Abstract: A vehicle includes a door, an inside door handle, and a door opening warning system having an object sensor that detects an object outside the vehicle in a door collision risk zone, a touchless sensor that detects if a part of a user is positioned in an opening zone at the inside door handle and a touch sensor that detects if the user touches the inside door handle, and a control unit. The control unit triggers a first alarm if the object sensor detects an object in the risk zone and the touchless sensor detects a part of the user in the opening zone at the inside door handle, and triggers a second alarm if the object sensor detects an object in the risk zone and the touch sensor detects that the user is touching the inside door handle.

Abstract: An edge device may receive, from a first device in a first vehicle moving towards an intersection, first-vehicle-provided-tracking information and, from a first sensor device at the intersection, sensor-provided-first-vehicle-tracking information. The edge device may receive, from a second device in a second vehicle moving towards the intersection, second-vehicle-provided-tracking information and, from a second sensor device at the intersection, sensor-provided-second-vehicle-tracking information. The edge device may determine whether vehicle-provided-tracking information matches the sensor-provided-tracking information for the first vehicle and/or the second vehicle. The edge device may determine whether the first vehicle and the second vehicle are predicted to collide and may provide, to one or more traffic control devices, one or more instructions to provide signals to the first vehicle and/or the second vehicle.

Abstract: A vehicular control system includes a plurality of cameras, a plurality of sensors and a central electronic control unit. Image data captured by at least one of the cameras and sensor data sensed by at least one of the sensors is provided to and processed at the central electronic control unit. Data relevant to a current geographical location of the vehicle is provided to and processed at the central electronic control unit. The central electronic control unit is operable to at least partially control the vehicle. The central electronic control unit may include a threat recognizer and a risk assessor. Responsive at least in part to threat recognition/evaluation by the threat recognizer and risk assessment by the risk assessor, the central electronic control unit may control at least one selected from the group consisting of (i) braking of the equipped vehicle and (ii) steering of the equipped vehicle.

Abstract: Systems and methods limit communication of redundant telematics data in V2X systems. A communications management device receives telematics data from multiple sources in a service area and calculates a trajectory each of the objects identified by the telematics data. The communications management device selects high vulnerability trajectories based on the calculated trajectories and identifies when the telematics data from different sources, of the multiple sources, corresponds to a same vehicle. When duplicate sources are determined to provide tracking data corresponding to the same vehicle, the communications management device reports (to a collision avoidance system) the tracking data from only the most accurate of the duplicate sources.

Abstract: A collision avoidance system of a vehicle includes a sensor configured to be disposed at a vehicle for sensing exterior and forwardly of the vehicle. A processor is operable to process sensor data captured by the sensor to determine the presence of a pedestrian ahead of the vehicle and at or moving towards a path of travel of the vehicle. The processor determines a time to intersection of projected paths of travel of the vehicle and pedestrian based on a determined distance to the pedestrian and determined speed of the pedestrian and speed of the vehicle. The system adjusts the speed of the vehicle so that the pedestrian will not be in the projected path of travel of the vehicle when the vehicle arrives where the projected path of travel of the vehicle intersects the projected path of travel of the pedestrian.

Abstract: A propulsion control system for a marine vessel includes a plurality of propulsion devices steerable to propel the marine vessel, at least one proximity sensor that determines a relative position of the marine vessel with respect to an object, wherein the at least one proximity sensor has a field of view (FOV). A controller is configured to identify a trigger condition for expanding the FOV of the at least one proximity sensor and control thrust and/or steering position of at least one of the plurality of propulsion devices to expand the FOV of the at least one proximity sensor by inducing a roll movement or a pitch movement of the marine vessel.

Abstract: A method for controlling a distance of a vehicle to a preceding vehicle, characterized in that, in a step of limiting, the distance is limited to a preselected distance target value and, in a step of setting, the distance target value is set using a change of an accelerator pedal angle.

Abstract: A system for assisting in aligning a vehicle for hitching with a trailer includes a vehicle steering system, a detection system outputting a signal including object position information of an area to a rear of the vehicle, and a controller. The controller receiving the object position data and identifying a targeted trailer and at least one additional object within the area to the rear of the vehicle and controlling the vehicle steering system to maneuver the vehicle during reversing to align a hitch ball mounted on the vehicle to a coupler of the targeted trailer. Upon aligning the hitch ball with the coupler, the controller presents an indication, if it is determined that the at least one additional object is within a threshold distance of a side of the targeted trailer.

Abstract: A rear-side alarm device and a rear-side alarm method thereof. The rear-side alarm device includes a sensor and a controller. The sensor monitors a rear area or a rear-side area of a vehicle. If a trailer connected to the vehicle is detected, the controller sets a system deactivation area and a system activation area, based on status information of the trailer, and controls an alarm operation to be performed by judging whether or not another vehicle has entered into the system activation area. The rear-side alarm device can operate normally even in the case in which a trailer is connected to a vehicle, thereby providing driving safety to a driver.

Abstract: The present disclosure provides a travel control device including: determining whether a lane change request from an adjacent vehicle for a lane change is detected; determining whether a safety condition is met with regard to an expected relationship between the adjacent vehicle and a subject vehicle; determining whether an acceptance condition is met; generating a target motion for the subject vehicle to allow the lane change by the adjacent vehicle to satisfy the safety condition; and executing a reception control that is a travel control of the automated driving in accordance with the target motion.

Abstract: An information providing system includes a server and an onboard device configured to transmit and receive information to and from the server. The onboard device includes an imaging unit configured to image a nearby vehicle and a first transmission unit configured to transmit nearby vehicle information including a captured image of the nearby vehicle or dangerous vehicle information including a position of a dangerous vehicle which is detected from the nearby vehicle information and position information of the onboard device to the server. The server includes a warning unit configured to generate warning information including the position of the dangerous vehicle for a vehicle near the dangerous vehicle and a second transmission unit configured to transmit the warning information.

Abstract: The present application provides autonomous vehicle based position detection method and apparatus, a device and a medium, where the method includes: identifying an obtained first visual perception image according to an underlying neural network layer in a slender convolution kernel neural network model to determine feature information of the target linear object image, and identifying the feature information of the target linear object image by using a high-level neural network layer in the slender convolution kernel neural network model to determine size information of the target linear object image; further, matching the size information of the target linear object image with preset coordinate system map information to determine a position of the autonomous vehicle. Embodiments of the present application can accurately determine the position of the autonomous vehicle.

Abstract: An object detection system includes plural first object detection apparatuses each including a transmitting portion configured to transmit a first wave motion of a first frequency included in plural first frequencies, the transmitting portion transmitting the first wave motion of the first frequency which is different from the other first object detection apparatus concurrently with the other first object detection apparatus. The system includes a receiving portion configured to receive a reception wave, a determination portion configured to determine a correspondence relationship between one or more second frequency of one or more second wave motion included in the reception wave and the plural first frequencies, and a detection portion configured to detect information related to the object.

Abstract: A vehicle and a control method of the vehicle obtains information on a target object in a blind spot based on a result of object detection by the vehicle itself and a result of object detection received from another vehicle. In particular, to the vehicle performs a safe driving control based on the obtained information on the target object. The method of controlling the vehicle may include: setting a blind spot around the vehicle based on the detection result of the object around the vehicle, and performing the safe driving control of the vehicle based on the degree of risk of the target object in the blind spot.

Abstract: Systems and methods for connected vehicle and mobile device communications are provided herein. An example method includes determining a distracted condition for at least one of a driver or a pedestrian by evaluating actions occurring within in a vehicle of the driver or on a mobile device of the pedestrian; determining a distraction level for either the driver or the pedestrian based on the actions occurring within in the vehicle or on the mobile device; and providing an alert message to the mobile device or a human machine interface of the vehicle based on the distraction level, the alert message warning of a distracted condition of the pedestrian or the driver.

Abstract: An acquisition unit acquires laser measurement data every time a laser measurement is performed. Based on the acquired laser measurement data, a first detection unit detects an attention angle, which is an irradiation angle corresponding to a measurement distance not included in a standard distance range. A second detection unit detects a warning angle, which is an irradiation angle matching the attention angle for a time longer than an allowable time. A warning unit issues a warning when the warning angle is detected.

Abstract: A method and apparatus of predicting a crash are provided. The method includes detecting coordinates of a target obstacle based on first coordinate information of a radar sensor and second coordinate information of an ultrasonic sensor, translating the second coordinate information to the coordinate system of the first coordinate information and generating translated coordinate information, estimating a position, speed, and acceleration of a vehicle if the vehicle is a predetermined distance from the target obstacle based on the translated coordinate information, determining a potential crash type between the vehicle and the target obstacle based on the estimated position, speed and acceleration of the vehicle, determining a three-dimensional movement trajectory based on the estimated position, speed and acceleration of the vehicle and the determined crash type, and predicting a crash point of the vehicle, a crash point of the target object and a crash time based on the three-dimensional movement trajectory.

Abstract: A safe-to-proceed system (10) for operating an automated vehicle proximate to an intersection (14) includes an intersection-detector (18), a vehicle-detector (20), and a controller (24). The intersection-detector (18) is suitable for use on a host-vehicle (12). The intersection-detector (18) is used to determine when a host-vehicle (12) is proximate to an intersection (14). The vehicle-detector (20) is also suitable for use on the host-vehicle (12). The vehicle-detector (20) is used to estimate a stopping-distance (22) of an other-vehicle (16) approaching the intersection (14). The controller (24) is in communication with the intersection-detector (18) and the vehicle-detector (20). The controller (24) is configured to prevent the host-vehicle (12) from entering the intersection (14) when the stopping-distance (22) indicates that the other-vehicle (16) will enter the intersection (14) before stopping.

Abstract: A collision avoidance and/or pedestrian detection system for a large passenger vehicle such as commuter bus, which includes one or more exterior and/or interior sensing devices positioned strategically around the exterior and interior of the vehicle for recording data, method for avoiding collisions and/or detecting pedestrians, and features/articles of manufacture for improving same, is described herein in various embodiments. The sensing devices may be responsive to one or more situational sensors, and may be connected to one or more interior and/or exterior warning systems configured to alert a driver inside the vehicle and/or a pedestrian outside the vehicle that a collision may be possible and/or imminent based on a path of the vehicle and/or a position of the pedestrian as detected by one or more sensing devices and/or situational sensors.

Abstract: Systems and methods for monitoring a position of a window. A window position monitoring system comprises a display screen, a camera configured to capture an image for display on the display screen, and a controller including a processor and a memory. The memory includes instructions that, when executed by the processor, cause the processor to: determine a position of the window; determine at least one of a camera setting of a camera and a display screen setting of the display screen based on the determined position of the window; capture, using the camera, at least one frame of a video of a real-world environment; and display, on the display screen, based on the at least one of the camera setting and the display screen setting, the at least one frame of the video of the real-world environment.

Abstract: A vehicle includes a body having two pillars spaced from each other. The body includes a rocker extending from one pillar to the other pillar. An airbag is fixed to the rocker and inflatable to an inflated position. The airbag extends outboard of the rocker in the inflated position. The airbag includes a first chamber and a second chamber substantially fluidly separated from each other. The first chamber extends farther outboard of the rocker in the inflated position than the second chamber.

Abstract: A method for warning a vulnerable road user. The method includes: receiving surroundings signals, which represent a surroundings of the road user and the road user, receiving communication address signals, which represent a communication address of a mobile communication device, which the road user is carrying, ascertaining on the basis of the surroundings signals whether there is an imminent danger to the road user, in the event of an imminent danger to a road user, generating communication message signals, which represent a communication message for the mobile communication device that is to be transmitted to the communication address of the mobile communication device, the communication message comprising a warning to the road user, outputting the generated communication message signals in order to transmit the communication message to the communication address of the mobile communication device. A device, a computer program, and a machine-readable storage medium, are also described.

Abstract: The disclosure includes embodiments of a lane change timing indicator for a connected vehicle. In some embodiments, a method includes determining a time and a path for an ego vehicle to change lanes. In some embodiments, the method includes displaying, on an electronic display device of the ego vehicle, one or more graphics that depict the time and the path.

Abstract: A method and a system for vehicle-to-pedestrian collision avoidance system, the system comprising participants consisting of Long-Term Evolution (LTE)-capable user equipment (UE) terminals physically linked to at least one vehicle and at least one pedestrian; wherein a spatiotemporal positioning of the terminals is determined from Long Term Evolution (LTE) cellular radio signals mediated by Long-Term Evolution (LTE) cellular base stations (BS) and a Location Service Client (LCS) server including an embedded Artificial Intelligence algorithm comprising a Recurrent Neural Network (RNN) algorithm and analyzes the spatiotemporal positioning of the terminals and determines the likely future trajectory and communicates the likely future trajectory of the participants to the terminals physically linked to the pedestrian; the terminals physically linked to the pedestrian include an embedded Artificial Intelligence algorithm comprising a Conditional Random Fields (CRFs) algorithm to determine if the likely future traj

Abstract: An emotion estimation apparatus includes a recording section that records one or more events that cause a change in an emotion of a person and prediction information for predicting, for each event, an occurrence of the event; an event predicting section that predicts the occurrence of the event, based on detection of the prediction information; and a frequency setting section that sets a frequency with which an estimation of the emotion is performed. If the occurrence of the event is predicted by the event predicting section, the frequency setting section sets the frequency to be higher than in a case where the occurrence of the event is not predicted, and also sets the frequency based on the content of the event.

Abstract: A vehicle driving controller includes a determination device that identifies a change in location of a reference interval. The reference interval has a certain magnitude value in power spectrum data received from a radar sensor, which detects a preceding vehicle in front of a host vehicle and determines whether there occurs an altitude difference between the host vehicle and the preceding vehicle based on the change in the location of the reference interval. The driving controller includes a correction device that corrects a power spectrum using a virtual layer when it is determined that the altitude difference exists between the host vehicle and the preceding vehicle. The driving controller includes a controller that controls the driving of the host vehicle based on the corrected spectrum.

Abstract: A method and system for controlling access to restricted sectors in airspace is disclosed. The method includes creating a multi-dimensional map of airspace, overlaying a sector having boundaries onto the map, wherein the sector contains a restricted flight zone and a buffer zone monitoring the flight of an unmanned aerial vehicle (UAV), sending a command to the UAV if the UAV enters the buffer zone; and generating a response if the UAV does not leave the sector based on the command.

Abstract: The present disclosure is provided to a collision avoidance apparatus, system and method, configured to calculate a velocity vector of a following vehicle located behind the vehicle in the driving lane in which the vehicle is driving based on at least one of the image data and the sensing data, calculate a collision risk range between the vehicle with the following vehicle based on the driving data and the velocity vector, calculate a time to collision within the collision risk range, and output a warning based on the time to collision. According to the present disclosure, warning and control may be performed according to the collision risk range and the time to collision calculated using the velocity vector of the following vehicle, thereby preventing collision with an adjacent following vehicle.

Abstract: A map of an area is generated from first sensor data of a first field of view including a blocked area and second sensor data of a second field of view having an unblocked view of the blocked area.

Abstract: A method for training an auto labeling device capable of performing automatic verification by using uncertainty scores of labels is provided. The method includes steps of: a learning device (a) inputting unlabeled training images into a trained object detection network and a trained convolution network to generate bounding boxes for training and feature maps for training; and (b) (i) instructing an ROI pooling layer to generate pooled feature maps for training, (ii) at least one of (ii-1) inputting the pooled feature maps for training into a first classifier to generate first class scores for training and first box uncertainty scores for training, and (ii-2) inputting the pooled feature maps for training into a second classifier to generate second class scores for training and second box uncertainty scores for training, and (iii) training one of the first classifier using first class losses and the second classifier using second class losses.

Abstract: The purpose of the present invention is to provide an object detection device which is capable of accurately estimating the height of a road surface and is capable of reliably detecting an object present on the road surface.

Abstract: A system, computer program product, and method for displaying a warning graphic for traffic congestion is provided. The method begins with determining a mobile device approaching a zone comprising traffic congestion. Next a first warning graphic of the traffic congestion is displayed to the at least one mobile device approaching that zone, the warning graphic displayed on the one mobile device. A location of the mobile device is tracked overtime approaching that zone. The mobile device approaching is updated that zone regarding a progress of the traffic congestion indicated in the first warning graphic by displaying at least one follow-up warning graphic on the at least one mobile device, provided prior to the traffic congestion being cleared.

Abstract: An autonomous driving display system includes: an autonomous driving control device to carry out autonomous driving; an autonomous driving display device display state of which is recognizable from outside a vehicle; and an in-vehicle device to make the autonomous driving control device carry out autonomous driving of the vehicle upon receiving an autonomous driving instruction signal instructing autonomous driving, monitor start of the autonomous driving of the vehicle by the autonomous driving control device, and control a display state of the autonomous driving display device to be in a display state different from a display state before the autonomous driving is started when the autonomous driving of the vehicle is started by the autonomous driving control device.

Abstract: A warning system implemented in a vehicle includes a proximity warning module and a display module. The proximity warning module is configured to detect an object approaching the vehicle and generate a proximity warning signal. The display module is configured to display a signal in the side area when the proximity warning signal is generated, where a position of the signal is determined based on a relative direction of the object to the vehicle.

Abstract: A radar device includes: a housing that includes an aperture in a front direction as a transmitting direction of an electromagnetic wave; a circuit board disposed in the housing such that a board surface extends along the front direction; an antenna unit that includes antenna elements being arrayed along a direction intersecting the front direction in a region on a side in the front direction of the circuit board, and that transmits the electromagnetic wave to the outside of the housing through the aperture and receives a reflected wave of the electromagnetic wave; and a dielectric lens that is disposed in the aperture of the housing to extend along a direction in which the antenna elements are arrayed and that has a semi-cylindrical shape or a parabolic-cylindrical shape projecting in the front direction.

Abstract: The invention relates to an active parking assist system (4) for a motor vehicle (2), wherein the parking assist system (4) comprises a plurality of capacitive ultrasonic converters (6a, 6b, 6c, 6d, 6e, 6f, 6g, 6h, 6i, 6j, 6k, 6l, 6m, 6n, 6o) which are essentially spaced evenly in a horizontally circumferential manner outside the motor vehicle (2).

Abstract: An imaging unit for a vehicle including: an imaging device configured to image an area outside of a vehicle; and a housing attached to a side portion of a vehicle body of the vehicle and configured to accommodate the imaging device, wherein the housing comprises a convex section formed by downwardly swelling a lower wall that faces a lower side of the vehicle, and a drain hole formed in the convex section and configured to discharge water entered the housing to an outside of the housing.

Abstract: Systems and methods for predicting traffic accident location. A method includes: obtaining a plurality of accident records, each accident record is associated with a target user terminal and includes a plurality of locations; determining a plurality of refined accident locations by, for each of the plurality of accident records, operating a first clustering procedure with the corresponding plurality of locations of the target user terminal as inputs of the first clustering procedure and assigning a first result of the first clustering procedure as a refined accident location of the plurality of locations of the target user terminal; and determining at least one accident-prone road section by operating a second clustering procedure with the plurality of refined accident locations corresponding to the plurality of accident records as the inputs of the second clustering procedure.

Abstract: A travel control method estimates a position of a moving object around a subject vehicle using a detector configured to detect the moving object. This method includes detecting a direction of travel in a lane from map information. The lane is located on a road on which the moving object travels. When the subject vehicle is changing directions, the position of the moving object is estimated on the basis of the direction of travel in the lane.

Abstract: A system and method for automated lane change control for autonomous vehicles are disclosed. A particular embodiment is configured to: receive perception data associated with a host vehicle; use the perception data to determine a state of the host vehicle and a state of proximate vehicles detected near to the host vehicle; determine a first target position within a safety zone between proximate vehicles detected in a roadway lane adjacent to a lane in which the host vehicle is positioned; determine a second target position in the lane in which the host vehicle is positioned; and generate a lane change trajectory to direct the host vehicle toward the first target position in the adjacent lane after directing the host vehicle toward the second target position in the lane in which the host vehicle is positioned.

Abstract: A system and method for automated lane change control for autonomous vehicles are disclosed. A particular embodiment is configured to: receive perception data associated with a host vehicle; use the perception data to determine a state of the host vehicle and a state of proximate vehicles detected near to the host vehicle; determine a first target position within a safety zone between proximate vehicles detected in a roadway lane adjacent to a lane in which the host vehicle is positioned; determine a second target position in the lane in which the host vehicle is positioned; and generate a lane change trajectory to direct the host vehicle toward the first target position in the adjacent lane after directing the host vehicle toward the second target position in the lane in which the host vehicle is positioned.

Abstract: One embodiment of the present invention discloses a process of providing a report predicting potential risks relating to an operator driving a vehicle using information obtained from various interior and exterior sensors, vehicle onboard computer (“VOC”), and cloud network. After activating interior and exterior sensors mounted on a vehicle operated by a driver for obtaining data relating to external surroundings and internal environment, the data is forwarded to VOC for generating a current fingerprint associated with the driver. The current fingerprint represents current driving status in accordance with the collected real-time data. Upon uploading the current fingerprint to the cloud via a communications network, a historical fingerprint which represents historical driving information associated with the driver is retrieved.

Abstract: A method for operating a highly or fully automated vehicle equipped with a plurality of driving functions and at least two, at least partially redundant sensors recording measurement data. In this method, the sensors are checked for their operativeness. The measurement data of the at least two, at least partially redundant sensors are based on different measuring principles, and at least one driving function is modified and/or deactivated as a function of the operativeness of the sensors.

Abstract: A method, device and system for controlling a vehicle passing through an intersection are provided. The method includes generating a reference running track based on coordinate of first road from which the vehicle is to leave and coordinate of second road into which the vehicle is to enter. The first road and the second road intersect at the intersection through which the vehicle is to pass. The method further includes generating, in real time, a preview track of vehicle based on coordinate of centroid of the vehicle, course of the vehicle and coordinate of the second road, comparing the preview track with the reference running track and acquiring running speed limit and turning angle of the vehicle based on the comparing, and controlling the vehicle to run at the turning angle and a running speed less than or equal to the running speed limit.

Abstract: An obstacle detection apparatus includes a distance measuring sensor, an imaging section, and a control section. Based on a reception result of reception waves obtained by the distance measuring sensor, the control section acquires an estimated reflection position in an illuminated region already irradiated with search waves. Based on the estimated reflection position, the control section recognizes an outer shape of an obstacle. Based on an imaging result obtained by the imaging section, the control section performs image recognition of the obstacle. Based on a recognition result of the outer shape of the obstacle and a result of the image recognition, the control section acquires a relative position, with respect to a vehicle, of a portion of the obstacle in an image captured by the imaging section included in a non-irradiated region ahead of the illuminated region in a vehicle traveling direction.

Abstract: A vehicle headlight device includes: a first illuminating unit for reflecting, at a first mirror that rotatingly operates, laser light that has exited from a laser light source and making the laser light into scanning light, and a illuminating the scanning light toward a first illumination region that is in front of a vehicle; second illuminating unit for reflecting, at a second mirror that is an aggregate of plural micro mirrors whose angles can be changed, visible light that has exited from a visible light source and making the visible light into information displaying light, and illuminating the information displaying light toward a second illumination region that is in front of the vehicle; and controlling unit for controlling turning-off and turning-on of the laser light source, rotating operation of the first mirror, and changing of the angles of the micro mirrors at the second mirror.

Abstract: Determining yaw parameter(s) (e.g., at least one yaw rate) of an additional vehicle that is in addition to a vehicle being autonomously controlled, and adapting autonomous control of the vehicle based on the determined yaw parameter(s) of the additional vehicle. For example, autonomous steering, acceleration, and/or deceleration of the vehicle can be adapted based on a determined yaw rate of the additional vehicle. In many implementations, the yaw parameter(s) of the additional vehicle are determined based on data from a phase coherent Light Detection and Ranging (LIDAR) component of the vehicle, such as a phase coherent LIDAR monopulse component and/or a frequency-modulated continuous wave (FMCW) LIDAR component.

Abstract: The present invention provides a vehicle control apparatus capable of traveling a vehicle by automated driving, comprising a state detection unit configured to detect a state of an occupant in the vehicle, and a control unit configured to set a traveling mode of automated driving in accordance with a change in the state of the occupant detected by the state detection unit.

Abstract: A monitoring device to monitor traffic behind and on both sides of a vehicle includes a radar detecting a traveling object in a pair of first detection areas that are defined behind and on left and right sides, respectively, of the vehicle and in a pair of second detection areas that are defined behind and on the left and right sides, respectively, of the vehicle, the second detection area having a width in a vehicle width direction narrower than a width of the first detection area; a detector detecting a moving direction in which the vehicle is moving and other moving indicators to find a moving status of the vehicle; and a processor causing the radar to switch to the second detection areas when the detected moving status indicates that the moving direction of the vehicle is straight ahead and at least one other moving indicator satisfies a prescribed condition.

Abstract: A method for training an auto labeling device performing verification using uncertainty scores of auto-labeled labels is provided. The method includes steps of: a learning device (a) (i) inputting first unlabeled images into a feature pyramid network (FPN) to generate first pyramid feature maps, (ii) allowing an object detection network to generate first bounding boxes, and (iii) training the object detection network and the FPN; (b) (i) allowing the FPN to generate second pyramid feature maps and allowing the object detection network to generate second bounding boxes, (ii) instructing an ROI pooling layer to generate pooled feature maps and inputting the pooled feature maps into at least one of a first classifier to generate first class scores and first box uncertainty scores, and a second classifier to generate second class scores and second box uncertainty scores and (iii) training one of the first classifier and the second classifier.