Determining the emotional state of a vehicle occupant

Abstract

A method for increasing the well-being of an occupant in a vehicle comprises training a plurality of machine learning models to predict the emotional state. The method may further comprise recording a plurality of sensor data using at least one vehicle sensor. The method may further comprise determining an occupant personality profile. The method may further comprise selecting a selected model from the plurality of machine learning models to predict the emotional state based at least in part on the occupant personality profile. The method may further comprise determining an emotional state of the occupant based at least in part on the plurality of sensor data using the selected model from the plurality of machine learning models to predict the emotional state.The method may further include adjusting operation of an autonomous driving system of the vehicle based at least in part on the emotional state of the occupant.

Description

introduction

[0001] The present disclosure relates to systems and methods for enhancing occupant well-being, and more particularly to systems and methods for enhancing occupant well-being based at least in part on identifying an occupant's mood or emotional state.

[0002] To increase an occupant's alertness, vehicles may be equipped with driver monitoring systems (DMS). DMS are configured to monitor one or more aspects of an occupant's mood or attention to influence the operation of the vehicle's systems. In some embodiments, DMS include one or more sensors (e.g., cameras) configured to monitor and / or measure one or more aspects of the occupant (e.g., facial expression, gaze direction, and / or the like). In one non-limiting example, the DMS may use signal processing algorithms to determine one or more aspects of the occupant's mood or attention. For example, a DMS may be used to disable a Society of Automotive Engineers (SAE) Level 3 automated driving system in response to determining that the occupant is not paying attention to the roadway.In another example, a DMS can be used to monitor driver fatigue and suggest that the driver take a break. However, current DMSs may not be configured to adaptively select signal-processing algorithms based on an occupant's predispositions to the vehicle or vehicle systems.

[0003] While current driver monitoring systems (DMS) serve their intended purpose, there is a need for a new and improved system and method to enhance the well-being of an occupant in a vehicle.

[0004] DE 10 2015 101 507 A1 describes an apparatus for use in adapting a controllable system based on user communication input, such as non-detailed or arbitrary input. The apparatus comprises a processor and a computer-readable storage medium including instructions that, when executed by the processor, cause the processor to perform operations for determining an objective system command corresponding to the user input for use in adapting the controllable system.

[0005] DE 10 2015 200 775 A1 describes a device, a driver assistance system, a means of transport, and a method for independently assessing the emotional state and cognitive load of a driver of a means of transport. The method comprises the following steps: - sensory detection of a physiological parameter of the driver, - computer-based detection of an interaction of the driver with a control element of the means of transport, - determining a traffic situation of the means of transport, and using the physiological parameter as well as the interaction and the traffic situation - determining a first assessment of the emotional state of the driver and a second assessment of the cognitive load of the driver based on an electronically stored reference.

[0006] DE 10 2021 101 451 A1 relates to a driver assistance system for a vehicle, in particular a motor vehicle. The driver assistance system comprises at least one optical sensor unit configured to optically detect an occupant of the vehicle and generate corresponding optical sensor data; at least one processor unit configured to determine at least one acceleration characteristic with respect to a body of the occupant based on the optical sensor data; and at least one control unit configured to control one or more vehicle functions based on the at least one acceleration characteristic in order to adjust an effect of driving dynamics on the occupant.

[0007] DE 11 2019 001 733 T5 describes a system for emotionally adaptive driving policies for automated vehicles, comprising a first plurality of sensors for detecting environmental information related to at least one occupant in a vehicle and a controller communicatively coupled to the plurality of sensors and comprising processing circuitry for receiving the environmental information from the first plurality of sensors, determining an emotional state of the at least one occupant from the environmental information, and implementing a driving policy based at least in part on the emotional state of the at least one occupant. Other examples may be described.

[0008] DE 10 2023 120 674 A1 describes a system, e.g., an autonomous vehicle, comprising a sensor array, a controller, and a computer-controlled device. The sensor array collects user data describing the current emotional state of a human user, e.g., a passenger of the representative autonomous vehicle. The controller executes a method for controlling the system. In particular, the controller identifies the user's psychological experience mode in response to the user data.

[0009] DE 10 2022 127 026 A1 describes systems and methods for supporting and executing automated control of climate comfort settings based on user classification within a group identity database. Methods for generating and deploying the group identity database are also disclosed. Description

[0010] The problem is solved by the subject matter according to claim 1. Further developments can be found in the subclaims.

[0011] According to several aspects, a method for increasing the well-being of an occupant in a vehicle is provided. The method may include training a plurality of machine learning models to predict the emotional state. The method may further include recording a plurality of sensor data using at least one vehicle sensor. The method may further include determining an occupant personality profile. The method may further include selecting a selected model from the plurality of machine learning models to predict the emotional state based at least in part on the occupant personality profile. The method may further include determining an emotional state of the occupant based at least in part on the plurality of sensor data using the selected model from the plurality of machine learning models to predict the emotional state.The method may further comprise adjusting operation of an autonomous driving system of the vehicle based at least in part on the emotional state of the occupant.

[0012] In another aspect of the present disclosure, training the plurality of machine learning models to predict emotional state may further comprise determining the occupant personality profile of each of a plurality of training occupants. The occupant personality profile of each of the plurality of training occupants is selected from a plurality of personality profiles. At least one of the plurality of training occupants has each of the plurality of personality profiles. Training the plurality of machine learning models to predict emotional state may further comprise providing one or more simulated driving scenarios to the plurality of training occupants. Training the plurality of machine learning models to predict emotional state may further comprise recording a plurality of training sensor data sets.Each of the plurality of training sensor data sets corresponds to one of the plurality of training occupants. Training the plurality of emotional state prediction machine learning models may further comprise training the plurality of emotional state prediction machine learning models based at least in part on the plurality of training sensor data sets and the occupant personality profile of each of the plurality of training occupants. Each of the plurality of emotional state prediction machine learning models corresponds to one of the plurality of personality profiles. The plurality of emotional state prediction machine learning models are configured to receive the plurality of sensor data as an input and provide the emotional state of the occupant as an output.

[0013] In another aspect of the present disclosure, determining the occupant personality profile of each of the plurality of training occupants may further comprise administering a survey to each of the plurality of training occupants. The survey includes one or more questions to determine a training occupant's attitude toward the autonomous driving system of the vehicle. Determining the occupant personality profile of each of the plurality of training occupants may further comprise determining the occupant personality profile of each of the plurality of training occupants based at least in part on one or more responses of each of the plurality of training occupants to the one or more survey questions.

[0014] In another aspect of the present disclosure, training the plurality of emotional state prediction machine learning models may further comprise training a low-confidence emotional state prediction machine learning model based on a first of the plurality of training sensor data sets corresponding to a first of the plurality of training occupants. The occupant personality profile of the first of the plurality of training occupants is a low-confidence occupant personality profile. Training the plurality of emotional state prediction machine learning models may further comprise training a medium-confidence emotional state prediction machine learning model based on a second of the plurality of training sensor data sets corresponding to a second of the plurality of training occupants.The occupant personality profile of the second of the plurality of training occupants is a medium-confidence occupant personality profile. Training the plurality of emotional state prediction machine learning models may further comprise training a high-confidence emotional state prediction machine learning model based on a third of the plurality of training sensor data sets corresponding to a third of the plurality of training occupants. The occupant personality profile of the third of the plurality of training occupants is a high-confidence personality profile.

[0015] In another aspect of the present disclosure, the method further comprises training a machine learning model to identify a personality profile based at least in part on the plurality of sensor data of each of the plurality of training occupants and the occupant personality profile of each of the plurality of training occupants. The personality profile identification machine learning model is configured to determine the occupant personality profile based at least in part on the plurality of sensor data.

[0016] In another aspect of the present disclosure, determining the occupant personality profile of the occupant may further comprise conducting a survey of the occupant. The survey includes one or more questions to determine an attitude of the occupant toward the autonomous driving system of the vehicle. Determining the occupant personality profile of the occupant may further comprise determining the occupant personality profile of the occupant based at least in part on one of: one or more responses of the occupant to the one or more survey questions; and the machine learning model for identifying the personality profile based at least in part on the plurality of sensor data.

[0017] In another aspect of the present disclosure, selecting the selected model from the plurality of emotional state prediction machine learning models based at least in part on the occupant personality profile may further comprise selecting a low-confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a low-confidence personality profile.Selecting the selected model from the plurality of emotional state prediction machine learning models based at least in part on the occupant personality profile may further comprise selecting a medium confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a medium confidence personality profile.Selecting the selected model from the plurality of emotional state prediction machine learning models based at least in part on the occupant personality profile may further comprise selecting a high confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a high confidence personality profile.

[0018] In another aspect of the present disclosure, determining the emotional state of the occupant may further comprise executing the selected model from the plurality of machine learning models to predict the emotional state. The selected model from the plurality of machine learning models to predict the emotional state is provided with the plurality of sensor data as an input. The selected model from the plurality of machine learning models to predict the emotional state provides the emotional state of the occupant as an output. The emotional state of the occupant includes at least one of the following: a low-stress emotional state, a medium-stress emotional state, and a high-stress emotional state.

[0019] In another aspect of the present disclosure, adjusting operation of the autonomous driving system of the vehicle may further comprise adjusting one or more driving parameters of the autonomous driving system in response to determining that the emotional state of the occupant is at least one of the following: the high distress emotional state and the medium distress emotional state.

[0020] In another aspect of the present disclosure, adjusting the one or more driving parameters of the autonomous driving system of the vehicle may further comprise adjusting the one or more driving parameters of the autonomous driving system in response to determining that the emotional state of the occupant is at least one of the following: the high distress emotional state and the medium distress emotional state. The one or more driving parameters include at least one of the following: a driving speed, a driving aggressiveness, a following distance, a maximum acceleration limit, and an activation state of an enhanced driver assistance feature.

[0021] According to several aspects, a system for enhancing the well-being of an occupant in a vehicle is provided. The system may include at least one vehicle sensor. The at least one vehicle sensor is operable to collect data about the occupant. The system may further include a system for autonomous driving of the vehicle. The system may further include a control unit for the vehicle in electrical communication with the at least one vehicle sensor and the autonomous driving system of the vehicle. The control unit of the vehicle is programmed to record a plurality of sensor data using the at least one vehicle sensor. The control unit of the vehicle is further programmed to determine an occupant personality profile of the occupant based at least in part on the plurality of sensor data. The occupant personality profile is one of a plurality of personality profiles.The vehicle controller is further programmed to select a selected one of a plurality of machine learning models for predicting the emotional state based at least in part on the occupant personality profile. Each of the plurality of machine learning models for predicting the emotional state corresponds to one of the plurality of personality profiles. The vehicle controller is further programmed to determine an emotional state of the occupant based at least in part on the plurality of sensor data using the selected one of the plurality of machine learning models for predicting the emotional state. The vehicle controller is further programmed to adjust operation of the autonomous driving system of the vehicle based at least in part on the emotional state of the occupant.

[0022] In another aspect of the present disclosure, the at least one vehicle sensor includes at least one camera configured to view a face of the occupant.

[0023] In another aspect of the present disclosure, to determine the personality profile of the occupant, the vehicle controller is further programmed to conduct a survey of the occupant. The survey includes one or more questions to determine an attitude of the occupant toward the autonomous driving system of the vehicle. To determine the occupant personality profile, the vehicle controller is further programmed to determine the occupant personality profile of the occupant based at least in part on one of: one or more responses of the occupant to the one or more survey questions; and a machine learning model to identify the personality profile based at least in part on the plurality of sensor data.

[0024] In another aspect of the present disclosure, in order to select the selected model from a plurality of emotional state prediction machine learning models, the vehicle controller is further programmed to select a low-confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a low-confidence personality profile.To select the selected model from a plurality of emotional state prediction machine learning models, the vehicle control unit is further programmed to select a medium confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a medium confidence personality profile.To select the selected model from a plurality of emotional state prediction machine learning models, the vehicle control unit is further programmed to select a high-confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a high-confidence personality profile.

[0025] In another aspect of the present disclosure, to determine the emotional state of the occupant, the vehicle's control unit is further programmed to execute the selected model from the plurality of machine learning models for predicting the emotional state. The selected model from the plurality of machine learning models for predicting the emotional state is provided with the plurality of sensor data as an input. The selected model from the plurality of machine learning models for predicting the emotional state provides the emotional state of the occupant as an output. The emotional state of the occupant includes at least one of the following: a low-stress emotional state, a medium-stress emotional state, and a high-stress emotional state.

[0026] In another aspect of the present disclosure, to adjust operation of the autonomous driving system, the vehicle's controller is further programmed to adjust one or more driving parameters of the vehicle's autonomous driving system in response to determining that the occupant's emotional state is at least one of the following: the high-stress emotional state and the medium-stress emotional state. The one or more driving parameters include at least one of the following: a driving speed, a driving aggressiveness, a following distance, a maximum acceleration limit, and an activation state of an enhanced driver assistance feature.

[0027] In another aspect of the present disclosure, the plurality of machine learning models for predicting emotional state are trained at least in part based on a plurality of training sensor datasets. Each of the plurality of training sensor datasets corresponds to one of the plurality of personality profiles.

[0028] According to several aspects, a method for increasing the well-being of an occupant in a vehicle is provided. The method may include recording a plurality of sensor data using at least one vehicle sensor. The method may further include determining an occupant personality profile of the occupant. The method may further include selecting a selected model from a plurality of machine learning models for predicting the emotional state based at least in part on the occupant personality profile. The method may further include determining an emotional state of the occupant based at least in part on the plurality of sensor data, wherein the selected model from the plurality of machine learning models is used to predict the emotional state.The method may further comprise adjusting operation of an autonomous driving system of the vehicle based at least in part on the emotional state of the occupant.

[0029] In another aspect of the present disclosure, the method further comprises determining the personality profile of each of a plurality of training occupants. The occupant personality profile of each of the plurality of training occupants is selected from a plurality of personality profiles. At least one of the plurality of training occupants has each of the plurality of personality profiles. The method further comprises providing one or more simulated driving scenarios to the plurality of training occupants. The method further comprises recording a plurality of training sensor data sets. Each of the plurality of training sensor data sets corresponds to one of the plurality of training occupants.The method further includes training the plurality of emotional state prediction machine learning models based at least in part on the plurality of training sensor data sets and the personality profile of each of the plurality of training occupants. Each of the plurality of emotional state prediction machine learning models corresponds to one of the plurality of personality profiles. The plurality of emotional state prediction machine learning models are configured to receive the plurality of sensor data as an input and provide the emotional state of the occupant as an output.

[0030] In another aspect of the present disclosure, training the plurality of emotional state prediction machine learning models may further include training a low-confidence emotional state prediction machine learning model based on a first of the plurality of training sensor data sets corresponding to a first of the plurality of training occupants. The occupant personality profile of the first of the plurality of training occupants is a low-confidence occupant personality profile. Training the plurality of emotional state prediction machine learning models may further include training a medium-confidence emotional state prediction machine learning model based on a second of the plurality of training sensor data sets corresponding to a second of the plurality of training occupants.The occupant personality profile of the second of the plurality of training occupants is a medium-confidence occupant personality profile. Training the plurality of emotional state prediction machine learning models may further comprise training a high-confidence emotional state prediction machine learning model based on a third of the plurality of training sensor data sets corresponding to a third of the plurality of training occupants. The occupant personality profile of the third of the plurality of training occupants is a high-confidence personality profile.

[0031] Further areas of applicability will become apparent from the present description. It should be understood that the description and specific examples are intended for purposes of illustration only and are not intended to limit the scope of the present disclosure. Short description of the drawings

[0032] The drawings described herein are for illustrative purposes only and are not intended to limit the scope of the present disclosure in any way. Fig. 1 is a schematic diagram of a system for enhancing the comfort of an occupant in a vehicle according to an exemplary embodiment; Fig. 2 is a schematic diagram of a driving simulation system according to an exemplary embodiment; Fig. 3 is a flowchart of the method for increasing the comfort of an occupant in a vehicle according to an exemplary embodiment; Fig. 4 is a flowchart of a method for training a plurality of machine learning models to predict emotional state according to an example embodiment; Fig.5A is a flowchart of a first exemplary embodiment of a method for determining an occupant personality profile of an occupant according to an exemplary embodiment; and Fig. 5B is a flowchart of a second exemplary embodiment of a method for determining an occupant personality profile of an occupant according to an exemplary embodiment. Detailed description

[0033] The following description is merely exemplary and is not intended to limit the present disclosure, application, or uses.

[0034] In aspects of the present disclosure, in-vehicle sensors may be used to detect a mood or emotional state of an occupant to determine the occupant's well-being with respect to actions of an autonomous driving system. However, vehicle occupants may have different predispositions regarding trust and confidence in autonomous driving systems, and detecting the emotional state without considering the occupant's predispositions may be less accurate. Therefore, the present disclosure provides a new and improved system and method for detecting the emotional state of a vehicle occupant with respect to the occupant's predispositions toward autonomous driving systems, thereby enabling actions to increase the occupant's well-being.

[0035] In Fig.1, a system for enhancing the comfort of an occupant in a vehicle is illustrated and generally designated by the reference numeral 10. The system 10 is illustrated with an example vehicle 12. Although a passenger car is depicted, the vehicle 12 may be any type of vehicle without departing from the scope of the present disclosure. The system 10 generally includes a vehicle control unit 14, at least one vehicle sensor 16, and an autonomous driving system of a vehicle 18.

[0036] The vehicle control unit 14 is used to perform a method 100 for increasing the comfort of an occupant in a vehicle, as described below. The vehicle control unit 14 includes at least one processor 20 and a non-volatile, computer-readable storage device or media 22. The processor 20 may be a custom or off-the-shelf processor, a central processing unit (CPU), a graphics processing unit (GPU), an auxiliary processor among a plurality of processors connected to the vehicle control unit 14, a semiconductor-based microprocessor (in the form of a microchip or chipset), a macroprocessor, a combination thereof, or generally a device for executing instructions. The computer-readable storage device or media 22 may include volatile and non-volatile memories, e.g.,Read-only memory (ROM), random-access memory (RAM), and keep-alive memory (KAM). KAM is persistent or non-volatile memory that can be used to store various operating variables while the processor 20 is powered off. The computer-readable storage device or media 22 can be implemented using a number of storage devices such as PROMs (programmable read-only memory), EPROMs (electrical PROMs), EEPROMs (electrically erasable PROMs), flash memory, or other electrical, magnetic, optical, or combination storage devices capable of storing data, some of which represent executable instructions used by the vehicle control unit 14 to control various systems of the vehicle 12.The vehicle control unit 14 may also consist of multiple control units that are electrically connected to one another. The vehicle control unit 14 may be connected to additional systems and / or control units of the vehicle 12 so that the vehicle control unit 14 can access data such as speed, acceleration, braking, and steering angle of the vehicle 12.

[0037] The vehicle control unit 14 is in electrical communication with the at least one vehicle sensor 16 and the autonomous driving system 18 of the vehicle. In an exemplary embodiment, the electrical communication is established, for example, via a CAN (Controller Area Network), a FLEXRAY network, a local area network (e.g., WiFi, Ethernet, and the like), a SPI (Serial Peripheral Interface), or the like. It is understood that various additional wired and wireless technologies and communication protocols for communicating with the vehicle control unit 14 are within the scope of the present disclosure.

[0038] The at least one vehicle sensor 16 is used to capture information about the occupant of the vehicle 12. Within the scope of the present disclosure, the occupant includes, as a non-limiting example, a driver, a passenger, and / or any additional persons in the vehicle 12. In one exemplary embodiment, the at least one vehicle sensor 16 includes at least one camera 24. In another exemplary embodiment, the at least one vehicle sensor 16 further includes a biometric sensor 26. In another exemplary embodiment, the at least one vehicle sensor 16 further includes one or more additional sensors 28. The at least one vehicle sensor 16 is in electrical communication with the vehicle control unit 14, as described above.

[0039] The camera 24 is used to capture images and / or videos of the occupant within the vehicle 12. In an exemplary embodiment, the camera 24 is a still and / or video camera positioned to view a face of the occupant within the vehicle 12. In one example, the camera 24 is mounted in a headliner of the vehicle 12 and has a view of the occupant of the vehicle 12. In another example, the camera 24 is mounted in, on, or near an instrument panel of the vehicle 12 and has a view of the occupant of the vehicle 12. In an exemplary embodiment, the camera 24 is used to capture one or more images of the occupant's face, enabling the determination of an emotional state of the occupant based at least in part on the one or more images, as explained in more detail below.

[0040] In some embodiments, camera 24 is part of a driver monitoring system (DMS). As a non-limiting example, a DMS is a system that uses sensors (e.g., camera 24) to monitor and analyze occupant behavior to ensure alertness, attention, and / or the like. It should be understood that camera 24 may include multiple cameras located at various locations within vehicle 12 without departing from the scope of the present disclosure. It should further be understood that cameras with various sensor types, including, for example, charge-coupled device (CCD) sensors, complementary metal oxide semiconductor (CMOS) sensors, and / or high dynamic range (HDR) sensors, are within the scope of the present disclosure. Furthermore, cameras with various lens types, such as wide-angle lenses and / or narrow-angle lenses, are also within the scope of the present disclosure.The camera 24 is in electrical communication with the vehicle control unit 14, as described above.

[0041] The biometric sensor 26 is used to perform biometric measurements of the occupant. Within the scope of the present disclosure, the biometric measurements include, for example, respiratory rate, heart rate, galvanic skin response, blood oxygen, body temperature, pupil dilation, brain activity, and / or the like. In an exemplary embodiment, the biometric sensor 26 includes at least one of the following: a respiratory rate sensor, a heart rate sensor, a galvanic skin response sensor, an electroencephalography (EEG) sensor, and / or the like. The biometric sensor 26 is in electrical communication with the vehicle control unit 14, as described above.

[0042] In one non-limiting example, the respiratory rate sensor is used to measure the respiratory rate (i.e., breathing) of the occupant. In one exemplary embodiment, the respiratory rate sensor is a pneumograph attached to the occupant's chest or abdomen. In another exemplary embodiment, the respiratory rate sensor is a non-contact infrared respiratory rate sensor mounted within the vehicle 12. In one exemplary embodiment, changes in respiratory rate may be associated with the occupant's emotional state. For example, an increased respiratory rate may indicate negative emotions such as stress or anger.

[0043] In one non-limiting example, the heart rate sensor is used to measure a heart rate of the occupant. In one exemplary embodiment, the heart rate sensor is an electrical sensor operable to detect a bioelectrical potential generated by electrical signals that control the expansion and contraction of the heart chambers. In another exemplary embodiment, the heart rate sensor is an optical sensor that uses light-based technology to measure a volume of blood pumped by the pumping action of the heart. In one non-limiting example, the heart rate sensor is disposed in a seat, armrest, steering wheel, and / or other surface that typically contacts the occupant within the vehicle 12. In one exemplary embodiment, changes in heart rate may be associated with the emotional state of the occupant.For example, an increased heart rate can indicate negative emotions such as stress or anger.

[0044] In one non-limiting example, the galvanic skin response sensor is used to measure a skin conductance of the occupant. In one example embodiment, the galvanic skin response sensor is an electrical sensor operable to measure electrical conductance between a plurality of electrodes in contact with the occupant's skin. In one non-limiting example, the galvanic skin response sensor is disposed in a seat, armrest, steering wheel, and / or other surface typically in contact with the occupant within the vehicle 12. In one example embodiment, changes in skin conductance (i.e., galvanic skin response) may be associated with the occupant's emotional state. For example, an increased skin conductance may indicate negative emotions such as stress, anger, or fear.

[0045] In one non-limiting example, the EEG sensor is used to measure the brainwave activity of the occupant. In one non-limiting example, the EEG sensor is disposed within a headrest of the vehicle 12. In one exemplary embodiment, various brainwave patterns may be associated with the emotional state of the occupant.

[0046] The one or more additional sensors 28 are used to capture additional information about the occupant. In an exemplary embodiment, the one or more additional sensors 28 include at least one of the following: a thermal imaging sensor operable to determine a skin temperature of the occupant (e.g., a facial skin temperature), a functional near-infrared spectroscopy (fNIRS) device, and an eye tracker operable to determine pupil diameter and gaze direction. The one or more additional sensors 28 are in electrical communication with the vehicle control unit 14, as described above.

[0047] The autonomous driving system of the vehicle 18 is used to assist the occupant to increase the occupant's attention and / or control the behavior of the vehicle 12. Within the scope of the present disclosure, the autonomous driving system of the vehicle 18 includes systems that provide any level of assistance to the occupant (e.g., blind spot warning, lane departure warning, and / or the like), as well as systems capable of driving the vehicle 12 autonomously under some or all conditions (e.g., automatic lane keeping, adaptive cruise control, fully autonomous driving, and / or the like). It is understood that all levels of driving automation defined, for example, by the Society of Automotive Engineers (SAE) J3016 (i.e., SAE LEVEL 0, SAE LEVEL 1, SAE LEVEL 2, SAE LEVEL 3, SAE LEVEL 4, and SAE LEVEL 5) are within the scope of the present disclosure.

[0048] In an exemplary embodiment, the autonomous driving system of the vehicle 18 is configured to detect and / or receive information about the surroundings of the vehicle 12 and process the information to provide assistance to the occupant. In some embodiments, the autonomous driving system of the vehicle 18 is a software module executing in the vehicle control unit 14. In other embodiments, the autonomous driving system of the vehicle 18 includes a separate autonomous driving system controller, similar to the vehicle control unit 14, that is capable of processing the information about the surroundings of the vehicle 12. In an exemplary embodiment, the autonomous driving system of the vehicle 18 can operate in a manual operating mode, a partially automated operating mode, and a fully automated operating mode.

[0049] Within the scope of the present disclosure, the manual mode of operation means that the autonomous driving system of the vehicle 18 warns or notifies the occupant, but does not directly intervene or control the vehicle 12. In one non-limiting example, the autonomous driving system of the vehicle 18 receives information from one or more perception sensors (i.e., one or more sensors operable to perceive information about the surroundings of the vehicle 12, such as one or more exterior cameras, a LIDAR sensor, a vehicle communication system, a global navigation satellite system, and / or the like). Using techniques such as computer vision, the autonomous driving system of the vehicle 18 understands the surroundings of the vehicle 12 and provides assistance to the occupant. For example, when the autonomous driving system of the vehicle 18 isone or more exterior cameras, a LIDAR sensor, a vehicle communication system, a global navigation satellite system, and / or the like) detects that the vehicle 12 is likely to collide with a distant vehicle, the autonomous driving system of the vehicle 18 may use an indicator to provide a warning to the occupant.

[0050] Within the scope of the present disclosure, the partially automated mode of operation means that the autonomous driving system of the vehicle 18 provides warnings or notifications to the occupant and, in certain situations, can directly intervene in or control the vehicle 12. In one non-limiting example, the autonomous driving system of the vehicle 18 is additionally in electrical communication with components of the vehicle 12, such as a braking system, a propulsion system (e.g., an internal combustion engine and / or an electric powertrain), and / or a steering system of the vehicle 12, such that the autonomous driving system of the vehicle 18 can control the behavior of the vehicle 12. In one non-limiting example, the autonomous driving system 18 of the vehicle can control the behavior of the vehicle 12 by applying the brakes of the vehicle 12 to avoid an impending collision.

[0051] In another non-limiting example, the autonomous driving system of the vehicle 18 may control the steering system of the vehicle 12 to provide an automatic lane keeping function. In another non-limiting example, the autonomous driving system of the vehicle 18 may control the braking system, the propulsion system, and the steering system of the vehicle 12 to temporarily drive the vehicle 12 to a predetermined destination. However, occupant intervention may be required at any time. In an exemplary embodiment, the autonomous driving system of the vehicle 18 may include additional components, such as a driver monitoring system (DMS), including, for example, an eye-tracking device configured to monitor an occupant's level of attention and ensure that the occupant is ready to assume control of the vehicle 12.

[0052] Within the scope of the present disclosure, the fully automated mode of operation means that the autonomous driving system of the vehicle 18 uses data from one or more perception sensors (e.g., one or more exterior cameras, a LIDAR sensor, a vehicle communication system, a global navigation satellite system, and / or the like) to understand the environment and control the vehicle 12 to drive the vehicle 12 to a predetermined destination without requiring control or intervention from the occupant.

[0053] The autonomous driving system of the vehicle 18 operates with a path planning algorithm configured to generate a safe and efficient trajectory for the vehicle 12 to navigate the environment surrounding the vehicle 12. In one exemplary embodiment, the path planning algorithm is a machine learning model trained to output control signals for the vehicle 12 based on input data acquired from the one or more perception sensors (e.g., one or more exterior cameras, a LIDAR sensor, a vehicle communication system, a global navigation satellite system, and / or the like). In another exemplary embodiment, the path planning algorithm is a deterministic algorithm programmed to output control signals for the vehicle 12 based on data acquired from the one or more perception sensors (e.g.,one or more external cameras, a LIDAR sensor, a vehicle communication system, a global navigation satellite system and / or the like).

[0054] As one non-limiting example, the path planning algorithm generates a sequence of waypoints or a continuous path that the vehicle 12 should follow to reach a destination while adhering to rules, regulations, and safety restrictions. The sequence of waypoints or the continuous path is created based at least in part on a detailed map and a current state of the vehicle 12 (i.e., position, speed, and orientation of the vehicle 12). The detailed map includes, for example, information about lane boundaries, road geometry, speed limits, traffic signs, and / or other relevant features. In an exemplary embodiment, the detailed map is stored in the media 22 of the vehicle control unit 14 and / or in a remote database or server.In another exemplary embodiment, the path planning algorithm performs perception and mapping tasks to interpret data collected by the one or more perception sensors (e.g., one or more exterior cameras, a LIDAR sensor, a vehicle communication system, a global navigation satellite system, and / or the like) and create, update, and / or augment the detailed map.

[0055] In an exemplary embodiment, one or more aspects of the operation of the autonomous driving system of the vehicle 18 are defined by one or more driving parameters of the autonomous driving system of the vehicle 18. In one non-limiting example, the driving parameters include, for example, a driving speed (i.e., a percentage of the road speed limit that the autonomous driving system of the vehicle 18 travels under normal conditions), a minimum following distance behind distant vehicles, a maximum acceleration limit, and / or the operating mode of the autonomous driving system of the vehicle 18 (i.e., the manual operating mode, the partially automated operating mode, or the fully automated operating mode). The value of one or more of the foregoing driving parameters may contribute to an overall driving aggressiveness of the autonomous driving system of the vehicle 18.Driving aggressiveness can be understood as the tendency of the autonomous driving system of the vehicle 18 to perform actions that are perceived as risky by the occupant (e.g., entering traffic with minimal gap / following distance).

[0056] In an exemplary embodiment, the driving parameters are controllable by at least the vehicle control unit 14, for example, in response to weather / road conditions, safety following distances, laws / regulations, the status of the vehicle's autonomous driving system 18, the vehicle's capabilities, the occupant's preference, the occupant's emotional state (as described in more detail below), and / or the like. It should be understood that the vehicle's autonomous driving system 18 may include any software and / or hardware module configured to operate in any combination of the manual operating mode, the partially automated operating mode, or the fully automated operating mode, as described above. The vehicle's autonomous driving system 18 is in electrical communication with the vehicle control unit 14, as described above.

[0057] In Fig.2 illustrates a schematic diagram of a driving simulation system 40. The driving simulation system 40 is used to train machine learning models by presenting simulated driving scenarios to a plurality of training occupants, as explained in more detail below. The driving simulation system 40 includes a driving simulator 42. The driving simulator 42 is a structure that enables the simulation of driving scenarios for training occupants. In an exemplary embodiment, the driving simulator 42 is a substantially stationary structure configured to look and feel like an interior of the vehicle 12. In one non-limiting example, the driving simulator 42 includes a steering wheel, an instrument panel, control pedals (e.g., a brake pedal and an accelerator pedal), one or more occupant seats, and / or the like.

[0058] The driving simulator 42 also includes the at least one vehicle sensor 16. The at least one vehicle sensor 16 in the driving simulator 42 is mounted and configured in the driving simulator 42 in a similar manner as in the vehicle 12. The driving simulator 42 further includes a simulator control unit 44 and a simulator display 50. The simulator control unit 44 includes at least one simulator processor 46 and a simulator non-transitory, computer-readable storage device or simulator medium 48. The description of the type and configuration given above for the vehicle control unit 14 also applies to the simulator control unit 44. In some examples, the simulator control unit 44 may differ from the vehicle control unit 14 in that the simulator control unit 44 is capable of higher processing speed, includes more memory, more inputs / outputs, and / or the like.By way of non-limiting example, the simulator processor 46 and simulator media 48 of the simulator control unit 44 are similar in structure and / or function to the processor 20 and media 22 of the vehicle control unit 14, as described above.

[0059] The simulator control unit 44 is in electrical communication with the at least one vehicle sensor 16 of the driving simulator 42 and the simulator display 50. In an exemplary embodiment, the electrical communication is established, for example, via a CAN (Controller Area Network), a FLEXRAY network, a local area network (e.g., WiFi, Ethernet, and the like), a SPI (Serial Peripheral Interface), or the like. It is understood that various additional wired and wireless techniques and communication protocols for communicating with the simulator control unit 44 are within the scope of the present disclosure.

[0060] The simulator display 50 is used to display one or more simulated driving scenarios. In an exemplary embodiment, the simulator display is arranged within the driving simulator 42 to mimic a view through a windshield of the vehicle 12. In one non-limiting example, the simulator controller 44 retrieves one or more videos containing simulated driving scenarios from the simulator media 48 and displays the one or more videos containing the simulated driving scenarios using the simulator display 50.

[0061] In Fig.3 illustrates a flowchart of the method 100 for increasing the well-being of an occupant in a vehicle. The method 100 begins at block 102 and proceeds to block 104. At block 104, a plurality of machine learning models are trained to predict the emotional state. In an exemplary embodiment, the plurality of machine learning models to predict the emotional state includes a low-confidence emotional state prediction machine learning model, a medium-confidence emotional state prediction machine learning model, and a high-confidence emotional state prediction machine learning model.It is understood that the plurality of emotional state prediction machine learning models may include various additional emotional state prediction machine learning models with different parameters and features without departing from the scope of the present disclosure. Training the plurality of emotional state prediction machine learning models is discussed in more detail below. After block 104, the method 100 proceeds to block 106.

[0062] In block 106, the vehicle control unit 14 uses the at least one vehicle sensor 16 to collect a plurality of sensor data. Within the scope of the present disclosure, the plurality of sensor data includes sensor data about the occupant, including, for example, sensor data collected by the camera 24, the biometric sensor 26, and / or the one or more additional sensors 28.In one non-limiting example, the sensor data about the occupant includes one or more images of the occupant's face captured by the camera 24, a respiratory rate, a heart rate, a galvanic skin response, a blood oxygen, a body temperature, pupil dilation, brain activity, and / or the like captured by the biometric sensor 26, a skin temperature of the occupant, a pupil diameter of the occupant, and / or an eye gaze direction of the occupant captured by the one or more additional sensors 28, and / or the like. In an exemplary embodiment, the plurality of sensor data is stored in the media 22 of the vehicle control unit 14 for later retrieval. After block 106, the method 100 proceeds to block 108.

[0063] In block 108, the vehicle controller 14 determines an occupant personality profile of the occupant. Within the scope of the present disclosure, the occupant's occupant personality profile is a metric that describes the occupant's attitude toward the autonomous driving system of the vehicle 18 (i.e., the occupant's level of trust in, comfort with, and / or confidence in the autonomous driving system of the vehicle 18). In an exemplary embodiment, the occupant personality profile is selected from a plurality of personality profiles. In an exemplary embodiment, the plurality of personality profiles includes at least: a low-trust occupant personality profile, a medium-trust occupant personality profile, and a high-trust occupant personality profile.It is understood that the plurality of personality profiles may include various additional personality profiles without departing from the scope of the present disclosure. The determination of the occupant's personality profile is discussed in more detail below. After block 108, the method 100 proceeds to block 110.

[0064] In block 110, the vehicle controller 14 selects one of the emotional state prediction machine learning models trained in block 104 (hereinafter referred to as the selected model from the plurality of emotional state prediction machine learning models). In an exemplary embodiment, the vehicle controller 14 selects the selected model from the plurality of emotional state prediction machine learning models based at least in part on the occupant personality profile determined in block 108. In one non-limiting example, the vehicle controller 14 selects the low-confidence emotional state prediction machine learning model in response to determining that the occupant personality profile is the low-confidence personality profile.The vehicle controller 14 selects the medium-confidence emotional state prediction machine learning model in response to determining that the occupant personality profile is the medium-confidence personality profile. The vehicle controller 14 selects the high-confidence emotional state prediction machine learning model in response to determining that the occupant personality profile is the high-confidence occupant personality profile. It should be understood that the vehicle controller 14 may select various additional emotional state prediction machine learning models from the plurality of emotional state prediction machine learning models without departing from the scope of the present disclosure. After block 110, the method 100 proceeds to block 112.

[0065] In block 112, the vehicle control unit 14 determines an emotional state of the occupant (hereinafter referred to as occupant emotional state). Within the scope of the present disclosure, the occupant emotional state is a metric that describes the level of negative emotion (e.g., stress, worry, confusion, fear, anger, and / or the like) of the occupant. In an exemplary embodiment, the occupant's emotional state is quantified as a value on a continuous scale (e.g., a continuous scale between zero and one hundred, where zero indicates very low negative emotion and one hundred indicates extremely high negative emotion).

[0066] In another exemplary embodiment, the emotional state of the occupant includes at least one of the following states: a low distress emotional state, a moderate distress emotional state, and a high distress emotional state. Within the scope of the present disclosure, the low distress emotional state describes an occupant who exhibits little to no negative emotions. Within the scope of the present disclosure, the moderate distress emotional state describes an occupant who exhibits a moderate level of negative emotions. Within the scope of the present disclosure, the high distress emotional state describes an occupant who exhibits a high level of negative emotions.

[0067] In an exemplary embodiment, to determine the emotional state of the occupant, the vehicle control unit 14 executes the selected model from the plurality of machine learning models for predicting the emotional state selected in block 110. In an exemplary embodiment, the selected model from the plurality of machine learning models for predicting the emotional state is provided with the plurality of sensor data recorded in block 106 as an input and provides the emotional state of the occupant as an output. The plurality of machine learning models for predicting the emotional state are discussed in more detail below. After block 112, the method 100 proceeds to block 114.

[0068] At block 114, in one exemplary embodiment, if the occupant's emotional state determined at block 112 and quantified on the continuous scale is less than a predetermined emotional state threshold, the method 100 transitions to a standby state at block 116. If the occupant's emotional state determined at block 112 and quantified on the continuous scale is greater than or equal to the predetermined emotional state threshold, the method 100 transitions to block 118, as explained in more detail below. In another exemplary embodiment, if the occupant's emotional state determined at block 112 is the low distress emotional state, the method 100 transitions to the standby state at block 116.If the occupant's emotional state determined in block 112 is the medium distress emotional state or the high distress emotional state, the method 100 proceeds to block 118.

[0069] In block 118, the vehicle controller 14 adjusts the operation of the autonomous driving system 18 of the vehicle. In an exemplary embodiment, to adjust the operation of the autonomous driving system of the vehicle 18, the vehicle controller 14 adjusts the driving parameters of the autonomous driving system of the vehicle 18. The vehicle controller 14 adjusts the driving parameters of the autonomous driving system of the vehicle 18 to improve the emotional state of the occupant (i.e., to reduce the level of negative emotions of the occupant). In one non-limiting example, the vehicle controller 14 reduces the driving speed, increases the minimum following distance behind distant vehicles, decreases the maximum acceleration limit, and / or decreases driving aggressiveness.

[0070] In another exemplary embodiment, the vehicle control unit 14 takes additional measures to improve the emotional state of the occupant. In one non-limiting example, the vehicle control unit 14 uses a display of the vehicle 12 to provide the occupant with notification of the functioning of the vehicle's autonomous driving system 18. In one non-limiting example, the vehicle control unit 14 controls the operation of a vehicle heating, ventilation, and air conditioning (HVAC) system. In one non-limiting example, the vehicle control unit 14 controls the operation of the vehicle's interior lighting. In one non-limiting example, the vehicle control unit 14 controls the operation of a regenerative braking system in the vehicle.In one non-limiting example, the vehicle control unit 14 regulates the operation of a global navigation satellite system (GNSS) of the vehicle 12 (e.g., suggesting an alternative navigation route to a destination). In one non-limiting example, the vehicle control unit 14 sets an activation state (i.e., activation or deactivation) of an enhanced driver assistance feature (e.g., a lane departure warning feature, an adaptive cruise control feature, a forward collision warning feature, and / or the like). It should be understood that the vehicle control unit 14 may take further actions to improve the emotional state of the occupant without departing from the scope of the present disclosure. After block 118, the method 100 transitions to the standby state at block 116.

[0071] In an exemplary embodiment, the vehicle control unit 14 repeatedly exits the standby state 116 and restarts the method 100 in block 102. In one non-limiting example, the vehicle control unit 14 exits the standby state 116 and restarts the method 100 after a timer, e.g., every three hundred milliseconds.

[0072] In Fig.4 is a flowchart of an exemplary embodiment 104a of block 104 (i.e., a method for training the plurality of machine learning models to predict the emotional state). It should be understood that the exemplary embodiment 104a may be performed with any computer system capable of providing simulated driving scenarios. It should further be understood that the exemplary embodiment 104a may be performed using crowdsourced data from actual vehicles and / or occupants. In one exemplary embodiment, discussed below, the exemplary embodiment 104a is performed with the driving simulator system 40 and a plurality of training occupants.

[0073] Within the scope of the present disclosure, the plurality of training occupants includes a plurality of driving-age individuals with different backgrounds, different personalities, different levels of driving experience, and / or experience with automated driving systems. In one non-limiting example, the plurality of training occupants is substantially representative of the general driver population. The exemplary embodiment 104a begins in block 402. In block 402, the simulator controller 44 determines the personality profile of each of the plurality of training occupants. In one exemplary embodiment, to determine the personality profile of each of the plurality of training occupants, the simulator controller 44 conducts an interview with each of the plurality of training occupants.In one non-limiting example, the survey includes one or more questions to determine a training occupant's attitude toward the autonomous driving system of the vehicle 18 (i.e., the occupant's level of confidence in, comfort with, and / or trust in the autonomous driving system of the vehicle 18). In an exemplary embodiment, the survey includes a plurality of questions, each of which the training occupant can respond to with one of the following: very disagree, moderately disagree, neither agree nor disagree, moderately agree, or very agree.

[0074] A set of sample survey questions is presented in Table 1. As a non-limiting example, the sample survey includes twenty-eight questions divided into three categories: agentic abilities, synchronized mental model, and willingness to use autonomy. Table 1: A series of sample survey questions. Agentic Capability The autonomous agent... Synchronized mental model The autonomous agent... Autonomy Willingness to use 1. has the ability to make some decisions independently 11. is generally on the "same side" as me 19. The car offers safety 2. has the authority to make decisions 12. Gives me what I need before I ask for it 20. The car behaves in an unexpected way 3. has the ability to control his actions 13.... and I don't feel in harmony with each other 21. I am suspicious of the car's intention or actions 4. always seems to need my help 14. is connected to me throughout the task 22. I have confidence in the car 5. I don't have to control it 100% of the time 15. reacts in an expected manner 23. I am suspicious of the car 6. can perform his tasks independently 16. adapts to new situations as if I do 24. The car is reliable 7. has control over his own actions 17. anticipates my actions 25. The car is trustworthy 8. implements the best course of action 18. anticipates my needs 26. The car's action may have harmful or injurious consequences 9. requires me to monitor him 27. I can trust the car 10. can work with little to no supervision 28. I feel comfortable while driving in the car In one exemplary embodiment, the survey is presented to each of the plurality of training occupants using the simulator display 50, and the responses are stored in the simulator media 48 of the simulator control unit 44. In another exemplary embodiment, the survey is presented to each of the plurality of training occupants outside of the simulation system 40 (e.g., online or on paper), and the responses are subsequently loaded into the simulator media 48 of the simulator control unit 44. After block 402, the exemplary embodiment 104a proceeds to block 404.

[0075] In block 404, the simulator controller 44 determines the personality profile of each of the plurality of training occupants based at least in part on one or more responses of each of the plurality of training occupants to the one or more questions of the survey administered in block 402. In an exemplary embodiment, the simulator controller 44 uses a deterministic algorithm to determine the occupant personality profile based directly on the survey responses to classify each of the plurality of occupants into the low-confidence occupant personality profile, the medium-confidence occupant personality profile, or the high-confidence occupant personality profile.

[0076] In another exemplary embodiment, to determine the occupant personality profile, the simulator controller 44 uses a machine learning model trained to classify each of the plurality of occupants into the low-confidence occupant personality profile, the medium-confidence occupant personality profile, or the high-confidence occupant personality profile based at least in part on the survey responses. After block 404, the exemplary embodiment 104a proceeds to blocks 406 and 408.

[0077] In block 406, the simulator control unit 44 uses the simulator display 50 to provide each of the plurality of training occupants with one or more simulated driving scenarios. In one non-limiting example, the one or more simulated driving scenarios include scenarios with varying weather conditions, varying traffic conditions, varying road conditions, and / or the like. In one non-limiting example, the one or more simulated driving scenarios include scenarios in which a simulated autonomous driving system controls behavior of the simulated vehicle. In one non-limiting example, the one or more simulated driving scenarios include at least one known non-stress-inducing scenario (e.g., a simulated drive on an open road in good weather conditions with no traffic) and at least one known stress-inducing scenario (e.g.,a simulated drive in bad weather conditions with high speed and heavy traffic). After block 406, the exemplary embodiment 104a proceeds to blocks 410 and 412, as explained in more detail below.

[0078] At block 408, concurrent with the playback of the simulated driving scenarios at block 406, the simulator controller 44 records a training sensor data set for each of the plurality of training occupants (i.e., a plurality of training sensor data sets). In an exemplary embodiment, each of the plurality of training sensor data sets includes data collected from one of the plurality of training occupants by the one or more vehicle sensors 16 in the simulator 42 during the one or more simulated driving scenarios. After block 408, the exemplary embodiment 104a proceeds to blocks 410 and 412.

[0079] In block 410, the simulator controller 44 trains the plurality of machine learning models to predict the emotional state based at least in part on the plurality of training sensor data sets collected in block 408, the occupant personality profile of each of the plurality of training occupants determined in block 404, and the content of the one or more simulated driving scenarios. In an exemplary embodiment, the content of the one or more simulated driving scenarios provides baseline information for training. For example, the physiological response of a training occupant (as measured by the one or more vehicle sensors 16) during replay of the known, non-stressful driving scenario may serve as a baseline for identifying the low-stress emotional state.In another example, the physiological response of a training occupant (measured by the one or more vehicle sensors 16) during replay of the known stress-inducing driving scenario may serve as a baseline for identifying the high-stress emotional state.

[0080] In an exemplary embodiment, each of the plurality of emotional state prediction machine learning models is trained using one or more of the plurality of training sensor datasets from training occupants with the same occupant personality profile. In a non-limiting example, the low-confidence emotional state prediction machine learning model is trained based on a first of the plurality of training sensor datasets corresponding to a first of the plurality of training occupants, wherein the occupant personality profile of the first of the plurality of training occupants is the low-confidence occupant personality profile.In one non-limiting example, the machine learning model is trained to predict the emotional state with medium confidence based on a second one of the plurality of training sensor data sets corresponding to a second one of the plurality of training occupants, wherein the occupant personality profile of the second one of the plurality of training occupants is a medium confidence occupant personality profile. In one non-limiting example, the machine learning model is trained to predict the emotional state with high confidence based on a third one of the plurality of training sensor data sets corresponding to a third one of the plurality of training occupants, wherein the occupant personality profile of the third one of the plurality of training occupants is a high confidence occupant personality profile.

[0081] The following description of the operation and training of an emotional state prediction machine learning model applies to each of the plurality of emotional state prediction machine learning models, provided that each of the plurality of emotional state prediction machine learning models is trained using one or more of the plurality of training sensor data sets from training occupants with the same occupant personality profile, as described above. In one non-limiting example, the emotional state prediction machine learning model comprises multiple layers, including an input layer and an output layer, and one or more hidden layers. The emotional state prediction machine learning model receives the plurality of sensor data recorded in block 106 as inputs. The inputs are then passed to the hidden layers.Each hidden layer applies a transformation (e.g., a nonlinear transformation) to the data and passes the result to the next hidden layer, up to the last hidden layer. The output layer generates the occupant's emotional state.

[0082] To train the machine learning model for emotional state prediction, a subset of the plurality of training sensor data sets collected in block 408 and the content of the one or more simulated driving scenarios are used. The content of the one or more simulated driving scenarios provides basic information, as described above. The algorithm is trained by adjusting the internal weights between the nodes in each hidden layer to minimize the prediction error. During training, an optimization technique (e.g., gradient descent) is used to adjust the internal weights and reduce the prediction error. The training process is repeated until the prediction error is minimized, and the thus-trained model is then used to classify new input data.

[0083] After sufficient training of the emotional state prediction machine learning model, the algorithm is able to accurately and precisely determine the occupant's emotional state based on the sensor data recorded in block 106. By adjusting the weights between the nodes in each hidden layer during training, the algorithm "learns" to recognize patterns in the sensor data that are indicative of the occupant's emotional state. In an exemplary embodiment, each of the plurality of emotional state prediction machine learning models is loaded into the media 22 of the vehicle control unit 14 for use by the vehicle control unit 14.As a non-limiting example, each of the plurality of emotional state prediction machine learning models may be further trained and / or retrained using real-world data collected by the vehicle controller 14 to improve an accuracy of each of the plurality of emotional state prediction machine learning models. After block 410, the exemplary embodiment 104a is complete, and the method 100 continues as described above.

[0084] In block 412, the simulator controller 44 trains a personality profile identification machine learning model. Within the scope of the present disclosure, the personality profile identification machine learning model is configured to determine the occupant personality profile based at least in part on the sensor data recorded in block 106, thereby enabling the occupant personality profile to be determined without interviewing each occupant.

[0085] As a non-limiting example, the personality profile identification machine learning model comprises multiple layers, including an input layer and an output layer, as well as one or more hidden layers. The personality profile identification machine learning model receives the plurality of sensor data recorded in block 106 as inputs. The inputs are then passed to the hidden layers. Each hidden layer applies a transformation (e.g., a nonlinear transformation) to the data and passes the result to the next hidden layer, up to the last hidden layer. The output layer generates the occupant personality profile.

[0086] To train the machine learning model to identify the personality profile, the plurality of training sensor datasets collected in block 408 and the occupant personality profile of each of the plurality of training occupants determined based on one or more responses of each of the plurality of training occupants to the one or more questions of the survey administered in block 402 are used. The algorithm is trained by adjusting the internal weights between the nodes in each hidden layer to minimize the prediction error. During training, an optimization technique (e.g., gradient descent) is used to adjust the internal weights and reduce the prediction error. The training process is repeated until the prediction error is minimized, and the resulting trained model is then used to classify new input data.

[0087] After sufficient training of the personality profile identification machine learning model, the algorithm is able to accurately and precisely determine the occupant personality profile based on the plurality of sensor data recorded in block 106. By adjusting the weights between the nodes in each hidden layer during training, the algorithm "learns" to recognize patterns in the sensor data that are indicative of the occupant's emotional state. In an exemplary embodiment, the personality profile identification machine learning model is loaded into the media 22 of the vehicle control unit 14 for use by the vehicle control unit 14.As a non-limiting example, the machine learning model for identifying the personality profile may be further trained and / or retrained using real data collected by the vehicle control unit 14 to improve the accuracy of the machine learning model for identifying the personality profile. After block 412, the exemplary embodiment 104a is complete, and the method 100 continues as described above.

[0088] Fig.5A shows a flowchart of a first exemplary embodiment 108a of block 108 (i.e., a method for determining the occupant's personality profile). The first exemplary embodiment 108a begins at block 502a. At block 502a, the vehicle controller 14 conducts a survey of the occupant. In one non-limiting example, the survey is conducted upon a new occupant's first entry into the vehicle 12. In another non-limiting example, the survey is repeated at regular and / or irregular intervals. In one non-limiting example, the survey includes one or more questions to determine a training occupant's attitude toward the autonomous driving system of the vehicle 18 (i.e., the level of confidence in the autonomous driving system of the vehicle 18, the occupant's comfort level, and / or the confidence in the autonomous driving system of the vehicle 18).In an exemplary embodiment, the survey contains a plurality of questions, each of which the training occupant can respond to with one of the following: very inaccurate, moderately inaccurate, neither inaccurate nor accurate, moderately accurate, or very accurate. A sample set of survey questions is shown above in Table 1.

[0089] In one exemplary embodiment, the survey is presented to the occupant via a display of the vehicle 12, and the responses are stored in the media 22 of the vehicle control unit 14. In another exemplary embodiment, the survey is presented to the occupant outside the vehicle 12 (e.g., on a mobile device), and the responses are subsequently loaded into the media 22 of the vehicle control unit 14. After block 502a, the first exemplary embodiment 108a proceeds to block 504a. In block 504a, the vehicle control unit 14 determines the occupant's occupant personality profile based at least in part on one or more of the occupant's responses to the one or more questions of the survey administered in block 504a.In an exemplary embodiment, the vehicle control unit 14 uses a deterministic algorithm based directly on the survey responses to determine the occupant personality profile to classify each of the plurality of occupants into the low confidence occupant personality profile, the medium confidence occupant personality profile, or the high confidence occupant personality profile.

[0090] In another exemplary embodiment, to determine the occupant's personality profile, the vehicle controller 14 uses a machine learning model trained to classify each of the plurality of occupants into the low-confidence occupant personality profile, the medium-confidence occupant personality profile, or the high-confidence occupant personality profile based at least in part on the survey responses. After block 504a, the first exemplary embodiment 108a is complete, and the method 100 continues as described above.

[0091] Fig.5B shows a flowchart of a second exemplary embodiment 108b of block 108 (i.e., a method for determining the occupant personality profile). The second exemplary embodiment 108b begins in block 502b. In block 502b, the vehicle controller 14 uses the personality profile identification machine learning model trained in block 412 to determine the occupant personality profile based at least in part on the plurality of sensor data recorded in block 106. After block 502b, the second exemplary embodiment 108b is complete, and the method 100 continues as described above.

[0092] It will be appreciated that either the first exemplary embodiment 108a of block 108, the second exemplary embodiment 108b of block 108, or both may be used to determine the occupant personality profile of the occupant during execution of block 108 in method 100.

[0093] The system 10 and method 100 of the present disclosure provide several advantages. By evaluating an occupant's attitude toward the autonomous driving system of the vehicle 18 (i.e., the occupant personality profile) and using a corresponding model from the plurality of machine learning models to predict the occupant's emotional state, the precision and accuracy of predicting the emotional state based on sensor data is increased. Furthermore, the system 10 and method 100 can be used to reassess the occupant personality profile periodically and / or in response to changing sensor data indicating a change in the occupant's personality profile. Subsequently, the vehicle controller 14 can take actions to improve occupant well-being based on the occupant's emotional state.Finally, the system 10 and the method 100 are applicable to various other manned automated vehicles, such as partially / fully automated aircraft.

[0094] The description of the present disclosure is merely exemplary, and variations that do not depart from the gist of the present disclosure are intended to be included within the scope of the present disclosure. Such variations should not be considered a departure from the spirit and scope of the present disclosure.

Claims

[1] A method for increasing the well-being of an occupant in a vehicle, the method comprising: Training a variety of machine learning models to predict emotional state; Recording a plurality of sensor data using at least one vehicle sensor; Determining an occupant personality profile; Selecting a selected model from the plurality of machine learning models to predict the emotional state based at least in part on the occupant personality profile; Determining an emotional state of the occupant based at least in part on the plurality of sensor data using the model selected from the plurality of machine learning models for predicting the emotional state; and Adjusting operation of an autonomous driving system of the vehicle based at least in part on the emotional state of the occupant, wherein training the plurality of machine learning models to predict the emotional state further comprises: Determining the occupant personality profile of each of a plurality of training occupants, wherein the occupant personality profile of each of the plurality of training occupants is selected from a plurality of personality profiles, and wherein at least one of the plurality of training occupants has each of the plurality of personality profiles; Providing one or more simulated driving scenarios to the plurality of training occupants; Recording a plurality of training sensor data sets, each of the plurality of training sensor data sets corresponding to one of the plurality of training occupants; and Training the plurality of machine learning models to predict the emotional state based at least in part on the plurality of training sensor data sets and the occupant personality profile of each of the plurality of training occupants, wherein each of the plurality of emotional state prediction machine learning models corresponds to one of the plurality of personality profiles, and wherein the plurality of emotional state prediction machine learning models is configured to receive the plurality of sensor data as an input and provide the emotional state of the occupant as an output, further comprehensive: Training a machine learning model to identify a personality profile based at least in part on the plurality of sensor data from each of the plurality of training occupants and the occupant personality profile of each of the plurality of training occupants, wherein the machine learning model to identify the personality profile is configured to determine the occupant personality profile based at least in part on the plurality of sensor data, wherein determining the occupant personality profile further comprises: Conducting a survey of the occupant, the survey comprising one or more questions to determine an attitude of the occupant regarding the autonomous driving system of the vehicle; and Determining the occupant personality profile based at least in part on at least one of: one or more responses of the occupant to the one or more survey questions and the machine learning model for identifying the personality profile based at least in part on the plurality of sensor data, wherein selecting the selected model from the plurality of machine learning models for predicting the emotional state based at least in part on the occupant personality profile further comprises: selecting a low-confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a low-confidence personality profile; Selecting a medium-confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a medium-confidence personality profile; and Selecting a high-confidence emotional state prediction machine learning model from the plurality of emotional state prediction machine learning models in response to determining that the occupant personality profile is a high-confidence personality profile. [2] The method of claim 1, wherein determining the occupant personality profile of each of the plurality of training occupants further comprises: Conducting a survey with each of the plurality of training occupants, the survey comprising one or more questions to determine a training occupant's attitude toward the autonomous driving system of the vehicle; and Determining the occupant personality profile of each of the plurality of training occupants based at least in part on one or more responses of each of the plurality of training occupants to the one or more survey questions. [3] The method of claim 1, wherein training the plurality of machine learning models to predict the emotional state further comprises: Training a machine learning model to predict the emotional state with low confidence based on a first of the plurality of training sensor datasets, corresponding to a first of the plurality of training occupants, wherein the occupant personality profile of the first of the plurality of training occupants is a low-confidence occupant personality profile; Training a machine learning model to predict the emotional state at medium confidence based on a second of the plurality of training sensor data sets corresponding to a second of the plurality of training occupants, wherein the occupant personality profile of the second of the plurality of training occupants is a medium confidence occupant personality profile; and Training a machine learning model to predict the emotional state with high confidence based on a third of the plurality of training sensor data sets corresponding to a third of the plurality of training occupants, wherein the occupant personality profile of the third of the plurality of training occupants is a high confidence occupant personality profile. [4] The method of claim 1, wherein determining the emotional state of the occupant further comprises: Executing the model selected from the plurality of machine learning models for predicting the emotional state, wherein the model selected from the plurality of machine learning models for predicting the emotional state is provided with the plurality of sensor data as an input, wherein the model selected from the plurality of machine learning models for predicting the emotional state provides the emotional state of the occupant as an output, and wherein the emotional state of the occupant comprises at least one of the following: a low distress emotional state, a medium distress emotional state, and a high distress emotional state. [5] The method of claim 4, wherein adjusting the operation of the autonomous driving system of the vehicle further comprises: Adjusting one or more driving parameters of the autonomous driving system of the vehicle in response to determining that the emotional state of the occupant is at least one of the following: the high distress emotional state and the medium distress emotional state. [6] The method of claim 5, wherein adjusting the one or more driving parameters of the autonomous driving system of the vehicle further comprises: Adjusting the one or more driving parameters of the autonomous driving system of the vehicle in response to determining that the emotional state of the occupant is at least one of the following: the high distress emotional state and the medium distress emotional state, wherein the one or more driving parameters include at least one of the following: a driving speed, a driving aggressiveness, a following distance, a maximum acceleration limit, and an activation state of an advanced driver assistance feature.

Images