Method for developing automated driving vehicle

Abstract

To reduce the effort of re-adapting a second automated driving vehicle having a second sensor specification inferior to a first sensor specification in at least one of the number, a type, and a detection performance, separately from a first automated driving vehicle.SOLUTION: In an automated driving system, out of recognition processing for recognizing an external environment, planning processing for performing route planning, and vehicle control processing for performing automated driving control, a machine learning model is applied to at least the recognition processing and the planning processing. A method for developing an automated driving vehicle includes: a first step of training the machine learning model by adapting at least the recognition processing and the planning processing, using a first automated driving vehicle including an external sensor having a first sensor specification; and a second step of applying the machine learning model trained in the first step to a second automated driving vehicle including an external sensor having a second sensor specification inferior to the first sensor specification in at least one of the number, a type, and a detection performance.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a method for developing an autonomous vehicle. [Background technology]

[0002] The detection system described in the patent document 1 has an analysis module in the form of a convolutional neural network. Images from the camera and sensor data from the sensor device are used as input data for the analysis module and are analyzed in real time by the detection system. As a result of the analysis, for example, the movement of the body of a first vehicle running in front of a second vehicle indicates the safety level of the second vehicle, and optical and / or acoustic warnings are output on a user interface of the second vehicle depending on the safety level. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2022-082510 Summary of the Invention [Problem to be solved by the invention]

[0004] In autonomous vehicles that recognize the external environment based on the detection results of external sensors, plan a route, and perform autonomous driving control based on the planned route, the input data for the route plan changes depending on, for example, the number, type, and detection performance of the external sensors. Therefore, conventionally, the route plan adaptation results could not be simply shared between autonomous vehicles with different external sensor configurations, and it was necessary to re-adapt the route plan separately to match the different input data, which required a huge amount of development work.

[0005] The present invention aims to reduce the effort required to separately refit a second autonomous vehicle having a second sensor specification in which at least one of the number, type, and detection performance is inferior to the first sensor specification from the first autonomous vehicle. [Means for solving the problem]

[0006] One aspect of the present invention is a method for developing an autonomous vehicle equipped with external sensors whose sensor specifications include number, type, and detection performance, wherein the autonomous vehicle is equipped with an autonomous driving system to which a machine learning model is applied that uses the detection results of the external sensors as input and outputs a route for at least the recognition and planning processes out of a recognition process that recognizes the external environment based on the detection results of the external sensors, a planning process that plans a route using the results of the recognition process, and a vehicle control process that performs autonomous driving control based on the planned route, and the method comprises a first step of training the machine learning model by adapting at least the recognition and planning processes using a first autonomous vehicle equipped with external sensors of a first sensor specification, and a second step of applying the machine learning model trained in the first step to a second autonomous vehicle equipped with external sensors of a second sensor specification in which at least one of the number, type, and detection performance is inferior to the first sensor specification.

[0007] In a method for developing an autonomous vehicle according to one aspect of the present invention, in a first step, a machine learning model is trained using a first autonomous vehicle having a sensor specification superior to that of a second autonomous vehicle. In a second step, the machine learning model trained in the first autonomous vehicle is applied to the second autonomous vehicle. This allows the second autonomous vehicle to output a route equivalent to that of the first autonomous vehicle using the trained machine learning model. This makes it possible to reduce the effort required to re-adapt the second autonomous vehicle having a second sensor specification in which at least one of the number, type, and detection performance is inferior to the first sensor specification, separately from the first autonomous vehicle.

[0008] In a method for developing an autonomous vehicle according to an embodiment, an autonomous driving system applies a machine learning model that uses detection results from an external sensor as input and outputs a control amount of autonomous driving control for recognition processing, planning processing, and vehicle control processing, and in a first step, a first autonomous driving vehicle may be used to train the machine learning model by adapting the recognition processing, planning processing, and vehicle control processing. In this case, the machine learning model including the vehicle control processing is trained on the first autonomous driving vehicle and then applied to the second autonomous driving vehicle. This allows the second autonomous driving vehicle to output a control amount of autonomous driving control equivalent to that of the first autonomous driving vehicle using the trained machine learning model. Effect of the Invention

[0009] According to the present invention, it is possible to reduce the effort required to separately refit a second autonomous vehicle having a second sensor specification in which at least one of the number, type, and detection performance is inferior to the first sensor specification from the first autonomous vehicle. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram illustrating an autonomous vehicle used to train a machine learning model in a method for developing an autonomous vehicle according to an embodiment. [Diagram 2] FIG. 1 is a block diagram illustrating an autonomous vehicle to which a trained machine learning model is applied in a method for developing an autonomous vehicle according to an embodiment. [Diagram 3] 1 is a flowchart illustrating a procedure of a development method for an autonomous driving vehicle according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, exemplary embodiments will be described with reference to the drawings. In the following description, the same or corresponding elements are denoted by the same reference numerals, and duplicated description may be omitted.

[0012] 1 is a block diagram illustrating an autonomous vehicle used for training a machine learning model in a method for developing an autonomous vehicle according to an embodiment. As shown in FIG. 1, the autonomous vehicle 10 is a vehicle such as a passenger car, and includes an autonomous driving system 10a that drives the autonomous vehicle 10 autonomously. The autonomous driving system 10a may be mounted on the autonomous vehicle 10, or may have a mode in which some of its functions are executed by a server that can communicate with the autonomous vehicle 10.

[0013] The autonomous driving system 10a of the autonomous driving vehicle 10 includes a first external sensor (external sensor) 11, an internal sensor 12, a map database 13, a GNSS receiver 14, an actuator 15, and an autonomous driving ECU [Electronic Control Unit] 100.

[0014] The first external sensor 11 is a detection device that detects the surrounding situation (external environment) of the autonomously driven vehicle 10. The sensor specifications (first sensor specifications) of the first external sensor 11 are, for example, the highest level of sensor specifications of the manufacturer of the autonomously driven vehicle 10. The sensor specifications include the number, type, and detection performance of the external sensors.

[0015] The first external sensor 11 includes at least a group of external cameras 11a that capture images of the area ahead of the autonomous vehicle 10. The first external sensor 11 may include a radar sensor. The first external sensor 11 may be configured to reconstruct various properties of the external environment in which the autonomous vehicle 10 is traveling (such as the position of the vehicle, the relative distance to other vehicles, the relative speed to other vehicles, the orientation of other vehicles, the shape of the lanes, and the lighting status of traffic lights).

[0016] The external camera group 11a is an imaging device that captures an image of the external environment of the autonomous vehicle 10. The external camera group 11a includes a plurality of cameras of different types and detection performance. Examples of the types of the external camera group 11a include a telephoto camera, a wide-angle camera, a monocular camera, and a stereo camera. Examples of the detection performance of the external camera group 11a include resolution and nighttime image quality. The stereo camera has two imaging units arranged to reproduce binocular parallax. The imaging information of the stereo camera also includes information in the depth direction. The external camera group 11a transmits imaging information related to the external environment of the autonomous vehicle 10 to the autonomous driving ECU 100.

[0017] The external camera group 11a includes, for example, a forward camera provided on the back side of the windshield of the autonomously driven vehicle 10. The external camera group 11a may also include a side camera provided on the door mirror of the autonomously driven vehicle 10. The external camera group 11a may also include a rear camera provided on the rear glass of the autonomously driven vehicle 10. The external camera group 11a may also include a camera that captures images of the external environment of the autonomously driven vehicle 10 for use in autonomous driving.

[0018] A radar sensor is a detection device that uses radio waves (e.g., millimeter waves) or light to detect objects around the autonomous vehicle 10. The radar sensor detects objects by transmitting radio waves or light to the periphery of the autonomous vehicle 10 and receiving the radio waves or light reflected by the objects. The radar sensor transmits information about the detected objects to the autonomous driving ECU 100.

[0019] The radar sensor includes a plurality of radar sensors with different types and detection capabilities. Examples of the types of radar sensors include LiDAR [Light Detection And Ranging] and millimeter wave radar. Examples of the detection capabilities of the radar sensor include resolution and measurable distance. The radar sensors here include, for example, LiDAR 11b and millimeter wave radar 11c.

[0020] The internal sensor 12 is a detection device that detects the traveling state of the autonomous driving vehicle 10. The internal sensor 12 includes a vehicle speed sensor. The vehicle speed sensor is a detector that detects the speed of the autonomous driving vehicle 10. As the vehicle speed sensor, for example, a known wheel speed sensor may be used. The vehicle speed sensor transmits the detected vehicle speed information to the autonomous driving ECU 100.

[0021] The internal sensor 12 may include a known acceleration sensor and a yaw rate sensor. The acceleration sensor is a detector that detects the acceleration of the autonomously driven vehicle 10. The acceleration sensor transmits acceleration information of the autonomously driven vehicle 10 to the autonomously driven ECU 100. The yaw rate sensor is a detector (e.g., a gyro sensor) that detects a yaw rate (rotational angular velocity) around the vertical axis of the center of gravity of the autonomously driven vehicle 10. The yaw rate sensor transmits the detected yaw rate information of the vehicle to the autonomously driven ECU 100.

[0022] The map database 13 is a storage device that stores map information. The map database 13 is formed in a storage medium such as a hard disk drive (HDD) mounted on the autonomous vehicle 10. The map information includes information such as the positions and shapes of roads, intersections, and branching points (for example, the positions and shapes of intersections, the widths of lanes, the positions of division lines and crosswalks, etc.). The map information may also include information on structures (such as the positions and types of traffic lights installed at intersections) and various types of traffic rule information (such as one-way streets, no right or left turns, and no turns). Note that a part of the map information included in the map database 13 may be stored in a storage device other than the HDD. The map database 13 may be formed in a computer in a facility such as a management center that can communicate with the autonomous vehicle 10.

[0023] The GNSS receiver 14 measures the position of the vehicle (for example, the latitude and longitude of the vehicle) by receiving signals from positioning satellites. The GNSS receiver 14 transmits the measured position information of the vehicle to the autonomous driving ECU 100.

[0024] The actuator 15 is a device used for automatic driving control of the automatic driving vehicle 10. The actuator 15 includes at least a drive actuator, a brake actuator, and a steering actuator. The drive actuator controls the amount of air supplied to the engine (throttle opening) in response to a control signal from the automatic driving ECU 100, for example, to control the driving force of the automatic driving vehicle 10. When the automatic driving vehicle 10 is a hybrid vehicle, in addition to the amount of air supplied to the engine, a control signal from the automatic driving ECU 100 is input to a motor serving as a power source to control the driving force. When the automatic driving vehicle 10 is an electric vehicle, a control signal from the automatic driving ECU 100 is input to a motor serving as a power source to control the driving force. The motor serving as a power source in these cases constitutes the actuator 15.

[0025] The brake actuator controls a brake system in response to a control signal from the autonomous driving ECU 100, and controls the braking force applied to the wheels of the autonomous driving vehicle 10. For example, a hydraulic brake system can be used as the brake system. The steering actuator controls the drive of an assist motor that controls the steering torque of the electric power steering system in response to a control signal from the autonomous driving ECU 100. As a result, the steering actuator controls the steering torque in response to a control signal from the autonomous driving ECU 100.

[0026] The autonomous driving ECU 100 is a main component of the autonomous driving system 10a. The autonomous driving ECU 100 executes autonomous driving control of the autonomous driving vehicle 10. The autonomous driving ECU 100 is an electronic control unit having a CPU [Central Processing Unit], a ROM [Read Only Memory], a RAM [Random Access Memory], a CAN [Controller Area Network] communication circuit, and the like. The autonomous driving ECU 100 is connected to a network that communicates using, for example, a CAN communication circuit, and is communicatively connected to the components of the autonomous driving vehicle 10 described above. The autonomous driving ECU 100, for example, operates the CAN communication circuit based on a signal output by the CPU to input and output data, stores the data in the RAM, loads a program stored in the ROM into the RAM, and executes the program loaded into the RAM by the CPU, thereby realizing an autonomous driving function. The autonomous driving ECU 100 may be composed of multiple electronic units.

[0027] Next, a description will be given of an example of a functional configuration of the autonomous driving ECU 100. The autonomous driving ECU 100 has a recognition processing unit 110, a planning processing unit 120, and a vehicle control processing unit .

[0028] The autonomous driving system 10a is configured to perform a recognition process that recognizes the external environment based on the detection results of external sensors installed in the autonomous driving vehicle 10, a planning process that plans a route using the results of the recognition process, and a vehicle control process that performs autonomous driving control based on the planned route.

[0029] In the autonomous driving system 10a, as an example, a machine learning model is applied to each of the recognition process, the planning process, and the vehicle control process. The machine learning models here include a first recognition model 111 of the recognition processing unit 110, a first planning model 121 of the planning processing unit 120, and a first vehicle control model 131 of the vehicle control processing unit 130.

[0030] The recognition processing unit 110 performs recognition processing based on the detection result of the first external sensor 11 using the first recognition model 111. The first recognition model 111 is, for example, a machine learning model that has been trained by deep learning to recognize objects (e.g., other vehicles and traffic lights) from images captured by the external camera group 11a, and a machine learning model that has been trained (e.g., deep learning) to recognize objects (e.g., lane lines) from detection results of a radar sensor. The neural network that constitutes the first recognition model 111 is, for example, a convolutional neural network [CNN: Convolutional neural network] including multiple layers including multiple convolution layers and pooling layers. The neural network may further be configured as a recurrent neural network [RNN: Recurrent neural network]. The recognition processing unit 110 inputs the detection result of the first external sensor 11 to the first recognition model 111 to obtain the recognition result of the object output from the first recognition model 111.

[0031] The recognition processing unit 110 recognizes the external environment of the autonomously driven vehicle 10 based on the object recognition results output from the first recognition model 111. The external environment includes the relative positions of surrounding objects with respect to the autonomously driven vehicle 10. The external environment may also include the relative speeds and movement directions of the surrounding objects with respect to the autonomously driven vehicle 10.

[0032] The recognition processing unit 110 recognizes the driving state of the autonomously driven vehicle 10 based on the detection result of the internal sensor 12. The driving state includes, for example, the vehicle speed of the autonomously driven vehicle 10, the acceleration of the autonomously driven vehicle 10, and the yaw rate of the autonomously driven vehicle 10. The recognition processing unit 110 acquires the road environment (for example, road shape, intersection shape, type of traffic light, etc.) around the autonomously driven vehicle 10 based on, for example, map information.

[0033] The recognition processing unit 110 acquires the current vehicle position of the autonomously driven vehicle 10, for example, based on position information from the GNSS receiving unit 14. The recognition processing unit 110 may acquire the current vehicle position of the autonomously driven vehicle 10 by SLAM technology. The recognition processing unit 110 may also acquire the current vehicle position of the autonomously driven vehicle 10 by other known methods.

[0034] The planning processing unit 120 generates a route plan using the first planning model 121. The first planning model 121 is a machine learning model in which machine learning (e.g., deep learning) is performed to output a route plan based on, for example, the object recognition result output from the first recognition model 111, the target route, the current vehicle position of the autonomous vehicle 10, and the traveling state of the autonomous vehicle 10. The neural network constituting the first planning model 121 may be a convolutional neural network or a recurrent neural network. The neural network constituting the first planning model 121 may be of the same type as the first recognition model 111, or may be of a different type from the first recognition model 111.

[0035] A route plan is a plan regarding a route along which the autonomous vehicle 10 will travel. The route plan is used during autonomous driving of the autonomous vehicle 10. The route plan may be continuous coordinate data on a map, or may be data including a target lateral position (position of a control target in the width direction of the traveling lane) relative to a target longitudinal position of the traveling lane (position of a control target in the extension direction of the traveling lane), or may be data including a target steering angle relative to the target longitudinal position of the traveling lane. The route plan may be time series data of a target lateral position relative to the lane or time series data of a target steering angle. A target steering torque may be used instead of the target steering angle.

[0036] The vehicle control processing unit 130 may perform vehicle control using a first vehicle control model 131. The first vehicle control model 131 is a machine learning model in which machine learning (e.g., deep learning) is performed to output a control amount of automatic driving control for automatically driving the automatic driving vehicle 10 according to the route plan output from the first planning model 121 and the driving state of the automatic driving vehicle 10. The neural network constituting the first vehicle control model 131 may be a convolutional neural network or a recurrent neural network. The neural network constituting the first vehicle control model 131 may be of the same type as at least one of the first recognition model 111 and the first planning model 121, or may be of a different type from the first recognition model 111 and the first planning model 121. The control amount of the automatic driving control may be, for example, data on the steering angle and vehicle speed of the automatic driving vehicle 10 set at predetermined intervals on the target route (steering angle plan and vehicle speed plan). The vehicle control processing unit 130 controls the actuator 15 so as to follow the route plan output from the first planning model 121, and executes autonomous driving of the autonomous vehicle 10.

[0037] In general, the lower the support level of autonomous driving control, the more likely it is that inexpensive external sensors will be used. Conversely, the higher the support level, the more expensive external sensors will be used or the number of external sensors will tend to increase. In this way, when the configuration of external sensors differs, it has not always been possible to reuse a system with a high support level. Therefore, even if an autonomous driving system with a high support level has been developed in advance, an autonomous driving system with a low support level must be developed separately, which can result in increased development costs.

[0038] Therefore, in the development method for an autonomous vehicle according to this embodiment, the first recognition model 111, the first planning model 121, and the first vehicle control model 131, which have been trained using an autonomous vehicle 10 equipped with a first external sensor 11, are capable of being applied to an autonomous vehicle 20 equipped with a second external sensor 21.

[0039] Fig. 2 is a block diagram illustrating an autonomous vehicle to which a trained machine learning model is applied in a method for developing an autonomous vehicle according to an embodiment. As shown in Fig. 2, the autonomous vehicle 20 has a similar configuration to the autonomous vehicle 10. The autonomous vehicle 20 is a vehicle such as a passenger car, and includes an autonomous driving system 20a that drives the autonomous vehicle 20 in an autonomous manner.

[0040] The autonomous driving system 20a of the autonomous driving vehicle 20 includes a second external sensor (external sensor) 21, an internal sensor 22, a map database 23, a GNSS receiver 24, an actuator 25, and an autonomous driving ECU200.

[0041] The sensor specifications (second sensor specifications) of the autonomous driving system 20a are inferior to the first sensor specifications in at least one of the number, type, and detection performance. "Inferior" in terms of number means a smaller number. "Inferior" in terms of type means a simpler structure among external sensors having different structures. "Inferior" in terms of detection performance means, for example, that detection performance is lower in external sensors having a common structure. "Inferior" in terms of detection performance may also mean that detection performance is lower in external sensors having different structures.

[0042] The second external sensor 21 has, for example, an external camera 21a and a millimeter wave radar 21c. The external camera 21a may have a smaller number of sensors than the external camera group 11a, for example, by omitting some cameras from the external camera group 11a. In this case, the external camera 21a may have the same sensor specifications as some of the cameras included in the external camera group 11a.

[0043] The LiDAR is omitted from the radar sensor in the second external sensor 21. The millimeter wave radar 21c may have the same number, type, and detection performance as the millimeter wave radar 11c.

[0044] In the autonomous driving system 20a, components other than the second external sensor 21 (the internal sensor 22, the map database 23, the GNSS receiver 24, and the actuator 25) may basically be configured in the same way.

[0045] In the autonomous driving system 20a, the autonomous driving ECU 200 has a recognition processing unit 210, a planning processing unit 220, and a vehicle control processing unit 230. In the autonomous driving system 20a, for example, the second recognition model 211, the second planning model 221, and the second vehicle control model 231 can be regarded as an integrated machine learning model unit that receives the detection result of the second external sensor 21 as an input, outputs a route, and controls the vehicle.

[0046] As the second recognition model 211 of the recognition processing unit 210, the first recognition model 111 trained in the automatic driving system 10a is applied. In this case, instead of a camera having the same specifications as the external camera 21a among the cameras included in the external camera group 11a, the external camera 21a can be used as an input of the first recognition model 111 trained in the automatic driving system 10a. In addition, it is possible that there is no input from the LiDAR as an input of the first recognition model 111 trained in the automatic driving system 10a.

[0047] The recognition processing unit 210 inputs the detection result of the second external sensor 21 to the second recognition model 211, thereby acquiring the recognition result of the object output from the second recognition model 211. This acquired result is the same as the result of inputting the detection result of the second external sensor 21 to the first recognition model 111, so the effort of adaptation in the second recognition model 211 is eliminated.

[0048] The first planning model 121 trained in the autonomous driving system 10a is applied as the second planning model 221 of the planning processing unit 220. Differences in the specifications of the second external sensor 21 are adjusted when inputting them to the second recognition model 211, so the second planning model 221 can output a route plan for the autonomous driving vehicle 20 based on the object recognition results output from the second recognition model 211, the target route, the current vehicle position of the autonomous driving vehicle 20, and the driving state of the autonomous driving vehicle 20.

[0049] The vehicle control processing unit 230 may perform vehicle control using a second vehicle control model 231. The second vehicle control model 231 may be the first vehicle control model 131 that has been trained in the autonomous driving system 10a. The vehicle control processing unit 230 automatically drives the autonomous vehicle 20 according to the route plan of the autonomous vehicle 20, for example, based on the route plan output from the second planning model 221 and the driving state of the autonomous vehicle 20.

[0050] Next, the procedure of a method for developing an autonomous vehicle will be described with reference to Fig. 3. Fig. 3 is a flowchart illustrating the procedure of a method for developing an autonomous vehicle according to an embodiment. The flowchart shown in Fig. 3 can be executed, for example, in a case where an autonomous vehicle 10 is developed first, and an autonomous vehicle 20 is developed in parallel with or after the development of the autonomous vehicle 10.

[0051] 3, in the method for developing an autonomous vehicle according to the embodiment, in S1 (first step), a machine learning model is trained by adapting at least the recognition process and the planning process using a first autonomous vehicle equipped with an external sensor of a first sensor specification. In this embodiment, for example, a first recognition model 111 and a first planning model 121 are trained by adapting the recognition process and the planning process using an autonomous vehicle 10 equipped with a first external sensor 11 having an external camera group 11a, a LiDAR 11b, and a millimeter wave radar 11c.

[0052] In S2 (second step), the machine learning model trained in the first step is applied to a second autonomous vehicle equipped with an external sensor of a second sensor specification. In this embodiment, for example, the first recognition model 111 and the first planning model 121 trained in step S1 are applied to an autonomous vehicle 20 equipped with a second external sensor 21 having an external camera 21a and a millimeter wave radar 21c. Then, the procedure in FIG. 3 is terminated.

[0053] As described above, in the method for developing an autonomous vehicle according to the embodiment, in step S1, at least the first recognition model 111 and the first planning model 121 are trained using the autonomous vehicle 10 having sensor specifications superior to those of the autonomous vehicle 20. In step S2, the first recognition model 111 and the first planning model 121 trained in the autonomous vehicle 10 are applied to the autonomous vehicle 20. As a result, in the autonomous vehicle 20, a route equivalent to that of the autonomous vehicle 10 can be output using the trained first recognition model 111 and the first planning model 121. Therefore, it is possible to reduce the effort of re-adapting the autonomous vehicle 20 having the second sensor specifications, in which at least one of the number, type, and detection performance is inferior to the first sensor specifications, separately from the autonomous vehicle 10.

[0054] In the method for developing an autonomous vehicle according to the embodiment, the autonomous driving system 10a applies a machine learning model that outputs a control amount of autonomous driving control using the detection results of an external sensor as input for recognition processing, planning processing, and vehicle control processing, and in step S1, the autonomous driving vehicle 10 is used to adapt the recognition processing, planning processing, and vehicle control processing, thereby training the first recognition model 111, the first planning model 121, and the first vehicle control model 131. As a result, the machine learning model including the vehicle control processing is trained in the autonomous driving vehicle 10 and applied to the autonomous driving vehicle 20. As a result, in the autonomous driving vehicle 20, the first recognition model 111, the first planning model 121, and the first vehicle control model 131, which are the trained machine learning models, can be used to output a control amount of autonomous driving control equivalent to that of the autonomous driving vehicle 10.

[0055] In addition, according to the method for developing an autonomous vehicle according to the embodiment, not only can the development costs of the autonomous vehicle 20 with the cheaper autonomous driving system 20a be reduced, but by learning the output of the expensive autonomous driving system 10a, performance may be improved compared to when the autonomous driving system 20a is developed alone.

[0056] [Variations] Although various exemplary embodiments have been described above, various omissions, substitutions, and modifications may be made without being limited to the above-described exemplary embodiments.

[0057] In the above embodiment, the first vehicle control model 131 trained in the autonomous vehicle 10 is applied to the autonomous vehicle 20, but the present invention is not limited to this example. The machine learning model may be trained by adapting at least the recognition process and the planning process. In this case, in the autonomous driving system, at least the machine learning model for the recognition process and the machine learning model for the planning process may be regarded as an integrated machine learning model unit that receives the detection results of an external sensor as input and outputs a route.

[0058] In the above embodiment, the second external sensor 21 having the external camera 21a and the millimeter wave radar 21c is illustrated, but the "second sensor specification in which at least one of the number, type, and detection performance is inferior to the first sensor specification" is not limited to this example. For example, when the resolution of the external camera is different from that of all the cameras included in the external camera group 11a, the resolution of the external camera can be converted to match the resolution of the cameras included in the external camera group 11a, and used as an input of a machine learning model trained in the automatic driving system 10a. Alternatively, when the image quality of the external camera is different from that of all the cameras included in the external camera group 11a, noise is added to the captured image of the external camera 21a so as to match the image quality of the cameras included in the external camera group 11a, and used as an input of a machine learning model trained in the automatic driving system 10a. The same applies to the resolution and noise of the laser radar. [Explanation of symbols]

[0059] 10, 20...Autonomous driving vehicle, 10a, 20a...Autonomous driving system, 11...First external sensor (external sensor), 21...Second external sensor (external sensor).

Claims

1. A method for developing an autonomous vehicle equipped with external sensors including number, type, and detection performance as sensor specifications, comprising: The autonomous driving vehicle is equipped with an autonomous driving system to which a machine learning model is applied that inputs the detection results of the external sensors and outputs the route for at least the recognition processing and the planning processing among a recognition processing that recognizes an external environment based on the detection results of the external sensors, a planning processing that uses the results of the recognition processing to plan a route, and a vehicle control processing that performs autonomous driving control based on the planned route, a first step of training the machine learning model by adapting at least the perception process and the planning process using a first autonomous vehicle equipped with the external sensor of a first sensor specification; a second step of applying the machine learning model trained in the first step to a second autonomous vehicle equipped with external sensors of a second sensor specification in which at least one of the number, type, and detection performance is inferior to that of the first sensor specification.

2. The autonomous driving system applies the machine learning model to the recognition processing, the planning processing, and the vehicle control processing, the machine learning model inputting the detection results of the external sensors and outputting a control amount of the autonomous driving control, 2. The method for developing an autonomous vehicle according to claim 1, wherein in the first step, the machine learning model is trained by adapting the recognition process, the planning process, and the vehicle control process using a first autonomous vehicle.

Images