Abstract: Example embodiments described herein therefore relate to an object-model based event detection system that comprises a plurality of sensor devices, to perform operations that include: generating sensor data at the plurality of sensor devices; accessing the sensor data generated by the plurality of sensor devices; detecting an event, or precursor to an event, based on the sensor data, wherein the detected event corresponds to an event category; accessing an object model associated with the event type in response to detecting the event, wherein the object model defines a procedure to be applied by the event detection system to the sensor data; and streaming at least a portion of a plurality of data streams generated by the plurality of sensor devices to a server system based on the procedure, wherein the server system may perform further analysis or visualization based on the portion of the plurality of data streams.

Abstract: A low power communication system provides efficient sensor data transmission using Bluetooth Low Energy (BLE) broadcasts to minimize power consumption while maximizing data flexibility and reliability. The system enables peripherals to transmit sensor data in a connectionless manner, utilizing Service Universally Unique Identifiers (UUIDs) and Tag-Length-Value (TLV) structures, for example.

Abstract: Video and/or security devices include a near field communication (NFC) module configured to communicate with other NPC-capable devices, such as cell phones, user ID tags or badges carried or worn by users. The NPC communications allows users to quickly and easily establish a connection with other devices, such as to track assets, perform configuration and testing, or to perform safety procedures.

Abstract: Disclosed are systems, methods, and non-transitory computer-readable media for modeling images obtained from a vehicle using prompts. The system receives an image from a camera of a vehicle, the image depicting one or more objects. The system receives a prompt defining a target to be detected in the image. The system analyzes the image and data associated with the prompt using a machine learning model to generate a prediction indicating whether the one or more objects depicted in the image match the target defined by the prompt. The system causes a message to be displayed to an operator representing the prediction.

Abstract: Multiple cameras within a vehicle, such as placed at multiple locations inside and/or outside a bus or semi-truck and trailer, may be chained together to provide a consolidated video stream that may be recorded and/or transmitted wirelessly to a cloud storage device. Rather that each of the cameras independently recording and/or streaming video content, each camera may be configured to deserialize an input video to identify video streams from separate cameras (e.g., such as based on associate virtual channel identifiers) and reserialize its own video stream with the other video streams in the received input. The serialized video stream is then transmitted to a next camera in the daisy chain and/or a video storage device.

Abstract: Methods for performing operations for improving driver safety across a fleet of vehicles are disclosed. A plurality of safety events pertaining to a driving of a fleet of vehicles by a plurality of drivers are detected. A subset of the events is identified. The subset corresponds to one or more safety events of the plurality of safety events involving one or more vehicles of the fleet of vehicles to which drivers have not been assigned. A user interface is generated for presentation on a client device, the user interface including an interactive user interface element for accessing the subset of the events. One or more user interface elements are provided for accepting or rejecting recommendations for assignments of one of the plurality of drivers to each of the vehicles. The recommendations are generated based on an application of a machine-learned model to images of faces captured.

Abstract: A vehicle dash cam may be configured to execute one or more neural networks (and/or other artificial intelligence), such as based on input from one or more of the cameras and/or other sensors associated with the dash cam, to intelligently detect safety events in real-time. Detection of a safety event may trigger an in-cab alert to make the driver aware of the safety risk. The dash cam may include logic for determining which asset data to transmit to a backend server in response to detection of a safety event, as well as which asset data to transmit to the backend server in response to analysis of sensor data that did not trigger a safety event. The asset data transmitted to the backend server may be further analyzed to determine if further alerts should be provided to the driver and/or to a safety manager.

Abstract: A method can include receiving, from a safety monitoring device on a vehicle, a raw event. The method can include receiving contextual information associated with the raw event, selecting rules to apply to the raw event, and applying the rules to the raw event. The method can include determining a pre-safety event, selecting, based on the pre-safety event and/or the contextual information, vendor safety event rules and customer safety event rules, determining a conflict between a customer safety event rule and a vendor safety event rule, and applying the vendor safety event rules and the customer safety event rules to the pre-safety event and the contextual information to determine a safety event, wherein the conflicting vendor safety event rule is not applied. The method can include making safety event information available to a user, receiving feedback from the user, and updating a safety event rule engine based on the feedback.

Abstract: A device may include a housing configured to be worn by a user, a sensor configured to detect a safety event, and a low-power communication module configured to transmit a low-power alert signal to a nearby central. The central may be configured to transmit a higher-power signal to a remote backend in response to receiving the low-power alert signal, initiating a safety protocol, such as a notification to a nearby user of a potential safety event.

Abstract: A method performed by an edge device (e.g., a dash camera) is disclosed to automatically trigger a security action. The method includes storing a face embedding of a first person that is expected to be inside a vehicle, capturing, using one or more cameras, one or more images of a face of a second person inside the vehicle, generating a face embedding of the second person based on analyzing the one or more images of the face of the second person using a face embedding machine learning model, determining whether the face of the second person matches a face of the first person based on comparing the face embedding of the second person to the face embedding of the first person using a face verification machine learning model, and triggering a security action in response to determining that the faces do not match.

Abstract: A diagnostic service is selectively started at opportune times, such as when the electrical vehicle is initially plugged in. When the electrical vehicle is turned off, there may not be any messages being sent on the communication bus since most vehicle electronic controllers are disabled to conserve power. But, when someone plugs in a charger, there is a surge of bus traffic for a short period of time. The system parses through the bus traffic and looks for a particular communication bus address. The presence of this address can be used as an indicator for an opportune time to start the diagnostic service to confirm whether the electric vehicle is charging or not and to log the state of charge of the electric vehicle.

Abstract: A system and method for predicting vehicle battery failures includes maintaining a rolling buffer of voltage data from a vehicle and analyzing the data during engine state transitions. The system detects transitions from engine off to on states and analyzes voltage data from a predetermined time period preceding the transition to identify minimum voltage values. These values are compared against configurable threshold voltages to assess battery health. The system generates time series models from historical voltage data to predict future voltage values and potential battery failures. The system generates customized alerts through a graphical user interface indicating predicted time until battery failure and recommended maintenance actions.

Abstract: Disclosed are systems, methods, and non-transitory computer-readable media for customized mobile device management. Mobile devices are used to aid vehicle operators but may also distract vehicle operators from focusing on safe operation of the vehicle. To alleviate this issue, a mobile device management system allows for generation and management of customized device management configurations that cause a mobile device to be configured into a restricted operating mode that restricts selected functionality of the mobile device that may distract the vehicle operator. For example, the restricted operating mode may restrict access to specified applications installed on the mobile device and/or restrict presentations of notifications.

Abstract: An electric vehicle (EV) charge management system is configured to monitor, manage, and optimize EV charging processes. The system can continuously monitor charge-related data from various sources, including the EVs themselves, charging infrastructure providers, schedule data, and/or other relevant systems, to identify charging issues in real-time. Users can define customized charge profiles for each EV or groups of EVs, such as target charge levels (e.g., 80%), low charge levels (e.g., 20%), charging schedules, maximum charging rates, and/or other criteria that are indicative of a potential charging issue. These charge profiles allow users to have greater control over the charging process, ensuring that their EVs are charged in a way that reduces fleet downtime (e.g., charging time in the middle of a driver's shift) and optimizes productivity of the fleet.

Abstract: A system receives vehicle metric data from a gateway device connected to a vehicle. The vehicle gateway device gathers data related to operation of the vehicle and/or location data. The system receives data from multiple vehicles in a fleet. The vehicle gateway device gathers vehicle metric data at a high frequency. Instead of transmitting a large amount of vehicle metric data at a fine level of granularity, the vehicle gateway device aggregates and buckets the vehicle metric data over a period of time (such as, every five minutes). The system uses the bucketed data for fleet management analysis.

Abstract: Systems and methods are provided for vehicle diagnostic aware route planning and navigation. A route management system receives diagnostic data from vehicle diagnostic harnesses and analyzes it along with historical vehicle data to determine vehicle parameters and potential routing conflicts. The system detects vehicle conditions through diagnostic codes or sensor data and determines appropriate route modifications based on factors including vehicle location, service requirements, and delivery schedules. Vehicle attributes are determined through analysis of both VIN data and diagnostic signals to create comprehensive vehicle profiles for routing purposes. When conflicts are detected, the system can modify routes to include service locations or identify replacement vehicles based on compatible configurations.

Abstract: A method and apparatus of utilizing an edge device comprising storing in a memory a pre-installed trained machine learning model, the machine learning model comprising a fixed base portion and a flexible portion. The method further comprises, when the edge device is installed, collecting data using one or more sensors on the edge device, and using the fixed base portion of the machine learning model on the data for a first aspect classification, while using the flexible portion of the machine learning model for a second aspect classification. The method further comprises calibrating the flexible portion of the machine learning model during the use.

Abstract: Systems and methods for providing real-time road condition information using a network of sensor-equipped vehicles. The system includes a plurality of vehicles equipped with dash cameras and other sensors that collect road condition data. A spatiotemporal index maintains processed sensor data associated with road segments, where personally identifiable information is removed and data quality is validated. The system analyzes the sensor data to detect various road network conditions including weather conditions, construction zones, and infrastructure status. The system monitors data freshness and automatically triggers collection of updated data when existing data exceeds age thresholds. User account information customizes the display of road conditions based on vehicle types and regulatory requirements.

Abstract: A mounting adapter for radar-based monitoring devices includes a body formed of radar-transparent material having first and second ends. The first end includes a mounting interface for coupling with a monitoring device, while the second end includes a threaded interface for engaging with vessel openings. A blind end portion between the interfaces maintains vessel sealing. The adapter may include a hexagonal engagement surface with integrated drainage features. The radar-transparent material enables accurate measurements while providing chemical resistance and mechanical integrity.

Abstract: A vehicle device may be configured to execute one or more event models, such as based on input from one or more of the audio capture devices, image capture devices, and/or other sensors associated with the vehicle device, to intelligently detect safety events, such as unsignaled lane departures, in real-time. Detection of a safety event may trigger an in-vehicle alert to make the driver aware of the safety risk. The vehicle device may include logic for determining which asset data to transmit to a backend server in response to detection of a safety event, as well as which asset data to transmit to the backend server in response to analysis of sensor data that did not trigger a safety event. The asset data transmitted to the backend server may be further analyzed to determine if any action should be taken by the user of the backend server.