Abstract: The present invention extends to methods, systems, and computer program products for distinguishing roadway surface lane markings for a vehicle to follow. Automated driving or driving assist vehicles utilize sensors to help the vehicle navigate on roadways or in parking areas. The sensors can utilize, for example, the painted surface markings to help guide the vehicle on its path. When ambiguity is detected between roadway surface markings, decision making algorithms identify a set of roadway surface markings for a vehicle to abide by. The sensors can also identify the location and trajectory of neighboring vehicles to increase confidence with respect to an identified set of roadway surface markings.

Abstract: A method is disclosed for using a camera on-board a vehicle to determine whether precipitation is failing near the vehicle. The method may include obtaining multiple images. Each of the multiple images may be known to photographically depict a “rain” or a “no rain” condition. An artificial neural network may be trained on the multiple images. Later, the artificial neural network may analyze one or more images captured by a first camera secured to a first vehicle. Based on that analysis, the artificial neural network may classify the first vehicle as being in “rain” or “no rain” weather.

Abstract: An autonomous vehicle includes interior sensors including an IR camera and a visible light camera. Images of an interior of the vehicle are captured using the cameras both before and after a passenger rides in the vehicle. The IR images from before and after are subtracted to obtain a difference image. Pixels above a threshold intensity a clustered. Clusters having an above-threshold size are determined to be anomalies. Portions of images from the visible light camera corresponding to the anomalies are sent to a dispatcher, who may then clear the vehicle to pick up another passenger or proceed to a cleaning station. Anomalies may be identified based on a combination of the IR images and visible light images.

Abstract: An autonomous vehicle includes both a LIDAR sensor and a camera. A point cloud from the LIDAR sensor is processed to remove points corresponding to a ground plane and points having a reflectivity below a reflectivity threshold. The remaining points are grouped into clusters. Clusters having points satisfying a flatness threshold are then converted into 2D pixel positions in the output of the camera. Regions of interest including these 2D pixel positions are then analyzed to detect and interpret any road signs present.

Abstract: A method for avoiding interference with a bus. The method includes detecting a bus and obtaining image data from the bus, such as information displayed on the bus. A deep neural network trained on bus images may process the information to associate the bus with a bus route and stop locations. Map data corresponding to the stop locations may also be obtained and used to initiate a lane change or safety response in response to proximity of the bus to a stop location. A corresponding system and computer program product is also disclosed and claimed herein.

Abstract: Example ride control systems and methods are described. In one implementation, a method receives a request to initiate a drive session to help a baby fall asleep in a vehicle. The method identifies a driving route for the drive session and implements the drive session by following the driving route. One or more vehicle-mounted sensors determine whether the baby in the vehicle has fallen asleep during the drive session.

Abstract: A scenario is defined that including models of vehicles and a typical driving environment. A model of a subject vehicle is added to the scenario and sensor locations are defined on the subject vehicle. Perception of the scenario by sensors at the sensor locations is simulated to obtain simulated sensor outputs. The simulated sensor outputs are annotated to indicate the location of obstacles in the scenario. The annotated sensor outputs may then be used to validate a statistical model or to train a machine learning model. The simulates sensor outputs may be modeled with sufficient detail to include sensor noise or may include artificially added noise to simulate real world conditions.

Abstract: A vehicle controller receives images from a camera upon arrival and upon departure. A location of the vehicle may be tracked and images captured by the camera may be tagged with a location. A departure image may be compared to an arrival image captured closest to the same location as the arrival image. A residual image based on a difference between the arrival and departure images is evaluated for anomalies. Attributes of the anomaly such as texture, color, and the like are determined and the anomaly is classified based on the attributes. If the classification indicates an automotive fluid, then an alert is generated. A machine learning algorithm for generating classifications from image data may be trained using arrival and departure images obtained by rendering of a three-dimensional model or by adding simulated fluid leaks to two-dimensional images.

Abstract: Methods, devices and apparatuses pertaining to identifying the candidate road segments for autonomous operations are described. A method may involve obtaining data of a first plurality of road segments that are permitted for one or more autonomous operations. The method may further include generating a road segment template based on the data of the first plurality of road segments, and reconstructing a second plurality of road segments based on the road segment template to obtain a plurality of reconstructed road segments.

Abstract: A vehicle controller receives images from a camera upon arrival and upon departure. A location of the vehicle may be tracked and images captured by the camera may be tagged with a location. A departure image may be compared to an arrival image captured closest to the same location as the arrival image. A residual image based on a difference between the arrival and departure images is evaluated for anomalies. Attributes of the anomaly such as texture, color, and the like are determined and the anomaly is classified based on the attributes. If the classification indicates an automotive fluid, then an alert is generated.

Abstract: A system and method for assisted or autonomous parking of a vehicle is disclosed. The method may begin when the vehicle approaches a feeder lane within a parking lot. At that point, a computer system may decide whether the vehicle should enter the feeder lane. The computer system may use at least one of machine learning, computer vision, and range measurements to determining whether a condition precedent for entering the feeder lane exists. The condition precedent may include an in-bound arrow on the feeder lane or parking lines and/or a parked vehicle adjacent the feeder lane defining a departure angle less than or equal to ninety degrees. If the condition precedent exists, the vehicle may enter the feeder lane. If the condition precedent does not exist, the vehicle may move on to another feeder lane.

Abstract: A method of recommending radio stations including the step of generating a plurality of driver models. Each driver model may correspond to a different driver and documents audio content listened to by that driver while driving. The plurality of driver models may be aggregated to generate a recommendation model correlating radio stations, audio content, and geographic regions. Thereafter, a request for recommendations may be received from a driver while the driver is in a familiar or unfamiliar geographic region. A driver model corresponding to the driver may be compared against other data contained within the recommendation model in order to identify one or more radio stations that present audio content within the geographic region that best matches the audio content documented within the driver model corresponding to the driver. The one or more radio stations may then be communicated to the driver.

Abstract: A controller for an autonomous vehicle receives audio signals from one or more microphones. The audio signals are input to a machine learning model that classifies the source of the audio features. For example, features may be classified as originating from a vehicle. A direction to a source of the audio features is determined based on relative delays of the audio features in signals from multiple microphones. Where audio features are classified with an above-threshold confidence as originating from a vehicle, collision avoidance is performed with respect to the direction to the source of the audio features. The direction to the source of the audio features may be correlated with vehicle images and/or map data to increase a confidence score that the source of the audio features is a parked vehicle with its engine running. Collision avoidance may then be performed with potential paths of the parked vehicle.

Abstract: A controller receives outputs from a plurality of sensors such as a camera, LIDAR sensor, RADAR sensor, and ultrasound sensor, which may be rearward facing. A probability is updated each time a feature in a sensor output indicates presence of an object. The probability may be updated as a function of a variance of the sensor providing the output and a distance to the feature. Where the variance of a sensor is directional, directional probabilities may be updated according to these variances and the distance to the feature. If the probability meets a threshold condition, actions may be taken such as a perceptible alert or automatic braking. The probability may be decayed in the absence of detection of objects. Increasing or decreasing trends in the probability may be amplified by further increasing or decreasing the probability.

Abstract: Systems, methods, and devices for predicting a driver's intention and future movements of a proximal vehicle, whether an automated vehicle or a human driven vehicle, are disclosed herein. A system for predicting future movements of a vehicle includes an intersection component, a camera system, a boundary component, and a prediction component. The intersection component is configured to determine that a parent vehicle is near an intersection. The camera system is configured to capture an image of the proximal vehicle. The boundary component is configured to identify a sub-portion of the image containing a turn signal indicator on the proximal vehicle. The prediction component is configured to predict future movement of the proximal vehicle through the intersection based on a state of the turn signal indicator.

Abstract: A system for detecting and identifying foliage includes a tracking component, a tracking parameters component, and a classification component. The tracking component is configured to detect and track one or more features within range data from one or more sensors. The tracking parameters component is configured to determine tracking parameters for each of the one or more features. The tracking parameters include a tracking age and one or more of a detection consistency and a position variability. The classification component is configured to classify a feature of the one or more features as corresponding to foliage based on the tracking parameters.

Abstract: A controller receives outputs form a plurality of sensors such as a camera, LIDAR sensor, RADAR sensor, and ultrasound sensor. Sensor outputs corresponding to an object are assigned to a tracklet. Subsequent outputs by any of the sensors corresponding to that object are also assigned to the tracklet. A trajectory of the object is calculated from the sensor outputs assigned to the tracklet, such as by means of Kalman filtering. For each sensor output assigned to the tracklet, a probability is updated, such as using a Bayesian probability update. When the probability meets a threshold condition, the object is determined to be present and an alert is generated or autonomous obstacle avoidance is performed with respect to an expected location of the object.

Abstract: Methods, devices and apparatuses pertaining to identifying the candidate road segments for autonomous operations are described. A method may involve obtaining data of a first plurality of road segments that are permitted for one or more autonomous operations. The method may further include generating a road segment template based on the data of the first plurality of road segments, and reconstructing a second plurality of road segments based on the road segment template to obtain a plurality of reconstructed road segments.

Abstract: The present invention extends to methods, systems, and computer program products for using virtual data to test and train parking space detection systems. Aspects of the invention integrate a virtual driving environment with sensor models (e.g., of a radar system) to provide virtual radar data in relatively large quantities in a relatively short amount of time. The sensor models perceive values for relevant parameters of a training data set. Relevant parameters can be randomized in the recorded data to ensure a diverse training data set with minimal bias. Since the driving environment is virtualized, the training data set can be generated alongside ground truth data. The ground truth data is used to annotate true locations, which are used to train a parking space classification algorithms to detect the free space boundaries.

Abstract: The present invention extends to methods, systems, and computer program products for distinguishing roadway surface lane markings for a vehicle to follow. Automated driving or driving assist vehicles utilize sensors to help the vehicle navigate on roadways or in parking areas. The sensors can utilize, for example, the painted surface markings to help guide the vehicle on its path. When ambiguity is detected between roadway surface markings, decision making algorithms identify a set of roadway surface markings for a vehicle to abide by. The sensors can also identify the location and trajectory of neighboring vehicles to increase confidence with respect to an identified set of roadway surface markings.