Abstract: Autonomous vehicle operational management may include an autonomous vehicle traversing a vehicle transportation network, which may include operating a scenario-specific operational control evaluation module instance, which is an instance of a scenario-specific operational control evaluation module instantiated for an occurrence of a distinct vehicle operational scenario by allocating computing resources to, and populating, the scenario-specific operational control evaluation module instance with operational environment information corresponding to the occurrence of the distinct vehicle operational scenario.

Abstract: A system for vehicle decision-making using sequential information probing generates a first world model for a feature of a vehicle operational scenario. The first world model is a copy of a second world model that represents the vehicle operational scenario using incomplete state information for the feature. The first world model represents the vehicle operational scenario using complete state information for the feature. A first expected reward is determined for a sequence of actions of an artificial intelligence (AI) agent using the first world model. A second expected reward is determined for the sequence of actions using the second world model. An expected loss in decision-making performance associated with the incomplete state information for the feature is calculated based on the first and second expected rewards. The expected loss is used to inform a decision-making process for controlling an autonomous vehicle to traverse a portion of a vehicle transportation network.

Abstract: A vehicle includes an on-hoard satellite navigation device, a telematics control unit, a non-transitory computer readable medium, and a processor. The on-board satellite navigation device is in communication with a global positioning system unit to acquire real-time information regarding conditions near the vehicle's vicinity. The telematics control unit is in wireless communications to a cloud services or a vehicle network to upload and receive crowdsourced information regarding conditions near the vehicle's vicinity. The non-transitory computer readable medium stores predetermined information regarding conditions near the vehicle vicinity. The processor is programmed to output one or more complexity values related to the vehicle's vicinity during vehicle travel based on one or more of the real-time information, the crowdsourced information and the predetermined information.

Abstract: A method of displaying road information in a host vehicle includes receiving with the host vehicle the road information acquired by another vehicle. The road information includes information about a portion of a road ahead of a current portion of the road upon which the host vehicle is traveling. An information presentation time is determined at which the road information is configured to be presented in the host vehicle. The road information is presented in the host vehicle at the determined information presentation time.

Abstract: System and method for dynamically refining custom classes using zero-shot image classifiers. The system uses a CLIP model to generate text embeddings of object descriptions and image embeddings of captured images, and determines similarity scores between the text and image embeddings. When a similarity score exceeds a threshold, the system notifies a user that a captured image includes the object, and the system stores the relevant text prompt and captured image as a custom class.

Abstract: A video of a driver of a vehicle is obtained. Based on subset of frames of the video, a series of body poses are identified. The series of body poses are identified by detecting landmark points associated with respective body parts of the driver. The landmark points correspond to coordinates of locations of pixels that represent at least one or more joints of the respective body parts of the driver in the subset of frames. A driver behavior is identified based on the series of the body poses. An assistive vehicle control action for the vehicle is output based on the driver behavior.

Abstract: Vehicle decision-making is analyzed and can be used to modify a decision-making process. For subsets of features comprising a vehicle operational scenario, a first value is generated that quantifies behavior of an artificial intelligence (AI) agent as the AI agent performs a sequence of actions within a first world model based on a complete set of observations for the subset of features. A first world model is a copy of a second world model for sequential decision making. A second value is generated that quantifies behavior of an AI agent as the AI agent performs a sequence of actions in the second world model based on an incomplete set of observations for the subset of features. A difference between the first and second values determines the impact of individual features on the AI agent within the second world model. A decision-making process of the AI agent can be updated.

Abstract: A vehicle traversing a vehicle transportation network may use a scenario-specific operational control evaluation model instance. The vehicle may receive operational environment data that includes data associated with an object external to the vehicle. The vehicle may determine, based at least in part on the data associated with the object, a final belief associated with an undesired outcome. The vehicle may construct a backward Monte Carlo tree search model based on the final belief and may determine, based on the backward Monte Carlo tree search model, a set of initial beliefs corresponding to a likelihood of leading to the final belief within a timestep threshold. A policy may be trained using the set of initial beliefs. Based on the trained policy, a set of candidate vehicle control actions may be determined, and a vehicle control action may be selected from the set of candidate vehicle control actions.

Abstract: System and method for responding to queries about objects in a cabin of a vehicle. The system detects a trigger that causes an in-cabin camera to capture video of the cabin, and the system generates a history of captions for at least selected frames of the video by a large multi-modal model (LMM). The system converts a spoken query received by a microphone to a text-based prompt and generates, by the LMM, a response to the prompt based on the history of captions. The response is converted to speech that is output to a speaker.

Abstract: A method for resolving an exception situation in autonomous driving includes receiving telemetry data from an autonomous vehicle (AV); identifying, using the telemetry data, features and feature values; identifying, using the features and the feature values, a solution to the exception situation; determining a confidence level of the solution; in response to the confidence level exceeding a threshold, transmitting the solution to the AV; and in response to the confidence level not exceeding the threshold, forwarding the solution to a mobility manager, obtaining, from the mobility manager, a validated solution, and transmitting the validated solution to the AV.

Abstract: Systems and methods are provide an incentive for a driver to improve power systems, safety systems, and autonomous driving systems of a vehicle. A method includes determining a learning goal for the vehicle. The method includes generating a request based on the learning goal. The method includes calculating a reward value. The method includes displaying a task on a user interface of the vehicle. The task may be based on the request, the reward value, or both. The method includes obtaining sensor data. The sensor data may be obtained based on an initiation of the task. The method includes determining progress of the task. The method includes transmitting a notification based on a determination that the task is completed.

Abstract: System and method for determining a drive mode of a vehicle, such as an eco mode or a sport mode, and retrieving a human-machine interface (HMI) policy based on the drive mode. The HMI policy describes output parameters of various HMI systems of the vehicle, such as an audio system or a haptic system. The system and method further includes causing an HMI system of the vehicle to implement the HMI policy, where such implementation includes generating outputs, such as sounds, lighting patterns, and haptic vibrations, according to the output parameters described by the HMI policy. The system and method may optionally include retrieving a driver profile for a driver of the vehicle, such that the retrieved HMI policy is also based on the driver profile.

Abstract: A vehicle includes a vehicle pedal, a vehicle engine and an electronic control unit. The vehicle engine generates a torque output of the vehicle in accordance with an operation of the vehicle pedal. The electronic control unit controls the vehicle pedal between a normal state, an increased resistance state and a decreased resistance state. The vehicle pedal is less sensitive to driver pressure in the increased resistance state. The vehicle pedal is more sensitive to driver pressure in the decreased resistance state.

Abstract: Decision-making for a vehicle uses a data determining interface between perception and decision-making. The interface receives, from at least two data sources, operational environment data representing objects external to the vehicle while the vehicle is traversing a vehicle transportation network. The operational environment data is modified to determine output data for data types needed to determine a candidate vehicle control action responsive to the distinct vehicle operation scenario identified using the operational environment data. The solution is the same candidate vehicle control action for the same conditions regardless of the data types available from the data sources. That is, even where different data sources are available, consistent output is provided to a decision-making system. Consistent decision-making results whether the vehicle has a high-definition map, a standard-definition map, or no map as a data source, for example.

Abstract: Intelligent eco mode optimization in a battery electric vehicle (BEV) includes collecting data from one or more systems of a vehicle in which the vehicle includes a battery. A predicted route is generated based on the collected data. The collected data includes a navigation map for a portion of a vehicle transportation network. A state of the vehicle is determined based on the collected data and the predicted route. A drive mode is determined, using a decision-making model, for the vehicle based on the state of the vehicle and the predicted route. The drive mode is either a first drive mode having a first acceleration curve or a second drive mode have a second acceleration curve and the second drive mode reduces a rate of discharge of the battery as compared to the first drive mode. The vehicle is set to use the drive mode.

Abstract: Systems and methods for adjusting vehicle components based on audible commands are disclosed herein. In an embodiment, the system includes a plurality of adjustable vehicle components, an audio device, and a controller. The audio device is configured to receive an audible command and generate corresponding command data. The controller is programmed to generate at least one confidence score based on the command data and (i) cause a first response to be output from the audio device after determining that the at least one confidence score does not meet a first threshold, (ii) cause a second response to be output from the audio device after determining that the at least one confidence score does not meet a second threshold, and (iii) cause an adjustment of at least one adjustable vehicle component after determining that the at least one confidence score meets the second threshold.

Abstract: Autonomous vehicle operations use learned constraints over beliefs to traverse a vehicle transportation network. Sensor data and user demonstration data are used to determine a belief path using a partially observable Markov decision process (POMDP) model. The belief path can be updated based on the learned constraints. Candidate actions are determined based on the POMDP model. The candidate actions are constrained by the updated belief path. An action is selected, and the vehicle traverses the vehicle network using the selected action.

Abstract: Monitoring driving characteristics by a first sensor and environmental characteristics by a second sensor, and determining, based on data of the first sensor and data of the second sensor, a deviant driving characteristic and an associated environmental characteristic. Further determining a deviant driving behavior based on the deviant driving characteristic and the associated environmental characteristic, for example, based on repeated instances of the deviant driving characteristic and the associated environmental characteristic. Indicating, to a driver of the vehicle, a corrective action for correcting the deviant driving characteristic, where the indication may be via a user interface of the vehicle or of a mobile device. In some implementations, the driver can review determined deviant driving behaviors via an application and provide to the application a driving goal for focused correction of certain deviant driving behaviors.

Abstract: Vehicle decision-making is analyzed and can be used to modify a decision-making process. For subsets of features comprising a vehicle operational scenario, a first value is generated that quantifies behavior of an artificial intelligence (AI) agent as the AI agent performs a sequence of actions within a first world model based on a complete set of observations for the subset of features. A first world model is a copy of a second world model for sequential decision making. A second value is generated that quantifies behavior of an AI agent as the AI agent performs a sequence of actions in the second world model based on an incomplete set of observations for the subset of features. A difference between the first and second values determines the impact of individual features on the AI agent within the second world model. A decision-making process of the AI agent can be updated.

Abstract: A system may detect a sound associated with a vehicle. The sound may be detected using a set of one or more microphones. The system may use a machine learning model to detect at least one of a squeak or a rattle based on the sound. The machine learning model may be trained using a plurality of sounds emitted in relation to one or more vehicles. The system may determine, based on detecting at least one of the squeak or the rattle, a location of the vehicle associated with at least one of the squeak or the rattle. In some implementations, the location may be determined based on detecting different sound levels between a first microphone and a second microphone, or detecting a time difference between the first and the second microphone, or by correlating the sound to a sound signature associated with a part of the vehicle.