Abstract: Disclosed are methods, devices, and computer-readable media for detecting lanes and objects in image frames of a monocular camera. In one embodiment, a method is disclosed comprising receiving a plurality of images; identifying a horizon in the plurality of images by inputting the plurality of images into a deep learning (DL) model (either stored on a local device or via a network call); determining one or more camera parameters based on the horizon; and storing or using the camera parameters to initialize a camera.

Abstract: Disclosed are methods, devices, and computer-readable media for detecting lanes and objects in image frames of a monocular camera. In one embodiment, a method is disclosed comprising receiving a plurality of images; identifying a horizon in the plurality of images by inputting the plurality of images into a deep learning (DL) model (either stored on a local device or via a network call); determining one or more camera parameters based on the horizon; and storing or using the camera parameters to initialize a camera.

Abstract: Disclosed are methods, devices, and computer-readable media for detecting lanes and objects in image frames of a monocular camera. In one embodiment, a method is disclosed comprising receiving a sample set of image frames; detecting a plurality of markers in the sample set of image frames using a convolutional neural network (CNN); fitting lines based on the plurality of markers; detecting a plurality of vanishing points based on the lines; identifying a best fitting horizon for the sample set of image frames via a RANSAC algorithm; computing an inverse perspective mapping (IPM) based on the best fitting horizon; and computing a lane width estimate based on the sample set of image frames using the IPM in a rectified view and the parallel line fitting.

Abstract: Disclosed are a multi-task training technique and resulting model for detecting distracted driving. In one embodiment, a method is disclosed comprising inputting a plurality of labeled examples into a multi-task network, the multi-task network comprising: a backbone network, the backbone network generating one or more feature vectors corresponding to each of the labeled examples, and a plurality of prediction heads coupled to the backbone network; minimizing a joint loss based on outputs of the plurality of prediction heads, the minimizing the joint loss causing a change in parameters of the backbone network; and storing a distraction classification model after minimizing the joint loss, the distraction classification model comprising the parameters of the backbone network and parameters of at least one of the prediction heads.

Abstract: Disclosed are a multi-task training technique and resulting model for detecting distracted driving. In one embodiment, a method is disclosed comprising inputting a plurality of labeled examples into a multi-task network, the multi-task network comprising: a backbone network, the backbone network generating one or more feature vectors corresponding to each of the labeled examples, and a plurality of prediction heads coupled to the backbone network; minimizing a joint loss based on outputs of the plurality of prediction heads, the minimizing the joint loss causing a change in parameters of the backbone network; and storing a distraction classification model after minimizing the joint loss, the distraction classification model comprising the parameters of the backbone network and parameters of at least one of the prediction heads.

Abstract: Disclosed are methods, devices, and computer-readable media for detecting lanes and objects in image frames of a monocular camera. In one embodiment, a method is disclosed comprising receiving a sample set of image frames; detecting a plurality of markers in the sample set of image frames using a convolutional neural network (CNN); fitting lines based on the plurality of markers; detecting a plurality of vanishing points based on the lines; identifying a best fitting horizon for the sample set of image frames via a RANSAC algorithm; computing an inverse perspective mapping (IPM) based on the best fitting horizon; and computing a lane width estimate based on the sample set of image frames using the IPM in a rectified view and the parallel line fitting.

Abstract: A computing device is configured to receive information representing an electronic drawing. The information is processed to locate and access the electronic drawing. Moreover, a template configuration file is accessed that corresponds with a category of the electronic drawing. Using the template configuration file, the at least one computing device can generate an item specification that includes information representing items to be reviewed for compliance with at least one rule. For each respective one of the items in the item specification, the respective item is determined to be present in the electronic drawing and complies with at least one rule identified in the template configuration file. Further, information is generated representing whether the respective item is present in the electronic drawing and complies with at least one rule identified in the template configuration file. The generated information is transmitted to at least one computing device.

Abstract: This disclosure presents a method, an apparatus, and a non-transitory computer readable medium to synchronize an active load with a microgrid having a plurality of distributed generators. The method comprises obtaining respective reference frames for the active load and each of the plurality of distributed generators. The method further comprises selecting the reference frame of a first distributed generator as a common reference frame for the microgrid. The method further comprises pooling the active load and the other distributed generators of the plurality of distributed generators on the common reference frame of the microgrid. The method further comprises tuning controller parameters of the active load and the plurality of distributed generators so that predefined grid voltage, frequency, and phase values of the microgrid are maintained.

Abstract: A computing device is configured to receive information representing an electronic drawing. The information is processed to locate and access the electronic drawing. Moreover, a template configuration file is accessed that corresponds with a category of the electronic drawing. Using the template configuration file, the at least one computing device can generate an item specification that includes information representing items to be reviewed for compliance with at least one rule. For each respective one of the items in the item specification, the respective item is determined to be present in the electronic drawing and complies with at least one rule identified in the template configuration file. Further, information is generated representing whether the respective item is present in the electronic drawing and complies with at least one rule identified in the template configuration file. The generated information is transmitted to at least one computing device.

Abstract: A method of cleaning a conducting film containing tin oxide from an insulating surface of an item for use in electroplating applications, comprises the steps of immersing the item in a cleaning fluid and irradiating the immersed item with light of wavelength in the range 100 nm-450 nm.

Abstract: Disclosed are a multi-task training technique and resulting model for detecting distracted driving. In one embodiment, a method is disclosed comprising inputting a plurality of labeled examples into a multi-task network, the multi-task network comprising: a backbone network, the backbone network generating one or more feature vectors corresponding to each of the labeled examples, and a plurality of prediction heads coupled to the backbone network; minimizing a joint loss based on outputs of the plurality of prediction heads, the minimizing the joint loss causing a change in parameters of the backbone network; and storing a distraction classification model after minimizing the joint loss, the distraction classification model comprising the parameters of the backbone network and parameters of at least one of the prediction heads.

Abstract: Disclosed are methods, devices, and computer-readable media for detecting lanes and objects in image frames of a monocular camera. In one embodiment, a method is disclosed comprising receiving a plurality of images; identifying a horizon in the plurality of images by inputting the plurality of images into a deep learning (DL) model (either stored on a local device or via a network call); determining one or more camera parameters based on the horizon; and storing or using the camera parameters to initialize a camera.

Abstract: Described are embodiments for training and using a close following classifier. In the example embodiments, a system includes a backbone network configured to receive an image; and at least one prediction head communicatively coupled to the backbone network, the at least one prediction head configured to receive an output from the backbone network, wherein the at least one prediction head includes a classifier configured to classify the image as including a close-following event, the classifier receiving the output of the backbone network and a vehicle speed as inputs.

Abstract: Disclosed are a multi-task training technique and resulting model for detecting distracted driving. In one embodiment, a method is disclosed comprising inputting a plurality of labeled examples into a multi-task network, the multi-task network comprising: a backbone network, the backbone network generating one or more feature vectors corresponding to each of the labeled examples, and a plurality of prediction heads coupled to the backbone network; minimizing a joint loss based on outputs of the plurality of prediction heads, the minimizing the joint loss causing a change in parameters of the backbone network; and storing a distraction classification model after minimizing the joint loss, the distraction classification model comprising the parameters of the backbone network and parameters of at least one of the prediction heads.

Abstract: Disclosed are a multi-task training technique and resulting model for detecting distracted driving. In one embodiment, a method is disclosed comprising inputting a plurality of labeled examples into a multi-task network, the multi-task network comprising: a backbone network, the backbone network generating one or more feature vectors corresponding to each of the labeled examples, and a plurality of prediction heads coupled to the backbone network; minimizing a joint loss based on outputs of the plurality of prediction heads, the minimizing the joint loss causing a change in parameters of the backbone network; and storing a distraction classification model after minimizing the joint loss, the distraction classification model comprising the parameters of the backbone network and parameters of at least one of the prediction heads.

Abstract: Improvements to intracardiac devices such as intracardiac blood pump assemblies, and associated methods. In one example, the present technology includes systems and methods for pacing the heart, and/or performing cardiac ablation using electrodes mounted on a portion of the intracardiac device. In another example, the present technology includes systems and methods for detecting mural thrombi in a patient's heart using electrical sensors or ultrasonic phased arrays mounted on the intracardiac device. In another example, the present technology includes systems and methods for detecting tissue changes and reactions in heart tissue during treatment using one or more temperature sensors. In another example, the present technology includes an improved distal tip for use with an intracardiac device.

Abstract: A method for improving an autonomous microgrid, and an autonomous microgrid that includes a plurality of inverter-based distributed generations. Each of the inverter-based distributed generations is coupled to a corresponding power droop controller, a corresponding voltage controller, and a corresponding current controller. The autonomous microgrid further includes a constant power load (CPL) coupled to one of the plurality of inverted-based distributed generations. The CPL includes a phase locked loop (PLL), a DC voltage controller and an AC current controller. Power-sharing coefficients, controller parameters of the controllers and gains of the PLL are defined based on a weighted objective function that is calculated through on a particle swarm optimization.

Abstract: Systems and methods for determining the positioning of intracardiac devices, such as intracardiac blood pump assemblies, using electrical sensors configured to sense electrical potential as it propagates through the heart. In one example, the present technology provides an intracardiac device with one or more electrical sensors mounted thereon, and a controller configured to determine the absolute or relative location of the intracardiac device based on the timing, shape, and/or amplitude of the electrical signals received from the one or more sensors.

Abstract: A personal emergency notification system which is incorporated into a badge holder or similar item. This allows for a user to have an emergency notification system attached to a badge, keycard, or other similar access pass that a user will otherwise need to have on their person at work or in another specific environment but otherwise is not necessary. Further, the system can have various automatic features when an emergency is detected in the work area, the system may also be able to recognize when the user is on site, near the site (e.g. in a mustering area or rally point), or offsite and safe. These automatic features may also be used during an evacuation drill to obtain valuable data on area evacuation.

Abstract: A computer-implemented device-to-device (D2D) communication method includes broadcasting, by a device, device information of the device at periodic time intervals using device-to-device (D2D) communication, receiving, by the device, the device information in an incoming new device broadcast sequence, and receiving, by the device, compiled device information in an incoming D2D broadcast sequence. The compiled device information includes device information of respective devices of a plurality of devices communicating using the D2D communication. The plurality of devices include the device and one or more additional devices. The computer-implemented method further includes storing, by the device, the compiled device information, determining, by the device, a broadcast sequence based on the compiled device information, and broadcasting, by the device, compiled device information of at least one device of the plurality of devices according to the determined broadcast sequence.