Abstract: Methods and systems for training an audio-based machine learning model to predict a health condition based on biological sounds emitted by a person. Audio data corresponding to biological sounds produced by the person is generated from a microphone. The audio data is segmented into a plurality of segments, each segment associated with a respective sound event. An audio-based machine learning model is executed on the plurality of segments. The audio-based machine learning model is configured to output, for each segment, a label of a medical condition and an associated a confidence score. The model is trained via active learning, in which a subset of the plurality of segments are selected based on their confidence score being below a threshold, and provided to a human for annotation.

Abstract: Systems and methods for detecting symptoms of occupant illness is disclosed herein. In embodiments, a storage is configured to maintain a visualization application and data from one or more sources, such as an image source. A processor is in communication with the storage and a user interface. The processor is programmed to receive data from the one or more sources, execute human-detection models based on the received data, execute activity-recognition models to recognize symptoms of illness based on the data from the one or more sources, determine a location of the recognized symptoms, and execute a visualization application to display information in the user interface. The visualization application can show a background image with an overlaid image that includes an indicator for each location of recognized symptom of illness. Additionally, data from the audio source, image source, and/or radar source can be fused.

Abstract: Systems and methods for detecting symptoms of occupant illness is disclosed herein. In embodiments, a storage is configured to maintain a visualization application and data from one or more sources, such as an image source. A processor is in communication with the storage and a user interface. The processor is programmed to receive data from the one or more sources, execute human-detection models based on the received data, execute activity-recognition models to recognize symptoms of illness based on the data from the one or more sources, determine a location of the recognized symptoms, and execute a visualization application to display information in the user interface. The visualization application can show a background image with an overlaid image that includes an indicator for each location of recognized symptom of illness. Additionally, data from the audio source, image source, and/or radar source can be fused.

Abstract: A workflow is described herein for training a radar data processing model with a reduced requirement for annotated training data. The training workflow leverages an image sensor, such as a camera, that is synchronized with a radar sensor to capture synchronous image and radar point cloud data. The training workflow utilizes a self-supervised knowledge distillation process to pre-train the radar data processing model on the unannotated synchronous image and radar point cloud data, using a pre-trained image model. Subsequently, the radar data processing model is fine-tuned with a limited set of annotated radar point cloud data, thereby greatly reducing the human annotation burden.

Abstract: A system includes a first and second transceiver configured to emit and receive a wireless signal, a controller in communication with the first and second transceiver, the controller configured to send instructions to the transceivers to initiate a Fine Time Measurement (FTM) on a wireless channel to a body part associated with a user during a healthy state of the user, store an FTM measurement associated with the healthy state of the user, send instructions to the transceiver to initiate the FTM on the wireless channel during an infected state of the user, store an FTM measurement associated with the infected state of the user, compare the FTM measurement associated with the healthy state of the user with the FTM measurement associated with the infected state of the user, and in response to the comparison exceeding a threshold value, output a notification.

Abstract: A vehicle system that includes a vehicle seat, wherein the vehicle seat includes a seat-back portion, a seat-bottom portion, a head rest portion, wherein the vehicle seat includes one or more acoustic sensors configured to retrieve an acoustic signal associated with a passenger seat, and a processor in communication with at least the one or more acoustic sensors, wherein the processor is programmed to identify an anomaly associated with the passenger utilizing the acoustic signal, wherein the anomaly is identified via pre-processing the acoustic signal and extracting one or more features associated with the acoustic signal in response to the pre-preprocessing and utilizing a classifier to classify one or more features associated with the acoustic signal as either a normal condition or the anomaly, and output a notification associated with the anomaly in response to identifying the anomaly.

Abstract: A system that includes one or more sensors installed in proximity to a machine configured to collect raw signals associated with an environment of the machine, are multi-layer spatial data that include time-stamp data. The system may include a processor in communication with the sensors and programmed to receive one or more raw signals, denoise the one or more raw signals to obtain a pre-processed signal, extract one or more features from the pre-processed signals, classify the one or more features to an associated class, wherein the associated class includes one or more of a normal class, abnormal class, or a potential-abnormal class, create fusion data by fusing the one or more features utilizing the associated class and the time-stamp data, and output a heat map on an overlaid image of the environment.

Abstract: A system that includes one or more sensors installed in proximity to a machine configured to collect raw signals associated with an environment of the machine, are multi-layer spatial data that include time-stamp data. The system may include a processor in communication with the sensors and programmed to receive one or more raw signals, denoise the one or more raw signals to obtain a pre-processed signal, extract one or more features from the pre-processed signals, classify the one or more features to an associated class, wherein the associated class includes one or more of a normal class, abnormal class, or a potential-abnormal class, create fusion data by fusing the one or more features utilizing the associated class and the time-stamp data, and output a heat map on an overlaid image of the environment.

Abstract: A computer-implemented system and method relates to travel time estimation using wireless signals. First communication data is received from a first wireless transceiver at a first location. The first communication data includes identification data of a mobile communication device. The second communication data includes second communication data from a second wireless transceiver at a second location. The second communication data includes the identification data of the mobile communication device. The first communication data and the second communication data, which are associated with the mobile communication device, are extracted using the identification data. Travel time data is generated for the mobile communication device with respect to a travel segment, defined between the first location and the second location, using the first communication data and the second communication data. The aggregated travel time data is generated for the travel segment.

Abstract: A system that includes one or more sensors installed in proximity to a machine configured to collect raw signals associated with an environment of the machine, are multi-layer spatial data that include time-stamp data. The system may include a processor in communication with the sensors and programmed to receive one or more raw signals, denoise the one or more raw signals to obtain a pre-processed signal, extract one or more features from the pre-processed signals, classify the one or more features to an associated class, wherein the associated class includes one or more of a normal class, abnormal class, or a potential-abnormal class, create fusion data by fusing the one or more features utilizing the associated class and the time-stamp data, and output a heat map on an overlaid image of the environment.

Abstract: Systems and methods for detecting symptoms of occupant illness is disclosed herein. In embodiments, a storage is configured to maintain a visualization application and data from one or more sources, such as an image source. A processor is in communication with the storage and a user interface. The processor is programmed to receive data from the one or more sources, execute human-detection models based on the received data, execute activity-recognition models to recognize symptoms of illness based on the data from the one or more sources, determine a location of the recognized symptoms, and execute a visualization application to display information in the user interface. The visualization application can show a background image with an overlaid image that includes an indicator for each location of recognized symptom of illness. Additionally, data from the audio source, image source, and/or radar source can be fused.

Abstract: In one embodiment a mobile device includes a long-range transceiver, a medium-range transceiver, and a controller. The controller is configured to receive, from the long-range transceiver, a target ID associated with a remote medium-range transceiver of a remote system, transmit, via the long-range transceiver, an ID, channel, and band of the medium-range transceiver of the remote system, receive packets from the remote medium-range transceiver on the channel, extract received signal strength indicator (RSSI) data from the packets, filter the RSSI data to obtain a maximum RSSI signal within a window of time, in response to the maximum RSSI signal exceeding a threshold, output a signal indictive of the remote system being less than a predetermined distance away.

Abstract: In one embodiment, a system for a vehicle includes a radio frequency (RF) transceiver having an identification (ID), and a processor coupled with the RF transceiver. The processor is configured to receive a request, via a first wireless connection, for the vehicle to travel to a location, in response to the vehicle being less than a predetermined distance from the location, receive RF packets, via a second wireless connection, from a target at the location, identify packets based on the ID of the RF transceiver, extract received signal strength indicator (RSSI) data from received signals associated with the identified packets, filter the RSSI data to obtain a maximum RSSI signal within a window of time, and in response to the maximum RSSI signal exceeding a threshold, output a signal indictive of the target being less than a predetermined distance from the vehicle.

Abstract: A vehicle side target location method includes receiving a request, via a first wireless connection, for the vehicle to travel to a location, in response to the vehicle being less than a predetermined distance from the location, receiving RF packets, via a second wireless connection having a multiple antenna radio frequency (RF) transceiver having an identification (ID), from a target at the location, identifying packets based on the ID of the RF transceiver, extracting channel state information (CSI) from received signals associated with the identified packets, determining a phase difference of subcarrier phase data of the received signals between each of the multiple antennae, filtering noise of the phase difference of subcarriers based on subcarrier selection to obtain multiple robust phase difference signals, and feeding the multiple robust phase difference signals to a classifier to obtain a side of the vehicle associated with the location of the target.

Abstract: In one embodiment, a vehicle side target location method includes receiving a request, via a first wireless connection, for the vehicle to travel to a location, in response to the vehicle being less than a predetermined distance from the location, receiving RF packets, via a second wireless connection having a multiple antenna radio frequency (RF) transceiver having an identification (ID), from a target at the location, identifying packets based on the ID of the RF transceiver, extracting channel state information (CSI) from received signals associated with the identified packets, determining an amplitude difference of subcarriers of the received signals between each of the multiple antennae, filtering noise of the amplitude difference of subcarriers based on subcarrier selection to obtain multiple robust amplitude difference signals, and feeding the multiple robust amplitude difference signals to a classifier to obtain a side of the vehicle associated with the location of the target.

Abstract: In one embodiment, a mobile device includes a long-range transceiver, a medium-range transceiver, and a controller. The controller is configured to receive, from the long-range transceiver, a medium-range ID associated with a remote medium-range transceiver, transmit, via the long-range transceiver, an ID, channel, and band of the medium-range transceiver, connect, via the medium-range transceiver, to a remote medium-range transceiver based on the medium-range ID, transmit, via the medium-range transceiver, a beacon of data packets at an interval.

Abstract: A vehicle system that includes a vehicle seat, wherein the vehicle seat includes a seat-back portion, a seat-bottom portion, a head rest portion, wherein the vehicle seat includes one or more acoustic sensors configured to retrieve an acoustic signal associated with a passenger seat, and a processor in communication with at least the one or more acoustic sensors, wherein the processor is programmed to identify an anomaly associated with the passenger utilizing the acoustic signal, wherein the anomaly is identified via pre-processing the acoustic signal and extracting one or more features associated with the acoustic signal in response to the pre-preprocessing and utilizing a classifier to classify one or more features associated with the acoustic signal as either a normal condition or the anomaly, and output a notification associated with the anomaly in response to identifying the anomaly.

Abstract: Systems and methods for detecting symptoms of occupant illness is disclosed herein. In embodiments, a storage is configured to maintain a visualization application and data from one or more sources, such as an audio source, an image source, and/or a radar source. A processor is in communication with the storage and a user interface. The processor is programmed to receive data from the one or more sources, execute human-detection models based on the received data, execute activity-recognition models to recognize symptoms of illness based on the data from the one or more sources, determine a location of the recognized symptoms, and execute a visualization application to display information in the user interface. The visualization application can show a background image with an overlaid image that includes an indicator for each location of recognized symptom of illness. Additionally, data from the audio source, image source, and/or radar source can be fused.

Abstract: A depth-based object-detection convolutional neural network is disclosed. The depth-based object-detection convolutional neural network described herein incorporates a base network and additional structure. The base network is configured to receive a depth image formatted as RGB image data as input, and compute output data indicative of at least one feature of an object in the RGB image data. The additional structure is configured to receive the output data of the base network as input, and compute predictions of the location of a region in the received depth image that includes the object and of a class of the object as output. An object detection device incorporating the depth-based object-detection convolutional neural network is operable in real time using an embedded GPU.

Abstract: A thermal state of a plurality of zones of the building is updated according to a building thermal model and information received from temperature sensors of the building. Predicted occupant counts for an upcoming plurality of time slots for each of the plurality of zones are updated using actual occupancy counts for each of the plurality of zones. A misprediction type distribution for the upcoming plurality of time slots for each of the plurality of zones is updated, the misprediction type distribution indicating misprediction for true negatives, false positives, false negatives, and true positives. A total misprediction cost expectation is updated according to the predicted occupant counts and the misprediction type distribution. HVAC power for each of the plurality of zones is determined to optimize occupant thermal comfort weighted according to the predicted occupant counts while minimizing the total misprediction cost expectation. HVAC operation is controlled per the HVAC power.