Information processing device, information processing program, and information processing method

Abstract

To accurately predict a state of a vehicle after a predetermined time from the time indicated by a time stamp attached by a sensor that measures the state of the vehicle. [Solution] The information processing device (2) comprises an acquisition unit (231) that acquires a state value indicating the state of a target vehicle (V1) together with a timestamp indicating the time at which the state was measured, a prediction unit (232) that, when a state value is input, inputs the state value of the target vehicle into a prediction model (222) that calculates a predicted state value indicating the state after a predetermined time has elapsed from the time indicated by the timestamp corresponding to the state value, thereby obtaining the predicted state value of the target vehicle calculated by the prediction model, and an output unit (21) that outputs information based on the predicted state value of the target vehicle, and the prediction unit uses a model that takes into account errors related to the time interval of measurements by the sensor as the prediction model.

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing program, and an information processing method. [Background technology]

[0002] Patent document 1 describes a server that includes an information receiving unit that receives, from a first vehicle having a function of performing first communication with the vehicle and a function of performing second communication between the vehicle and a moving body, position information and speed information of the first vehicle, information on the detection results of a non-responsive object located on the road that does not respond to the second communication from the first vehicle, and information on the measurement results of the position and speed of the non-responsive object; a vehicle information storage unit that stores the position information and speed information of the first vehicle received from the first vehicle, information on the detection results of the non-responsive object, and information on the measurement results of the position and speed of the non-responsive object; and an information processing unit that determines whether the non-responsive object is a vehicle based on the information on the detection results of the non-responsive object, and if it is determined that the non-responsive object is a vehicle, stores the information on the measurement results of the position and speed of the non-responsive object in the vehicle information storage unit as position information and speed information of a vehicle that is not capable of communication. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-160281 Summary of the Invention [Problem to be solved by the invention]

[0004] The time required for a sensor to send its measurement results to a server (communication delay) varies. For this reason, the sensor typically attaches a timestamp to the measurement results, and the server makes predictions based on the time indicated by the timestamp. However, with conventional technology, even using a timestamp, it is difficult to accurately predict the state at the same time as the actual sensor measurement. [Means for solving the problem]

[0005] The inventors of the present disclosure have focused on the uncertainty of timestamps, such as the resolution of the timestamps and the synchronization accuracy between the sensor and the server, as the reason why the use of timestamps has a limit to the improvement of prediction accuracy. The inventors then came up with the technical idea of ​​incorporating the uncertainty of the timestamps into the process noise matrix, leading to the present disclosure.

[0006] In order to solve the above problem, an information processing device according to one aspect of the present disclosure includes an acquisition unit that acquires a state value indicating the state of a target vehicle together with a timestamp indicating the time at which the state was measured, a prediction unit that, when a state value is input, inputs the state value of the target vehicle into a prediction model that calculates a predicted state value indicating the state after a predetermined time has elapsed from the time indicated by the timestamp corresponding to the state value, thereby obtaining the predicted state value of the target vehicle calculated by the prediction model, and an output unit that outputs information based on the predicted state value of the target vehicle, wherein the prediction unit uses a model that takes into account errors related to the time interval of measurements by the sensor as the prediction model.

[0007] Furthermore, an information processing program according to another aspect of the present disclosure causes a computer to execute an acquisition process of acquiring a state value indicating a state of a target vehicle together with a timestamp indicating the time at which the state was measured, a prediction process of, upon input of a state value, inputting the state value of the target vehicle into a prediction model that calculates a predicted state value indicating the state after a predetermined time has elapsed since the time indicated by the timestamp corresponding to the state value, thereby obtaining a predicted state value of the target vehicle calculated by the prediction model, and an output process of outputting information based on the predicted state value of the target vehicle, wherein in the prediction process, a model that takes into account an error related to the time interval of measurement by the sensor is used as the prediction model. Note that a computer-readable recording medium having an information processing program recorded thereon is also within the scope of the present disclosure.

[0008] In addition, an information processing method according to another aspect of the present disclosure includes an acquisition step in which a computer acquires a state value indicating the state of a target vehicle together with a timestamp indicating the time at which the state was measured; a prediction step in which the computer inputs the state value of the target vehicle into a prediction model that, when a state value is input, calculates a predicted state value indicating the state after a predetermined time has elapsed from the time indicated by the timestamp corresponding to the state value, thereby obtaining the predicted state value of the target vehicle calculated by the prediction model; and an output step in which the computer outputs information based on the predicted state value of the target vehicle, wherein in the prediction step, the computer uses a model that takes into account errors related to the time interval of measurements by the sensor as the prediction model. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a schematic diagram illustrating an example of a schematic configuration of an information providing system according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a schematic diagram showing another example of the schematic configuration of the system. [Figure 3] FIG. 2 is a block diagram showing an example of the functional configuration of the system. [Figure 4] 10 is a flowchart illustrating an example of a flow of an information providing method according to an embodiment of another aspect of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0010] <Information Providing System 100> The information providing system 100 according to this embodiment of one aspect of the present disclosure will be described in detail below.

[0011] [Configuration of information providing system 100] The information provision system 100 is a system for providing information about a target vehicle traveling on road R to other vehicles also traveling on road R. As shown in FIG. 1 , the information provision system 100 includes a sensor 1 and an information processing device 2. The information provision system 100 according to this embodiment further includes a calculation device 3. The other vehicle is a so-called connected car configured to be able to communicate with the information processing device 2. The target vehicle may be a connected car, or may be a so-called non-connected car that does not have the function of communicating with the information processing device 2. The information provision system 100 may include a plurality of sensors 1. The information processing device 2 may be provided outside the vehicle or inside the vehicle.

[0012] [Sensor 1] The sensor 1 according to this embodiment is provided outside the target vehicle V1. The sensor 1 according to this embodiment is provided near the road R on which the target vehicle V1 travels. The vicinity of the road R includes, for example, the side of the road R, above the road R, and buildings and fences facing the road. As shown in FIG. 2, the sensor 1 may be provided as an on-board sensor in another vehicle V2 traveling on the road R on which the target vehicle V1 travels. The sensor 1 may also be provided inside the target vehicle V1. The sensor 1 repeatedly detects the state of the target vehicle V1. The sensor 1 according to this embodiment is a Light Detection and Ranging (LiDAR) that detects the distance from the sensor 1 to the target vehicle V1. That is, the sensor 1 according to this embodiment detects the distance from the sensor 1 to the target vehicle V1. The sensor 1 may also be a camera, radar, or the like that captures the target vehicle V1. When the information provision system 100 includes a plurality of sensors 1, all of the sensors 1 may detect the state of the target vehicle V1, or some of the sensors 1 may detect the state of a vehicle other than the target vehicle V1. When the sensor 1 according to this embodiment detects the state (distance) of the target vehicle V1, it transmits the detection result to the calculation device 3.

[0013] [Calculation device 3] The calculation device 3 includes a calculation unit 31. The calculation unit 31 calculates a state value indicating the state of the target vehicle V1 based on the detection result of the sensor 1. The state value is a numerical value indicating the state of the target vehicle V1. As described above, the sensor 1 detects the distance from the sensor 1 to the target vehicle V1. Therefore, the calculation unit 31 according to this embodiment calculates at least one of the position, azimuth angle, speed, and vehicle length of the target vehicle V1 as the state value based on the distance. The calculation device 3 repeatedly calculates the state value. After calculating the state value, the calculation device 3 transmits the state value to the information processing device 2. Note that the calculation unit 31 may be included in the sensor 1 or the information processing device 2. In this case, the information provision system 100 does not need to include the calculation device 3.

[0014] [Information processing device 2] As shown in FIG. 3, the information processing device 2 includes a communication unit 21, a storage unit 22, and a calculation unit .

[0015] [Communications Department 21] The communication unit 21 communicates with the calculation device 3. The communication unit 21 according to this embodiment also communicates with a vehicle having a communication function. The communication unit 21 according to this embodiment is configured with a communication module. Note that the communication unit 21 may be configured to communicate with the sensor 1.

[0016] [Storage unit 22] The storage unit 22 stores an information processing program 221. The information processing program 221 is a program that causes a computer to function as the information processing device 2. The storage unit 22 according to this embodiment stores a prediction model 222. Details of the prediction model 222 will be described later. The storage unit 22 can also store various calculation results (predicted state values, etc.) of the calculation unit 23. The storage unit 22 according to this embodiment is configured with a semiconductor memory, a hard disk drive, etc. The storage unit 22 may be physically divided into sections for storing different contents (information processing program 221, prediction model 222, calculation results).

[0017] [Calculation unit 23] The calculation unit 23 includes an acquisition unit 231 and a prediction unit 232. The calculation unit according to this embodiment further includes a generation unit 233, a determination unit 235, a correction unit 236, and an output processing unit 237. The calculation unit 23 according to this embodiment is configured with a processor (the information processing device 2 is configured with a computer). Therefore, the functions of the control blocks 231 to 237 are realized by the calculation unit 23 executing an acquisition process, a prediction process, a generation process, a determination process, and an output process in accordance with the information processing program 221 stored in the storage unit 22.

[0018] (Acquisition part 231) The acquisition unit 231 executes an acquisition process. In the acquisition process, the acquisition unit 231 acquires, from the sensor 1, a state value indicating the state of the target vehicle, along with a timestamp indicating the time at which the state was measured. The acquisition unit 231 according to this embodiment acquires a state value received by the communication unit 21 from the calculation device 3 each time the calculation device 3 transmits the state value. When multiple received state values ​​are accumulated in the queue, the acquisition unit 231 acquires the state value in order from the oldest state value indicated by the timestamp. Furthermore, the acquisition unit 231 according to this embodiment acquires a second state value from an on-board sensor provided in a vehicle traveling on the road R on which the target vehicle V1 is traveling each time the on-board sensor transmits the second state value. When multiple received second state values ​​are accumulated in the queue, the acquisition unit 231 acquires the second state value in order from the oldest state value indicated by the timestamp. The second state value is a numerical value indicating the state of the target vehicle V1. The vehicle on which the on-board sensor is provided may be the target vehicle V1 or another vehicle V2. The second state value acquired by the acquisition unit 231 according to this embodiment includes the position (coordinates) of the other vehicle V2.

[0019] (Prediction unit 232) The prediction unit 232 executes a prediction process. In the prediction process, the prediction unit 232 inputs a state value of the target vehicle V1 into the prediction model 222 and obtains a predicted state value of the target vehicle V1 calculated by the prediction model 222. The prediction unit 232 obtains a predicted state value every time the acquisition unit 231 obtains a state value.

[0020] When a state value is input, the prediction model 222 is configured to calculate a predicted state value indicating the state after a predetermined time (time step) has elapsed from the time indicated by the timestamp corresponding to the state value. The predetermined time can be, for example, the elapsed time from the time the state corresponding to the state value is measured to the time the state corresponding to the next state value is measured. In this way, by having the prediction model 222 calculate a predicted state value indicating the state after a predetermined time has elapsed from the time of measurement rather than the time of acquisition, the influence of delay between the time of measurement and the time of acquisition can be eliminated from the prediction. As a result, the generation unit 233, described later, can reproduce a digital twin with higher accuracy.

[0021] Specifically, the prediction model 222 is a model that takes into account errors related to the time intervals of measurements by the sensors. The prediction model 222 according to this embodiment is a model that takes into account errors based on the time resolution of the sensors or time synchronization errors as errors related to the time intervals. Specifically, as shown in FIG. 4, the prediction model 222 according to this embodiment has a state equation that describes the state of the target vehicle V1 and includes a term for a process noise matrix (Q') that defines the amount of process noise.

[0022] The process noise matrix (Q') can be calculated as follows: First, a vector x', which is a combination of a state vector x and a time step Δt, is set as shown in the following formula (1).

number

number

number

number

number

[0023] The prediction model 222 adds a noise component (σ Δt 2 ) This noise component is the variance (σ Δt 2 ) Specifically, the noise component (variance) according to this embodiment is expressed by the following formula (6): Here, it is assumed that Δt1 and Δt2 follow different independent probability distributions.

number

[0024] The variance is the error δ ti (i=1, 2), that is, different values ​​are defined depending on the type of sensor 1 with which the information processing device 2 is communicating. For example, the timestamp resolution (r t ) is relatively low and this causes an error, the variance is determined based on the timestamp resolution of the sensor 1. Specifically, Var(δt i ) is expressed as the following equation (7).

number

[0025] On the other hand, if the synchronization accuracy between the sensor 1 and the information processing device 2 is low (for example, there is a difference between the timers of the sensor 1 and the information processing device 2), and an error occurs due to this, the variance is the standard deviation of the time synchronization with the sensor (σ t Specifically, Var(δt i ) is expressed as the following equation (8).

number

[0026] In addition, Var(δt i ) may be the sum, maximum value, or weighted average of the value on the right side of equation (7) and the value on the right side of equation (8). The prediction model 222 may be configured to output, as a predicted state value, a numerical value indicating the state after a predetermined time has elapsed since the acquisition unit 231, which will be described later, acquired the state value. The prediction model 222 may also be stored in another storage device (not shown) that is different from the information processing device 2.

[0027] As described above, the acquisition unit 231 further acquires a second state value. Therefore, the prediction unit 232 according to this embodiment inputs a state value and a second state value associated with the state value to the prediction model 222, and obtains a predicted state value of the target vehicle V1 calculated by the prediction model 222. Note that the prediction unit 232 may be configured to input only the second state value to the prediction model 222 and obtain a predicted state value calculated by the prediction model 222.

[0028] When the acquisition unit 231 acquires a state value, the prediction unit 232 inputs the state value from one cycle before (the most recent state value) among the state values ​​held by the information processing device 2 up to that point into the prediction model 222 configured as described above, thereby obtaining a predicted state value at the time the state value was acquired. When attempting to reproduce multiple target vehicles V1, or the target vehicle V1 and another vehicle V2, in a single digital twin, the prediction unit 232 inputs the state values ​​(second state values) acquired from each vehicle V1, V2 into the prediction model to obtain a predicted state value for each vehicle V1, V2. In this case, the time of the corrected predicted state value generally differs for each vehicle. Therefore, to reproduce multiple target vehicles V1, or the target vehicle V1 and another vehicle V2, in a single digital twin, it is necessary to align the time of the predicted state value of each vehicle with the time of the digital twin. Therefore, in this case, the prediction model 222 according to this embodiment sets the predetermined time as the elapsed time from each time indicated by the state value of each vehicle to the time of the digital twin.

[0029] (Judgment Department 235) Each time the acquisition unit 231 acquires a state value, the determination unit 235 determines whether or not there is a correspondence between the newly acquired state value and the latest predicted state value calculated by the prediction model 222. The presence or absence of the correspondence includes at least one of whether or not the ID of the state value is the same as the ID of the predicted state value, and whether or not the distance (Euclidean distance, Mahalanobis distance, etc.) between the position of the target vehicle V1 indicated by the state value and the position of the target vehicle V1 indicated by the predicted state value is less than a threshold. Then, when the determination unit 235 determines that there is a correspondence between the state value and the predicted state value, it associates the new state value with the latest predicted state value. As described above, the acquisition unit 231 further acquires a second state value. Therefore, the determination unit 235 according to this embodiment also determines whether or not there is a correspondence between the newly acquired second state value and the latest predicted state value each time the acquisition unit 231 acquires a second state value. Then, when the determination unit 235 determines that there is a correspondence between the second state value and the predicted state value, it also associates the new second state value with the latest predicted state value. On the other hand, when the determination unit 235 determines that there is no correspondence between the state value (second state value) and the predicted state value, it determines that the vehicle that has transmitted the newly acquired state value is a new vehicle that has not been targeted so far.

[0030] (correction unit 236) The correction unit 236 corrects the predicted state value associated with the state value based on the state value so that the predicted state value approaches (matches) the state value. At that time, the correction unit 236 also corrects the timestamp of the predicted state value so that the timestamp approaches (matches) the timestamp of the state value. The correction unit 236 according to this embodiment performs the correction using a Kalman filter. Note that the correction unit 236 may be configured to adjust the degree of correction taking into account the reliability of the state value. The reliability can be set based on the characteristics of the sensor 1. Furthermore, the correction of the predicted state value may be performed by a device other than the information processing device 2. In this case, the calculation unit 23 does not need to include the correction unit 236.

[0031] (Generation unit 233) The generation unit 233 executes a generation process. In the generation process, the generation unit 233 generates a digital twin that reproduces, in a virtual space, the target vehicle V1 and the surrounding environment of the target vehicle V1 after a predetermined time has elapsed since the state corresponding to the state value was measured, based on the predicted state value. The generation unit 233 generates a digital twin each time the prediction model 222 generates a predicted state value. The generation unit 233 according to this embodiment generates a digital twin based on the predicted state values ​​generated by the prediction model 222 and corrected by the correction unit 236. Note that the generation of the digital twin may be performed by a device other than the information processing device 2. In this case, the calculation unit 23 does not need to include the generation unit 233.

[0032] (output processing unit 237) The output processing unit 237 executes output processing. In the output processing, the output processing unit 237 according to the present embodiment controls the communication unit 21. As a result, the communication unit 21 transmits (outputs) information based on the predicted state values ​​of the target vehicle V1 to at least another vehicle V2. That is, the communication unit 21 according to the present embodiment also functions as an output unit. The information based on the predicted state values ​​includes the digital twin generated by the generation unit 233. Note that the output processing unit 237 may be configured to transmit the information based on the predicted state values ​​to something other than the other vehicle V2 (such as the target vehicle V1 or a remote monitoring device). Furthermore, when the information processing device 2 is provided inside a vehicle, the output processing unit 237 may be configured to supply the information based on the predicted state values ​​to a display unit or the like of the vehicle. In this case, a terminal or the like of the information processing device connected to the vehicle serves as the output unit.

[0033] [Operational Effects of Information Processing Device 2 (Information Providing System 100)] In the information processing device 2 (information provision system 100) described above, the prediction unit 232 uses, as the prediction model 222, a model that takes into account errors related to the time intervals of measurements by the sensor 1. As a result, the predicted state value absorbs errors that occur between the timestamp and the actual measurement time. Therefore, the information processing device 2 (information provision system 100) can accurately predict the state a predetermined time after the time indicated by the timestamp attached by the sensor 1 that measures the state of the target vehicle V1. Note that the problem of errors occurring between the timestamp and the actual measurement time is particularly likely to occur when multiple sensors 1 are present. Therefore, the information processing device 2 (information provision system 100) can further achieve the above-mentioned effects when multiple sensors 1 are present. Furthermore, the information processing device 2 (information provision system 100) can make more accurate predictions, particularly when the error in the timestamp is large compared to the traveling speed of the target vehicle V1.

[0034] Use of this disclosure can realize safer autonomous driving technology (autonomous vehicles, control devices, remote monitoring devices, etc.), which can contribute to achieving Sustainable Development Goals (SDGs), such as Goal 9 "Build resilient infrastructure, promote inclusive and sustainable industrialization, promote innovation and build resilient infrastructure," and Goal 11 "Make cities and towns inclusive and sustainable."

[0035] <Information processing method S100> Next, an information processing method S100 according to an embodiment of another aspect of the present disclosure will be described in detail.

[0036] [Flow of information processing method S100] 4, the information processing method S100 includes an acquisition step S1, a prediction step S2, and an output step S3. The information processing method S100 according to this embodiment further includes a generation step S4, a first determination step S6, a correction step S7, and a second determination step S8.

[0037] [Acquisition step S1] In the first acquisition step S1, a computer acquires a state value indicating the state of the target vehicle together with a timestamp indicating the time when the state was measured. In the acquisition step S1 according to this embodiment, the computer further acquires a second state value from an on-board sensor provided on a vehicle traveling on the road R on which the target vehicle V1 is traveling. The computer that acquires the state value may be the information processing device 2 or another device.

[0038] [Prediction step S2] After acquiring the state values, the process proceeds to the prediction step S2. In the prediction step S2, the computer inputs the state values ​​of the target vehicle V1 into the prediction model 222 to obtain a predicted state value of the target vehicle V1 calculated by the prediction model 222. In the prediction step S2, the computer uses, as the prediction model 222, a model that takes into account errors related to the time intervals of measurements by the sensors. As described above, in the acquisition step S1, a second state value is further acquired. Therefore, in the prediction step S2 according to this embodiment, the computer inputs the state values ​​and the second state values ​​associated with the state values ​​into the prediction model 222 to obtain a predicted state value of the target vehicle V1 calculated by the prediction model 222. The computer that acquires the predicted state value may be the information processing device 2 or another device.

[0039] [First decision step S6] After acquiring the state value, the process proceeds to the first determination step S6. In the first determination step S6, each time the computer acquires a state value in the acquisition step S1, the computer determines whether or not there is a correspondence between the newly acquired state value and the latest predicted state value calculated by the prediction model 222. If it is determined that there is a correspondence between the state value and the predicted state value (step S6: YES), the computer associates the new state value with the latest predicted state value. As described above, in the acquisition step S1, the computer further acquires a second state value. Therefore, in the first determination step S6 according to this embodiment, each time the computer acquires a second state value in the acquisition step S1, the computer also determines whether or not there is a correspondence between the newly acquired second state value and the latest predicted state value. If it is determined that there is a correspondence between the second state value and the predicted state value, the computer also associates the new second state value with the latest predicted state value. The computer that performs the association may be the information processing device 2 or another device.

[0040] On the other hand, if it is determined in the first determination step S6 that there is no correspondence between the newly acquired state value and the latest predicted state value (the newly acquired state value cannot be associated with the latest predicted state value) (step S6: NO), the computer determines that the vehicle that has transmitted the newly acquired state value is a new vehicle that has not been considered as a target vehicle until now. In this case, the computer starts the information processing method (steps S1 to S8) for the new vehicle as the target vehicle V1. Note that, after determining in the first determination step S6 that there is no correspondence between the state value and the predicted state value, it may further determine whether the determination that there is no correspondence has been repeated a predetermined number of times. Then, if it is determined that the determination that there is no correspondence has been repeated a predetermined number of times, the computer may determine that the vehicle that has transmitted the newly acquired state value is a new vehicle that has not been considered as a target vehicle until now.

[0041] [Correction step S7] After associating the state value with the predicted state value, the process proceeds to a correction step S7. In the correction step S7, the computer corrects the predicted state value associated with the state value based on the state value so that the predicted state value approaches (matches) the state value. At this time, in the correction step S7, the timestamp of the predicted state value is also corrected so that the timestamp approaches (matches) the timestamp of the state value. The computer that corrects the predicted state value may be the information processing device 2 or another device.

[0042] [Generation step S4] After obtaining the predicted state values, the process proceeds to generation step S4. In generation step S4, the computer generates a digital twin in a virtual space based on the predicted state values, which reproduces the target vehicle V1 and the surrounding environment of the target vehicle V1 after a predetermined time has elapsed since the state corresponding to the state value was measured. In generation step S4, a digital twin is generated each time the prediction model 222 generates a predicted state value. In generation step S4 according to this embodiment, a digital twin is generated based on the predicted state values ​​generated by the prediction model 222 and corrected in correction step S7. The computer that generates the digital twin may be the information processing device 2 described above, or may be another device.

[0043] [Output step S3] After obtaining the predicted state values, the process proceeds to output step S3. In output step S3, the computer outputs information based on the predicted state values ​​of the target vehicle V1. In output step S3 according to this embodiment, the computer transmits information based on the predicted state values ​​of the target vehicle V1 (such as a digital twin or information indicating that action to avoid a rear-end collision will be taken) to at least another vehicle V2. The computer that outputs the information based on the predicted state values ​​may be the information processing device 2 or another device. Note that in output step S3, the information based on the predicted state values ​​may be output to something other than the other vehicle V2 (such as the target vehicle V1 or a remote monitoring device).

[0044] [Second decision step S8] After outputting the information based on the predicted state values, the process proceeds to a second determination step S8. In the second determination step S8, the computer determines whether or not it is necessary to continue generating the predicted state values ​​of the target vehicle V1. If it is determined that it is not necessary to continue (step S8: NO), the information processing method S100 ends. On the other hand, if it is determined that it is necessary to continue (step S8: YES), the process returns to step S1. The computer that determines whether or not to continue may be the information processing device 2 or another device.

[0045] [Effects of information processing method S100] According to the information processing method S100 described above, similar to the information processing device 2 (information provision system 100), it is possible to accurately predict the state of the target vehicle V1 a predetermined time after the time indicated by the timestamp attached by the sensor 1 that measures the state of the target vehicle V1.

[0046] <Modification> The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present disclosure.

[0047] For example, the predicted state value generated by the prediction unit 232 may include acceleration. In this case, the calculation unit 23 may include another determination unit that executes a determination process when there is another vehicle V2 traveling ahead of the target vehicle V1 in the same direction as the target vehicle V1. In the determination process, the other determination unit determines whether or not there is a possibility that the target vehicle V1 will collide with the other vehicle V2 from the rear, based on the acceleration (predicted state value). Then, when the other determination unit determines that there is a possibility that the target vehicle V1 will collide with the other vehicle V2 from the rear, the output unit may output information that indicates that action to avoid the collision will be taken.

[0048] Furthermore, the information processing program may be stored not temporarily but on one or more computer-readable storage media. Each unit may or may not have a storage medium. In the latter case, the information processing program may be supplied to each unit via any wired or wireless transmission medium.

[0049] Furthermore, some or all of the functions of the control blocks 231 to 237 included in the arithmetic unit can be realized by logic circuits. For example, the scope of the present disclosure also includes integrated circuits in which logic circuits that function as the control blocks 231 to 235 are formed. In addition, the functions of each unit can also be realized by, for example, a quantum computer.

[0050] <Summary> An information processing device according to aspect 1 of the present disclosure includes an acquisition unit that acquires a state value indicating the state of a target vehicle together with a timestamp indicating the time at which the state was measured; a prediction unit that, when a state value is input, inputs the state value of the target vehicle into a prediction model that calculates a predicted state value indicating the state after a predetermined time has elapsed from the time indicated by the timestamp corresponding to the state value, thereby obtaining the predicted state value of the target vehicle calculated by the prediction model; and an output unit that outputs information based on the predicted state value of the target vehicle, wherein the prediction unit uses a model that takes into account errors related to the time interval of measurements by the sensor as the prediction model.

[0051] An information processing device according to aspect 2 of the present disclosure may be configured such that, in aspect 1 above, the prediction unit uses, as the prediction model, a state equation that describes the state of the target vehicle, the state equation including a term of a process noise matrix that defines the amount of process noise, and the components of the process noise matrix include noise components that represent the error.

[0052] An information processing device according to aspect 3 of the present disclosure may be configured in the above aspect 2 such that the noise component is a variance of the predetermined time and is defined based on the timestamp resolution of the sensor.

[0053] An information processing device according to aspect 4 of the present disclosure may be configured in the above aspect 2 such that the noise component is a variance over the specified time period and is defined based on the standard deviation of time synchronization with the sensor.

[0054] An information processing device according to aspect 5 of the present disclosure may be configured such that, in any of aspects 1 to 4 above, the acquisition unit acquires a second state value indicating the state of the target vehicle from an onboard sensor provided on a vehicle traveling on the road on which the target vehicle is traveling, and the prediction unit inputs the state value and the second state value into the prediction model to obtain a predicted state value of the target vehicle calculated by the prediction model.

[0055] An information processing program according to aspect 6 of the present disclosure causes a computer to execute an acquisition process that acquires a state value indicating the state of a target vehicle together with a timestamp indicating the time at which the state was measured; a prediction process that, when a state value is input, inputs the state value of the target vehicle into a prediction model that calculates a predicted state value indicating the state after a predetermined time has elapsed from the time indicated by the timestamp corresponding to the state value, thereby obtaining the predicted state value of the target vehicle calculated by the prediction model; and an output process that outputs information based on the predicted state value of the target vehicle, wherein in the prediction process, a model that takes into account errors related to the time interval of measurements by the sensor is used as the prediction model.

[0056] An information processing method according to aspect 7 of the present disclosure includes an acquisition step in which a computer acquires a state value indicating the state of a target vehicle together with a timestamp indicating the time at which the state was measured; a prediction step in which the computer inputs the state value of the target vehicle into a prediction model that calculates a predicted state value indicating the state after a predetermined time has elapsed since the time indicated by the timestamp corresponding to the state value, thereby obtaining the predicted state value of the target vehicle calculated by the prediction model; and an output step in which the computer outputs information based on the predicted state value of the target vehicle, wherein in the prediction step, the computer uses a model that takes into account errors related to the time interval of measurements by the sensor as the prediction model. [Explanation of symbols]

[0057] 100 Information Provision System 1 sensor 2. Information processing equipment 21 Communication unit (output unit) 22 Memory section 221 Information Processing Program 222 Prediction Model 23 Arithmetic section 231 Acquisition Department 232 Prediction Department 233 Generation part 235 Judgment Department 236 Correction Unit 237 Output Processing Unit 3. Calculation device 31 Calculation section S100 Information processing method S1 Acquisition step S2 Prediction step S3 Output Step S4 Generation Step S6 First decision step S7 Correction step S8 Second decision step V1 Target vehicle V2 Other vehicles

Claims

1. an acquisition unit that acquires a state value indicating a state of a target vehicle together with a timestamp indicating a time when the state was measured; a prediction unit that, when a state value is input, inputs the state value of the target vehicle into a prediction model that calculates a predicted state value indicating a state after a predetermined time has elapsed from the time indicated by the time stamp corresponding to the state value, and obtains the predicted state value of the target vehicle calculated by the prediction model; an output unit that outputs information based on the predicted state value of the target vehicle; Equipped with The prediction unit may include, as the prediction model: a state equation describing a state of the target vehicle, the state equation including a process noise matrix term that defines an amount of process noise; The components of the process noise matrix include a noise component representing an error related to a time interval of measurement by a sensor. Using a model, Information processing device.

2. The noise component is the predetermined time variance, defined based on the timestamp resolution of the sensor; The information processing device according to claim 1 .

3. The noise component is the predetermined time variance, It is determined based on the standard deviation of time synchronization with the sensor. The information processing device according to claim 1 .

4. the acquisition unit acquires a second state value indicating a state of the target vehicle from an on-board sensor provided in a vehicle traveling on a road on which the target vehicle is traveling; the prediction unit inputs the state value and the second state value into the prediction model to obtain a predicted state value of the target vehicle calculated by the prediction model; The information processing device according to claim 1 .

5. On the computer, an acquisition process for acquiring a state value indicating a state of the target vehicle together with a timestamp indicating the time when the state was measured; a prediction process in which, when a state value is input, the state value of the target vehicle is input into a prediction model that calculates a predicted state value indicating a state after a predetermined time has elapsed from the time indicated by the timestamp corresponding to the state value, and the predicted state value of the target vehicle calculated by the prediction model is obtained; an output process for outputting information based on the predicted state value of the target vehicle; Execute In the prediction process, the prediction model is a state equation describing a state of the target vehicle, the state equation including a process noise matrix term that defines an amount of process noise; The components of the process noise matrix include a noise component representing an error related to a time interval of measurement by a sensor. Using a model, Information processing program.

6. an acquisition step in which the computer acquires a state value indicating a state of the target vehicle together with a timestamp indicating the time when the state was measured; a prediction step in which, when a state value is input, the computer inputs the state value of the target vehicle into a prediction model that calculates a predicted state value indicating a state after a predetermined time has elapsed from the time indicated by the timestamp corresponding to the state value, thereby obtaining a predicted state value of the target vehicle calculated by the prediction model; an output step in which a computer outputs information based on the predicted state value of the target vehicle; Including, In the prediction step, the computer sets the prediction model as a state equation describing a state of the target vehicle, the state equation including a process noise matrix term that defines an amount of process noise; The components of the process noise matrix include a noise component representing an error related to a time interval of measurement by a sensor. Using a model, Information processing methods.

Images