Abstract: The present disclosure relates to techniques to implement a vehicle action using a distributed model distributed across the vehicle and a remote node. A local portion of the distributed model at the vehicle may generate a local output model based on vehicle event data collected at the vehicle. The local output model may be sent from the vehicle at a first location to the remote node at a second location. The remote node may generate a remote output model based on the local output model using the remote portion of the distributed model. The vehicle action may be determined based on inspecting a reconstructed version of the vehicle event data included in the remote output model. The determined vehicle action may be implemented at the vehicle. The distributed model may facilitate the transmission of vehicle event data across multiple locations while securing the transmission of personally-identifiable information.

Abstract: A method for determining distraction of a driver of a vehicle, including sampling sensor measurements at an onboard system of the vehicle; generating an output indicative of a distracted state; determining that the driver of the vehicle is characterized by the distracted state, based on the output; generating, at a second distraction detection module of a remote computing system, a second output indicative that the driver is characterized by the distracted state, based on the sensor measurements; computing a distraction score, at a scoring module of the remote computing system, in response to generating the second output and based on the sensor measurements and the distracted state.

Abstract: The present embodiments relate to efficient localization of a vehicle within a lane region. An imaging device onboard the vehicle may capture an image stream depicting an environment surrounding the vehicle. An image may be inspected to identify pixel bands indicative of lane markings of a road depicted in the image. Based on the identified pixel bands, a lane region indicating a lane of a roadway can be extracted. A lane drift metric may be generated that indicates an offset of the vehicle relative to a center of the lane region. An output action may be initiated based on the offset indicated in the lane drift metric. The lane region can be translated from a first frame to a second frame providing a top-down perspective of the vehicle within the lane region using a transformation matrix to assist in efficiently deriving the lane drift metric.

Abstract: Systems and methods for driving. A driving data set for each of plurality of human-driven vehicles is determined. For each driving data set, exterior scene features of an exterior scene of the respective vehicle are extracted from the exterior image data. A driving response model is trained based on the exterior scene features and the vehicle control inputs from the selected driving data sets.

Abstract: A method for determining event data including: sampling a first data stream within a first time window at a first sensor of an onboard vehicle system coupled to a vehicle, extracting interior activity data from the first data stream; determining an interior event based on the interior activity data; sampling a second data stream within a second time window at a second sensor of the onboard vehicle system; extracting exterior activity data from the second image stream; determining an exterior event based on the exterior activity data; correlating the exterior event and the interior event to generate combined event data; automatically classifying the combined event data to generate an event label; and automatically labeling the first time window of the first data stream and the second time window of the second data stream with the combined event label to generate labeled event data.

Abstract: A method for determining event data including: sampling a first data stream within a first time window at a first sensor of an onboard vehicle system coupled to a vehicle, extracting interior activity data from the first data stream; determining an interior event based on the interior activity data; sampling a second data stream within a second time window at a second sensor of the onboard vehicle system; extracting exterior activity data from the second image stream; determining an exterior event based on the exterior activity data; correlating the exterior event and the interior event to generate combined event data; automatically classifying the combined event data to generate an event label; and automatically labeling the first time window of the first data stream and the second time window of the second data stream with the combined event label to generate labeled event data.

Abstract: A method for safe route determination, including determining a geographic risk map for a geographic region, wherein the geographic risk map is determined based on a vehicle event dataset aggregated from a plurality of vehicles enabled with an onboard vehicle system, and automatically determining a route between two locations in the geographic region based on the geographic risk map.

Abstract: A system and method is provided for recommending food items based on a set of instructions. A first set of instructions are executed to receive a first set of input parameters associated with plurality of attributes of the entity. Further, a second set of input parameters are received from a second entity and are associated with the first set of input parameters of the entity. The received first set of input parameters and the received second set of input parameters are analyzed to determine at least one of a health label for the entity. Then, a health score is assigned for at least one of the health label for the entity. The health score is assigned based on a food item to be recommended. Upon, the assigned health score lying within a predefined threshold, the food item is recommended to the entity.

Abstract: A method for determining distraction of a driver of a vehicle, including sampling sensor measurements at an onboard system of the vehicle; generating an output indicative of a distracted state; determining that the driver of the vehicle is characterized by the distracted state, based on the output; generating, at a second distraction detection module of a remote computing system, a second output indicative that the driver is characterized by the distracted state, based on the sensor measurements; computing a distraction score, at a scoring module of the remote computing system, in response to generating the second output and based on the sensor measurements and the distracted state.

Abstract: A method for safe route determination, including determining a geographic risk map for a geographic region, wherein the geographic risk map is determined based on a vehicle event dataset aggregated from a plurality of vehicles enabled with an onboard vehicle system, and automatically determining a route between two locations in the geographic region based on the geographic risk map.

Abstract: A method for near-collision detection, including determining a risk map for a vehicle and automatically detecting a near-collision event with an object based on vehicle behavior relative to the risk map.

Abstract: A method for safe route determination, including determining a geographic risk map for a geographic region, wherein the geographic risk map is determined based on a vehicle event dataset aggregated from a plurality of vehicles enabled with an onboard vehicle system, and automatically determining a route between two locations in the geographic region based on the geographic risk map.

Abstract: A method for determining distraction of a driver of a vehicle, including sampling sensor measurements at an onboard system of the vehicle; generating an output indicative of a distracted state; determining that the driver of the vehicle is characterized by the distracted state, based on the output; generating, at a second distraction detection module of a remote computing system, a second output indicative that the driver is characterized by the distracted state, based on the sensor measurements; computing a distraction score, at a scoring module of the remote computing system, in response to generating the second output and based on the sensor measurements and the distracted state.

Abstract: The present embodiments relate to efficient localization of a vehicle within a lane region. An imaging device onboard the vehicle may capture an image stream depicting an environment surrounding the vehicle. An image may be inspected to identify pixel bands indicative of lane markings of a road depicted in the image. Based on the identified pixel bands, a lane region indicating a lane of a roadway can be extracted. A lane drift metric may be generated that indicates an offset of the vehicle relative to a center of the lane region. An output action may be initiated based on the offset indicated in the lane drift metric. The lane region can be translated from a first frame to a second frame providing a top-down perspective of the vehicle within the lane region using a transformation matrix to assist in efficiently deriving the lane drift metric.

Abstract: The present disclosure relates to techniques to implement a vehicle action using a distributed model distributed across the vehicle and a remote node. A local portion of the distributed model at the vehicle may generate a local output model based on vehicle event data collected at the vehicle. The local output model may be sent from the vehicle at a first location to the remote node at a second location. The remote node may generate a remote output model based on the local output model using the remote portion of the distributed model. The vehicle action may be determined based on inspecting a reconstructed version of the vehicle event data included in the remote output model. The determined vehicle action may be implemented at the vehicle. The distributed model may facilitate the transmission of vehicle event data across multiple locations while securing the transmission of personally-identifiable information.

Abstract: Systems and methods for distributed event detection. Sensor data is synchronized, and an on-board detector included in a vehicle detects an event from the synchronized sensor data. The synchronized sensor data is transmitted to a remote system. At the remote system, a remote detector detects the event and generates a remote label from at least a data subset of the synchronized sensor data.

Abstract: A method for determining event data including: sampling a first data stream within a first time window at a first sensor of an onboard vehicle system coupled to a vehicle, extracting interior activity data from the first data stream; determining an interior event based on the interior activity data; sampling a second data stream within a second time window at a second sensor of the onboard vehicle system; extracting exterior activity data from the second image stream; determining an exterior event based on the exterior activity data; correlating the exterior event and the interior event to generate combined event data; automatically classifying the combined event data to generate an event label; and automatically labeling the first time window of the first data stream and the second time window of the second data stream with the combined event label to generate labeled event data.

Abstract: A method for determining event data including: sampling a first data stream within a first time window at a first sensor of an onboard vehicle system coupled to a vehicle, extracting interior activity data from the first data stream; determining an interior event based on the interior activity data; sampling a second data stream within a second time window at a second sensor of the onboard vehicle system; extracting exterior activity data from the second image stream; determining an exterior event based on the exterior activity data; correlating the exterior event and the interior event to generate combined event data; automatically classifying the combined event data to generate an event label; and automatically labeling the first time window of the first data stream and the second time window of the second data stream with the combined event label to generate labeled event data.

Abstract: A method for digital environment reconstruction, including: capturing data using an onboard vehicle system; determining a vehicle event and an event class of the vehicle event; extracting a set of agent parameters from the data; generating a virtual model based on the set of agent parameters; and providing the virtual model to a remote system.

Abstract: Systems and methods for driving. A driving data set for each of plurality of human-driven vehicles is determined. For each driving data set, exterior scene features of an exterior scene of the respective vehicle are extracted from the exterior image data. A driving response model is trained based on the exterior scene features and the vehicle control inputs from the selected driving data sets.