, each including one or more pieces of second phenomenon information Ie2 (described later). It should be noted that details of this hypothesis inference process will be described later.

[0032] The derivation unit 23 is a unit that derives the cost C of each first phenomenon information Ie1 included in the aforementioned observation information Im. It should be noted that details regarding the derivation process of the cost C of such first phenomenon information Ie1 will be described later.

[0033] The decision unit 24 determines the unit with the lowest total cost among multiple hypotheses H, namely the maximum likelihood hypothesis Hm, based on the cost C of each first phenomenon information Ie1 included in the aforementioned observation information Im and the cost C of each second phenomenon information Ie2 included in the aforementioned hypotheses H. It should be noted that the details of the decision-making process for such a maximum likelihood hypothesis Hm will be described later.

[0034] The notification unit 25 is a unit that performs a predetermined notification action (notification processing) to the occupants 9 (driver, etc.) of the vehicle 1. Such a notification unit 25 may be configured to include, for example, a display unit (such as a display of various types) that performs a predetermined display action as such a notification action, and an audio output unit (such as a speaker of various types) that performs a predetermined sound output action as such a notification action.

[0035] The driving control unit 26 is a unit that controls the driving actions of the vehicle 1 (performs driving control processing) and performs comprehensive control related to the driving of the vehicle 1. Specifically, the driving control unit 26 controls, for example, the drive system, braking system, and steering system in the vehicle 1.

[0036] In this embodiment, the notification unit 25 and the driving control unit 26 respectively utilize the maximum likelihood hypothesis Hm determined by the decision unit 24 (based on the second phenomenon information Ie2 included in the maximum likelihood hypothesis Hm) to provide driving assistance to the vehicle 1. Specifically, the notification unit 25 performs notification processing using such a maximum likelihood hypothesis Hm, and the driving control unit 26 performs driving control processing using such a maximum likelihood hypothesis Hm.

[0037] (A-2. Battery control unit 112, communication control unit 113, camera control unit 114) The battery control unit 112 is a unit that controls the operation of the battery 12 (charging operation, discharging operation, etc.). The communication control unit 113 is a unit that controls the communication operation of the communicator 13. The camera control unit 114 is a unit that controls the shooting operation of the camera 14.

[0038] (B. Battery 12, Communicator 13, Camera 14) Battery 12 is a component that functions as a power source for vehicle 1, and is constructed using various rechargeable batteries such as lithium-ion batteries. Communicator 13 is a device that enables various communications with external parts of vehicle 1 (e.g., information processing devices such as servers located outside vehicle 1, other vehicles besides vehicle 1, etc.).

[0039] Camera 14 is a component that acquires photographic data related to the exterior (surrounding environment of vehicle 1, etc.) and interior (occupants of vehicle 1, etc.) of vehicle 1. It should be noted that the photographic data acquired in this way is supplied, for example, to a driver assistance device 111 (such as the aforementioned information acquisition unit 21) within the vehicle control unit 11.

[0040] [Actions and their functions / effects] Next, the actions and functions / effects of this embodiment will be described in detail.

[0041] Figure 3 A flowchart illustrates an example of driving assistance processing in this embodiment (a processing example in driving assistance device 111). Figure 4 , Figure 5 These are used for Figure 3 A schematic diagram illustrating the outline of the hypothetical inference (step S3 described later) is provided. It should be noted that... Figure 5 In this context, on a road with road signs St (green signs) on three driving lanes L1 to L3, there is a situation where a vehicle 7 is traveling in front of the vehicle being processed (vehicle 1, which is traveling in driving lane L2) and a parked vehicle 8 is parked in front of the vehicle 7.

[0042] (A. Summary of the hypotheses and inferences) In the Figure 3 Before explaining the overall driver assistance process shown, a summary of the hypothetical inferences will be provided.

[0043] First, hypothetical reasoning refers to the method of applying rules to deduce premises from a conclusion. On the other hand, deduction refers to the method of deduce conclusions based on rules and premises. Thus, hypothetical reasoning and deduction have different directions of inference; hypothetical reasoning is also called backward reasoning, and deduction is also called forward reasoning. Applying such hypothetical reasoning to vehicle driver assistance processing (hazard prediction) is as follows: That is, by applying the aforementioned background knowledge information Ibk (equivalent to "rules") to the observed information Im and the phenomena representing danger (equivalent to "conclusion"), the aforementioned hypothesis H (equivalent to "premise") is deduced.

[0044] Furthermore, the degree to which the hypothetical H derived in this way is consistent with the observed information Im (the goodness of the proof) can be output as the cost C. That is, it can be said that a relatively low cost C is a relatively good proof. When outputting such a cost C, the following two "weights" are used.

[0045] • The “weight” representing the importance of each piece of evidence information Ir is registered in the background knowledge information Ibk. The higher the importance of the evidence information Ir, the higher the “weight” is set.

[0046] • The “weight” representing the reliability R of each piece of information (first phenomenon information Ie1) included in the observation information Im is registered in the observation information Im. The higher the reliability R of the first phenomenon information Ie1, the lower the “weight” is set.

[0047] Here, refer to Figure 4 , Figure 5 This section provides a specific example of how such a generalized assumption is handled. It should be explained that... Figure 4 The information content and cost C values ​​shown are merely examples and can be arbitrarily set. Furthermore, Figure 4 The notation “^” shown represents logical AND (“AND”).

[0048] First, in the observation information Im obtained through various sensors, etc., a cost C is assigned to each piece of first phenomenon information Ie1 according to the weight of the aforementioned representation reliability R. Figure 4 In the example shown, the cost C = $10 is assigned to the first phenomenon information Ie1, which indicates "there is a vehicle ahead" and "there is a green sign," respectively. Additionally, in this... Figure 4 In the example shown, since the first phenomenon information Ie1 indicating "danger has occurred" is uncertain (low reliability R) information that has not been directly observed, its cost C is set to 10 times the value mentioned above (cost C = $100). It should be noted that in... Figure 5 In the example shown, "danger occurred" indicates a situation where vehicle 1 comes into contact with vehicle 7 traveling in front (see reference). Figure 5(Arrow P1 in the image).

[0049] Next, hypothesis H is derived by applying the background knowledge information Ibk to such observation information Im. Figure 4 In the example shown, the background knowledge information Ibk includes multiple pieces of evidence information Ir (representing "there is a vehicle moving forward", "there is a parked vehicle", and "there is a green sign"), and the weights (weights Wr) representing the importance of these pieces of evidence are set to 0.3, 0.6, and 0.3, respectively. Additionally, in Figure 4 In the example shown, hypothesis H includes multiple second phenomenon information Ie2 (representing "there is a vehicle moving forward", "there is a parked vehicle", "there is a green sign").

[0050] Here, the total cost (the sum of costs C) obtained in assumption H is thus realized in... Figure 4 In the example shown, the cost becomes $120 ($30 + $60 + $30), which is 1.2 times the value before the inference (the cost C = $100 in "danger occurs"). That is, the total cost increases because hypothesis H is not observed information. This increased total cost is allocated to the multiple pieces of second-phenomenal information Ie2 that constitute the factors (text) of hypothesis H. At this point, the importance (weight Wr) of each of the aforementioned pieces of evidence information Ir is reflected.

[0051] Next, when there is the same type of information (text) between the first phenomenon information Ie1 included in the observation information Im and the second phenomenon information Ie2 included in hypothesis H, a simplification process is performed as follows: In this simplification process, information with relatively high cost C is removed from the first phenomenon information Ie1 and the second phenomenon information Ie2, and only information with relatively low cost C is used. Figure 4 In the example shown, information with relatively low cost C (the first phenomenon information Ie1 indicating "there is a vehicle in front") is used through the simplification process shown by reference numeral U1. Additionally, information with relatively low cost C (the first phenomenon information Ie1 indicating "there is a green marker") is used through the simplification process shown by reference numeral U2.

[0052] Thus, the hypothesis H that minimizes total cost (and becomes the best proof) is determined, namely the maximum likelihood hypothesis Hm. Figure 4 In the example shown, the maximum likelihood hypothesis Hm includes multiple second phenomenon information Ie2 (representing "there is a vehicle moving forward", "there is a green sign", "there is a parked vehicle"), and the total cost becomes $80 ($10 + $10 + $60).

[0053] (B. Driver Assistance) Next, based on the summary of the above-mentioned hypothetical inferences, we will... Figure 3 The driving assistance processing of this embodiment will be described in detail.

[0054] (S1: Acquisition and processing of observation information Im) In Figure 3 In the series of processing examples shown, firstly, the information acquisition unit 21 acquires observation information Im, including the aforementioned first phenomenon information Ie1, as one of various types of information related to vehicle 1 (step S1). It should be noted that the observation information Im includes not only information at the current moment, but may also include past information (information from the present to the past few seconds, etc.).

[0055] (S2: Derivation and processing of the cost C of the first phenomenon information Ie1) Next, the derivation unit 23 derives the cost C of each first phenomenon information Ie1 included in the aforementioned observation information Im (step S2). It should be noted that the derivation unit 23 may, for example, periodically (on a daily basis, etc.) update the value of the cost C of each such first phenomenon information Ie1 (see the following description). Figure 6 ).

[0056] Here, Figure 6 The flowchart shows Figure 3 A detailed example of the derivation process (step S2) of the cost C of the first phenomenon information Ie1 shown.

[0057] In Figure 6 In the processing example shown, firstly, the derivation unit 23 determines whether information on the reliability R of each first phenomenon information Ie1 in the observation information Im has been obtained (step S21). Here, if it is determined that information on the reliability R of each first phenomenon information Ie1 has been obtained (step S21: yes), the derivation unit 23 derives the cost C of each first phenomenon information Ie1 based on the information on the reliability R of each first phenomenon information Ie1 (step S22). This should be explained, and will be discussed later. Figure 3 Step S3 in the process.

[0058] On the other hand, if it is determined that the reliability R of each first phenomenon information Ie1 has not been obtained (step S21: No), the derivation unit 23 derives the cost C of each first phenomenon information Ie1 as follows. That is, in this case, the derivation unit 23 derives the cost C of each first phenomenon information Ie1 based on a predetermined parameter Pr related to the identification of the object 6 by the information acquisition unit 21 (step S23). Specific examples of such a predetermined parameter Pr will be described later. Figures 7-14 (This should be explained before proceeding to the following explanation.) Figure 3Step S3 in the process.

[0059] (S3: Hypothesis Inference Processing) The derivation process of the cost C of the first phenomenon information Ie1 mentioned above ( Figure 3 After step S2), the above-mentioned hypothesis inference is performed through hypothesis inference unit 22, thereby inferring multiple hypotheses H ( Figure 3 Step S3). Specifically, the hypothesis inference unit 22 applies background knowledge information Ibk to the observation information Im obtained in step S1 to make hypothesis inferences, thereby inferring multiple hypotheses H, including the second phenomenon information Ie2.

[0060] (S4: Decision processing for the maximum likelihood hypothesis Hm) Next, the decision unit 24 performs the above-mentioned maximum likelihood hypothesis Hm decision process (step S4). Specifically, the decision unit 24 determines the hypothesis H with the lowest total cost among the multiple hypotheses H, i.e., the maximum likelihood hypothesis Hm, based on the cost C of each first phenomenon information Ie1 included in the observation information Im and the cost C of each second phenomenon information Ie2 included in the multiple hypotheses H obtained in step S3.

[0061] (S5: Driver Assistance Processing) Next, the notification unit 25 and the driving control unit 26 respectively use the maximum likelihood assumption Hm determined in step S4 to perform driving assistance for vehicle 1 (step S5). Specifically, the notification unit 25 uses such a maximum likelihood assumption Hm to perform the various notification processes described above, and the driving control unit 26 uses such a maximum likelihood assumption Hm to perform the various driving control processes described above.

[0062] above, Figure 3 The series of processing examples shown has ended.

[0063] (C. Specific examples of parameter Pr) Next, a specific example (specific embodiment) of the predetermined parameter Pr will be described in detail.

[0064] Figure 7 It shows Figure 6 The example shown is a specific instance of the predetermined parameter Pr. Additionally, Figure 8 , Figure 9 They are used for Figure 7 The diagram shown below illustrates the occlusion rate (px) as described later. Figure 10 , Figure 11 They are used for Figure 7 The diagram shown below illustrates the relative distance dx, which will be discussed later. Figures 12-14 They are used for Figure 7 The diagram shown below illustrates the determination of duration tx.

[0065] like Figure 7 As shown, in this embodiment, the predetermined parameter Pr mentioned above includes at least one of the following: occlusion rate px, relative distance dx, and determination duration tx.

[0066] • Occlusion rate (px) of object 6 (identified object) • The relative distance dx between vehicle 1 (object vehicle) and the identified object 6 • Determination duration tx (the length of the period during which object 6 is continuously determined to be in the same category) (C-1. Occlusion rate in pixels) First, except Figure 7 In addition, refer to Figure 8 , Figure 9 The above occlusion rate (px) will be explained.

[0067] As the occlusion rate px of the object 6 being identified increases, the cost C=Cx (e.g., proportionally) of the first phenomenon information Ie1 increases (see reference). Figure 7 Specifically, in Figure 8 In the example, as the occlusion rate increases to px=25%, 50%, and 100%, the cost (proportional) increases to Cx=$25, $50, and $100.

[0068] Here, in the export section 23, the process of exporting the cost Cx based on the occlusion rate px is as follows (steps S31 to S35 below: refer to...). Figure 9 ).

[0069] S31: Determine the category of the object 6 identified by the information acquisition unit 21. S32: Based on the category determination result, estimate the object area a (all) in the case where the entire area of ​​the identified object 6 can be identified (for example, a predefined table may also be used). S33: Calculate the area of ​​the region in the object 6 that can be recognized (by the information acquisition unit 21), i.e., the recognition area a (vis). S34: Based on the object area a(all) and the recognition area a(vis), the occlusion rate px of the recognized object 6 is calculated using the following formula (1-1).

[0070] • S35: Use the following formula (1-2) to calculate the cost Cx mentioned above.

Mathematical Formula 1

[0071] (C-2. Relative distance dx) Next, except Figure 7 In addition, refer to Figure 10 , Figure 11 The relative distance dx mentioned above will be explained.

[0072] As the relative distance dx between such vehicle 1 and the identified object 6 increases (see reference 23), the output section 23 continues to expand. Figure 10 The relative distances dx = d1, d2, and d3 shown in the examples of near, medium, and far distances are processed as follows. That is, as the relative distance dx increases, the cost C = Cx of the first phenomenon information Ie1 (e.g., in an exponential function relationship) increases (see reference). Figure 7 Specifically, in Figure 10 In the example, as the relative distance increases to dx = 10 [m], 100 [m], and 125 [m], the cost (in an exponential function) increases to Cx = $1, $37, and $100. It should be noted that... Figure 10 The example is a numerical example in the case of dmin=10 and α=25 in equation (2-1) as explained below.

[0073] Here, in the export section 23, the process of exporting the cost Cx based on the occlusion rate px is as follows (steps S41 and S42 below).

[0074] S41: Determine the relative distance dx of the identified object 6. •S42: Use the following equations (2-1) and (2-2) to calculate the cost Cx mentioned above (refer to...). Figure 11 (dmin, dmax: the minimum and maximum values ​​of the relative distance dx that can be identified by the information acquisition unit 21, α: a constant obtained by equation (2-2))

Mathematical Formula 2

[0075] (C-3. ​​Determine the duration tx) Next, except Figure 7 In addition, refer to Figures 12-14 The determination of the duration tx mentioned above will be explained.

[0076] First, as mentioned above, the determination duration tx is the length of the period during which the identified object 6 is continuously determined to be of the same category. Specifically, in Figure 12In the example (an example of the correspondence between time t and category determination value J), ​​the period during which the category determination value J of the identified object 6 continuously represents the same value (the period during which it is continuously determined to be of the same category) is defined as the determination duration tx. As such determination duration tx increases, the cost C=Cx of the first phenomenon information Ie1 decreases (for example, in an inversely proportional relationship) (see reference). Figure 7 Specifically, in Figure 13 , Figure 14 In the example, as the decision duration increases to tx = 0.1[s], 0.5[s], and 1.0[s], the cost (inversely proportionally) decreases to Cx = $100, $20, and $10. It should be noted that... Figure 13 , Figure 14 The example becomes a numerical example in the case of β=10 in equation (3-1) as explained below.

[0077] Here, in the export section 23, the process of exporting the cost Cx based on the occlusion rate px is as follows (steps S51 to S53 below).

[0078] S51: Determine the category of object 6 (calculate the category determination value J mentioned above). S52: Based on the above definition, determine the duration tx. S53: The cost Cx mentioned above is calculated using the following equations (3-1) and (3-2) (tmin: the minimum value of the determination duration tx that can be calculated in the derivation section 23, β: a constant obtained by equation (3-2)).

Mathematical Expression 3

[0079] (D. Functions and effects) Thus, in this embodiment, by applying background knowledge information Ibk to the observation information Im, including the first phenomenon information Ie1, hypothesis inference is performed, thereby inferring multiple hypotheses H, including the second phenomenon information Ie2. Furthermore, based on the cost C of each first phenomenon information Ie1 and the cost C of each second phenomenon information Ie2, the maximum likelihood hypothesis Hm among the multiple hypotheses H is determined. At this time, if the reliability R of each first phenomenon information Ie1 is obtained, the cost of each first phenomenon information Ie1 is derived based on this reliability R. On the other hand, if the reliability R of each first phenomenon information Ie1 is not obtained, the cost C of each first phenomenon information Ie1 is derived based on a predetermined parameter Pr related to the identification of the object 6.

[0080] Therefore, even when the reliability R of each first phenomenon information Ie1 in the observation information Im is unavailable, the cost C of each first phenomenon information Ie1 in the observation information Im can be derived using the predetermined parameter Pr described above. As a result, in this embodiment, the convenience of determining the maximum likelihood hypothesis Hm can be improved.

[0081] Furthermore, in this embodiment, the predetermined parameter Pr mentioned above includes at least one of the occlusion rate px, relative distance dx, and decision duration tx, as described below. That is, by utilizing these various parameters Pr, the cost C of each first phenomenon information Ie1 can be derived more appropriately. As a result, the convenience of determining the maximum likelihood hypothesis Hm can be further improved.

[0082] <2. Variations> While the present disclosure has been described above with examples of several implementation methods and embodiments, the present disclosure is not limited to these embodiments and various modifications are possible.

[0083] For example, the configuration (form, arrangement, number, etc.) of the various components in a vehicle or the like is not limited to the configuration described in the above embodiments. That is, the configuration of these components may also be other forms, arrangements, numbers, etc. Specifically, for example, in the above embodiments, an example of a driving assistance device according to an embodiment of the present disclosure being installed in a vehicle for driving assistance has been described, but the present disclosure is not limited to that example. That is, at least some of the elements (information acquisition unit, hypothesis inference unit, derivation unit, and decision unit, etc.) within the driving assistance device may, for example, be installed inside a server (information processing device) or the like installed outside the target vehicle.

[0084] Furthermore, in the above embodiments, various processing examples (such as the overall driving assistance processing described above, the derivation processing of the cost C of the first phenomenon information Ie1, the hypothesis inference processing, and the decision processing of the maximum likelihood hypothesis Hm, etc.) have been specifically listed and described, but the methods described in the above embodiments are not limited to those described in the above embodiments; for example, other methods may also be used. In addition, in the above embodiments, specific examples (embodiments) of the predetermined parameter Pr have been described in detail, but the methods are not limited to those specific examples; other types of parameters (such as the color, material, shape, etc. of the object to be identified) may also be used as the predetermined parameter Pr.

[0085] Furthermore, the values, ranges, and magnitude relationships of the various parameters described in the above embodiments are not limited to those described in the above embodiments, and may be other values, ranges, and magnitude relationships.

[0086] Furthermore, in the above embodiments, examples of vehicles 1 equipped with a battery 12 as a power source have been described. That is, examples of vehicles 1 being electric vehicles (EVs) or hybrid electric vehicles (HEVs) have been described, but the examples are not limited to these. For example, vehicles 1 could also be gasoline vehicles or fuel cell vehicles.

[0087] Furthermore, the series of processes described in the above embodiments can be executed by hardware (circuit), by software (program), or by a combination of hardware and software. In the case of software execution, the software consists of a group of programs used to enable the computer to perform various functions. Each program can be pre-installed on the computer for use, or it can be installed on the computer from a network or recording medium.

[0088] Furthermore, the various examples described so far can be applied in any combination.

[0089] It should be noted that the effects described in this specification are merely illustrative and not limiting; other effects may also be present.

[0090] Furthermore, this disclosure can take the following form. (1) A driving assistance device is a device that uses hypothetical reasoning to assist the driving of a target vehicle, the driving assistance device comprising: The information acquisition unit is configured to acquire observational information, including information about one or more first phenomena. The hypothesis inference section is configured to infer the hypothesis by applying background knowledge information, including multiple pieces of evidence, to the observation information, thereby inferring multiple hypotheses, each including one or more second phenomenon information. The derivation unit is configured to derive the cost of each piece of information about the first phenomenon included in the observation information; and The decision-making unit is configured to determine the hypothesis with the lowest total cost, i.e., the maximum likelihood hypothesis, based on the cost of each piece of first phenomenon information included in the observation information and the cost of each piece of second phenomenon information included in each hypothesis. The derived part is configured as follows: Having obtained the reliability information for each of the first phenomena in the observation information, the cost of each of the first phenomena is derived based on the reliability information. Without obtaining information on the reliability of each of the first phenomenon information in the observation information, the cost of each of the first phenomenon information is derived based on predetermined parameters related to the identification of the object by the information acquisition unit. (2) According to the driving assistance device described in (1) above, The predetermined parameters include at least one of the following: the occlusion rate of the identified object, the relative distance between the object vehicle and the identified object, and the length of the period during which the identified object is continuously classified as the same category, i.e., the determination duration. (3) According to the driving assistance device described in (2) above, The driving assistance device includes the occlusion rate as the predetermined parameter. The output section is configured such that the cost of the first phenomenon information increases as the occlusion rate increases. (4) According to the driving assistance device described in (2) or (3) above, The driving assistance device includes the relative distance as the predetermined parameter. The derived section is configured such that the cost of the first phenomenon information increases as the relative distance increases. (5) According to any one of (2) to (4) above, the driving assistance device, The driving assistance device includes the determination duration as the predetermined parameter. The derivation unit is configured to reduce the cost of the first phenomenon information as the determination duration increases. (6) According to any one of (1) to (5) above, The derived part is configured to periodically update the cost value for each piece of information about the first phenomenon. (7) According to any one of (1) to (6) above, the driving assistance device, The driving assistance device also includes a control unit configured to provide driving assistance to the target vehicle using the maximum likelihood assumption. (8) A vehicle having any one of the driving assistance devices described in (1) to (7) above. (9) A driving assistance method is a method of using hypothesis reasoning to assist the driving of a target vehicle, the driving assistance method comprising: Acquire observational information, including information on one or more primary phenomena; By applying background knowledge information, including multiple pieces of evidence, to the observation information, the hypothesis is inferred, thereby inferring multiple hypotheses, each including one or more second phenomenon information. The cost of deriving each piece of information about the first phenomenon included in the observation information; and Based on the cost of each piece of first phenomenon information included in the observation information and the cost of each piece of second phenomenon information included in each of the hypotheses, the hypothesis with the lowest total cost is determined, which is the maximum likelihood hypothesis. The cost of exporting each piece of information about the first phenomenon includes: Having obtained information on the reliability of each of the first phenomena in the observation information, the cost of each of the first phenomena is derived based on the reliability information; and Without obtaining information on the reliability of each of the first phenomenon information in the observation information, the cost of each of the first phenomenon information is derived based on predetermined parameters related to the identification of the object being identified.

[0100] Figure 1 , Figure 2 The vehicle control unit 11 shown can be implemented by a circuit including at least one semiconductor integrated circuit, such as at least one processor (e.g., a central processing unit (CPU)), at least one application-specific integrated circuit (ASIC), and / or at least one field-programmable gate array (FPGA). The at least one processor can be configured to execute by reading instructions from at least one non-transitory and tangible computer-readable medium. Figure 1 , Figure 2 This refers to all or part of the various functions in the vehicle control unit 11 shown. Such media can take various forms, including, but are not limited to, various magnetic media such as hard disks, various optical media such as CDs or DVDs, and various semiconductor memories (i.e., semiconductor circuits) such as volatile or non-volatile memory. Volatile memory can include DRAM and SRAM. Non-volatile memory can include ROM and NVRAM. ASIC is dedicated to executing... Figure 1 , Figure 2 The vehicle control unit 11 shown contains all or some of the various functions of an integrated circuit (IC). An FPGA is designed to be configured after manufacturing to execute... Figure 1 , Figure 2 All or some of the functions in the vehicle control unit 11 shown.

Claims

1. A driving assistance device, characterized in that, It is a device that uses hypothetical reasoning to assist the driving of a target vehicle, the driving assistance device comprising: The information acquisition unit is configured to acquire observational information, including information about one or more first phenomena. The hypothesis inference section is configured to infer the hypothesis by applying background knowledge information, including multiple pieces of evidence, to the observation information, thereby inferring multiple hypotheses, each including one or more second phenomenon information. The derivation unit is configured to derive the cost of each piece of information about the first phenomenon included in the observation information; and The decision-making unit is configured to determine the hypothesis with the lowest total cost, i.e., the maximum likelihood hypothesis, based on the cost of each piece of first phenomenon information included in the observation information and the cost of each piece of second phenomenon information included in each hypothesis. The derived part is configured as follows: Having obtained the reliability information for each of the first phenomena in the observation information, the cost of each of the first phenomena is derived based on the reliability information. Without obtaining information on the reliability of each of the first phenomenon information in the observation information, the cost of each of the first phenomenon information is derived based on predetermined parameters related to the identification of the object by the information acquisition unit.

2. The driving assistance device according to claim 1, characterized in that, The predetermined parameters include at least one of the following: the occlusion rate of the identified object, the relative distance between the object vehicle and the identified object, and the length of the period during which the identified object is continuously classified as the same category, i.e., the determination duration.

3. The driving assistance device according to claim 2, characterized in that, The driving assistance device includes the occlusion rate as the predetermined parameter. The output section is configured such that the cost of the first phenomenon information increases as the occlusion rate increases.

4. The driving assistance device according to claim 2, characterized in that, The driving assistance device includes the relative distance as the predetermined parameter. The derived section is configured such that the cost of the first phenomenon information increases as the relative distance increases.

5. The driving assistance device according to claim 2, characterized in that, The driving assistance device includes the determination duration as the predetermined parameter. The derivation unit is configured to reduce the cost of the first phenomenon information as the determination duration increases.

6. The driving assistance device according to any one of claims 1 to 5, characterized in that, The derived part is configured to periodically update the cost value for each piece of information about the first phenomenon.

7. The driving assistance device according to any one of claims 1 to 5, characterized in that, The driving assistance device also includes a control unit configured to provide driving assistance to the target vehicle using the maximum likelihood assumption.

8. A vehicle, characterized in that, It has a driving assistance device as described in any one of claims 1 to 5.

9. A driving assistance method, characterized in that, This is a method for providing driving assistance to a target vehicle using hypothetical reasoning, the driving assistance method including: Acquire observational information, including information on one or more primary phenomena; By applying background knowledge information, including multiple pieces of evidence, to the observation information, the hypothesis is inferred, thereby inferring multiple hypotheses, each including one or more second phenomenon information. The cost of deriving each piece of information about the first phenomenon included in the observation information; and Based on the cost of each piece of first phenomenon information included in the observation information and the cost of each piece of second phenomenon information included in each of the hypotheses, the hypothesis with the lowest total cost is determined, which is the maximum likelihood hypothesis. The cost of exporting each piece of information about the first phenomenon includes: Having obtained information on the reliability of each of the first phenomena in the observation information, the cost of each of the first phenomena is derived based on the reliability information; and Without obtaining information on the reliability of each of the first phenomenon information in the observation information, the cost of each of the first phenomenon information is derived based on predetermined parameters related to the identification of the object being identified.

Images