Abstract: A method includes receiving, from a mobile device disposed within a vehicle, a set of sensor measurements collected from an accelerometer of the mobile device during a first time period and converting the set of sensor measurements into a frequency domain. The method also includes filtering the set of sensor measurements to eliminate high frequency sensor measurements and defining a set of contiguous windows based on a remaining sensor measurements in the set of sensor measurements. Each contiguous window of the set of contiguous windows represents a contiguous portion of the remaining sensor measurements. The method further includes generating, for each contiguous window of the set of contiguous windows, a set of features by resampling the remaining sensor measurements of the contiguous window at one or more predefined frequencies and generating an estimated speed of the vehicle during the first time period using the set of features.

Abstract: Techniques for using machine learning-based continuous monitoring to improve crash detection and response system accuracy are provided. In examples, a driving event is detected as it occurs at a first time from driving data collected by a plurality of sensors of a mobile device disposed in a vehicle during a trip. After a second time, a crash prediction model is executed on driving data leading up to the second time to generate a first crash classification. In response to determining that the first crash classification meets first criterion, an escalation action sequence is initiated based on a severity of the crash prediction. After a third time, the crash model is executed on driving data leading up to the third time to generate a second classification, from which it is determined whether to terminate the escalation action sequence prior to an automatic execution of an action.

Abstract: A method includes receiving, from a mobile device disposed within a vehicle, a set of sensor measurements collected from an accelerometer of the mobile device during a first time period and converting the set of sensor measurements into a frequency domain. The method also includes filtering the set of sensor measurements to eliminate high frequency sensor measurements and defining a set of contiguous windows based on a remaining sensor measurements in the set of sensor measurements. Each contiguous window of the set of contiguous windows represents a contiguous portion of the remaining sensor measurements. The method further includes generating, for each contiguous window of the set of contiguous windows, a set of features by resampling the remaining sensor measurements of the contiguous window at one or more predefined frequencies and generating an estimated speed of the vehicle during the first time period using the set of features.

Abstract: A method includes receiving, from a mobile device disposed within a vehicle, a set of sensor measurements collected from an accelerometer of the mobile device during a first time period and converting the set of sensor measurements into a frequency domain. The method also includes filtering the set of sensor measurements to eliminate high frequency sensor measurements and defining a set of contiguous windows based on a remaining sensor measurements in the set of sensor measurements. Each contiguous window of the set of contiguous windows represents a contiguous portion of the remaining sensor measurements. The method further includes generating, for each contiguous window of the set of contiguous windows, a set of features by resampling the remaining sensor measurements of the contiguous window at one or more predefined frequencies and generating an estimated speed of the vehicle during the first time period using the set of features.

Abstract: Apparatuses and methods for predicting a crash using estimated vehicle speed. A set of sensor measurements are received from a mobile device disposed within a vehicle. A set of contiguous windows based on the sensor measurements may be defined. Each contiguous window represents a contiguous portion of the sensor measurements. A set of sensor measurements may be defined for each contiguous window. A trained neural network may execute, using the set of features, to generate one or more speed predictions. A vehicle crash prediction may be generated using the speed prediction. The vehicle crash prediction may then be transmitted to a remote device.

Abstract: Techniques are disclosed for virtual tagging of vehicles that include generating an association between a user of a mobile device and the mobile device. The techniques include receiving a first set of measurements from one or more sensors of the mobile device while the mobile device is positioned in a first vehicle during a trip and training a machine-learning model using the first set of measurements. The techniques further include receiving a second set of measurements from the one or more sensors of the mobile device and determining, by executing the machine-learning model using the third set of measurements, that the mobile device is positioned in the first vehicle or a second vehicle.

Abstract: Techniques are disclosed for predicting a confidence of a total loss event. A mobile device detects a crash event using one or more sensors of a mobile device. The mobile device records a first set of data from the one or more sensors of the mobile device. The mobile device generates a first feature vector including the first set of data and vehicle data that includes an identifier of a vehicle. The mobile device generates a second feature vector using the first set of data and additional data types. The mobile device predicts a confidence of a total loss event by generating a first confidence value from a first machine-learning model using the first feature vector and a second confidence value from a second machine-learning model using the second feature vector.

Abstract: Apparatuses and methods for predicting a crash using estimated vehicle speed. A set of sensor measurements are received from a mobile device disposed within a vehicle. A set of contiguous windows based on the sensor measurements may be defined. Each contiguous window represents a contiguous portion of the sensor measurements. A set of sensor measurements may be defined for each contiguous window. A trained neural network may execute, using the set of features, to generate one or more speed predictions. A vehicle crash prediction may be generated using the speed prediction. The vehicle crash prediction may then be transmitted to a remote device.