Abstract: Example embodiments relate to a sensor unit with rotating housing and spoiler for enhanced airflow. An example device includes one or more sensors configured to sense one or more aspects of an environment surrounding the device. The device also includes a housing that at least partially surrounds the one or more sensors. The housing and the one or more sensors are configured to rotate about a shared axis. The housing includes an inlet configured to act as an air intake for an airflow through the housing. The airflow is configured to cool the one or more sensors while the one or more sensors are operating. Further, the device includes a spoiler positioned on or near the inlet. The spoiler is configured to increase an air pressure near the inlet or promote laminar flow near the inlet in order to promote the airflow through the housing.

Abstract: Described is a system and method for robotic perception and action. The system includes a mobile autonomous platform programmed to receive a visual input related to an environment and perform a movement sequence that affects the environment based on the visual input. A neural network is in communication with the mobile autonomous platform. The neural network includes a vision autoencoder trained, based on movements of the mobile autonomous platform, to encode visual features of an object in the visual input onto a visual latent space. In addition, the neural network includes an action autoencoder trained, based on movements of the mobile autonomous platform, to encode action features of the movement sequence onto an action latent space. The neural network also includes a mapping layer between the vision autoencoder and the action autoencoder. The mapping layer decodes the visual inputs into an action output of the mobile autonomous platform.

Abstract: Embodiments relate to a method for presetting a compensation angle for steering clearance, an apparatus for performing same, and an autonomous agricultural machine in which the apparatus is installed, the method comprising: driving an agricultural machine while a steering wheel is in a reference state; supplying a control signal to a steering control motor so that the steering wheel rotates in one direction in the reference state; receiving, by a calculation unit, a measurement result of the control signal from a current sensor; and determining a compensation angle for the steering clearance on the basis of a current measurement result of the control signal.

Abstract: A lane change track planning method includes obtaining a current vehicle speed of a vehicle; determining a sampling interval that is of a lane change control parameter and that corresponds to the current vehicle speed; performing sampling in the sampling interval of the lane change control parameter to obtain a lane change control parameter set; determining a change with time of the lane change control parameter set; and planning a lane change track that is of the vehicle and that corresponds to the lane change control parameter set.

Abstract: An autonomous vehicle platform system and method configured to perform various in-season management tasks, including selectively applying fertilizer, mapping growth zones and seeding cover crop within an agricultural field, while self-navigating between rows of planted crops and beneath the canopy of the planted crops on the uneven terrain of an agricultural field, allowing for an ideal in-season application of fertilizer to occur once the planted crop is well established and growing rapidly, in an effort to limit the loss of fertilizer.

Abstract: The parking assist apparatus includes a sensor that detects an object by receiving a reflected wave reflected by the object by a transmitted wave, and a controller that executes parking assist control for parking the vehicle in a parking space set based on a detection result of the sensor. The controller provides a predetermined margin to a size of a peripheral object, which is an object detected by the sensor and exists around the parking space, in a provisional direction in which the parking space is narrowed. The controller executes parking assist control using a size of the peripheral object to which the margin is provided. When the size of the peripheral object newly detected by the sensor becomes larger than the size to which the margin is provided, the controller executes parking assist control using the size of the peripheral object newly detected by the sensor.

Abstract: An accuracy measurement method of an autonomous mobile vehicle, a calculating device, and an autonomous mobile vehicle are provided. The accuracy measurement method includes a distance calculating step, a regression center calculating step, and an average calculating step. The distance calculating step includes a controlling step, a light beam emitting step, an image capturing step, an image analyzing step, and a converting step. The regression center calculating step is performed after the distance calculating step is repeatedly performed by at least two times. The accuracy measurement method is performed to obtain an X-axis offset in an X-axis direction, a Y-axis offset in a Y-axis direction, and an angle deflection of an autonomous mobile vehicle.

Abstract: A utility vehicle includes: a travel structure including a front wheel, a rear wheel, a steering structure mounted to the front wheel, and a drive source that drives the front wheel and/or the rear wheel; circuitry that controls the travel structure to effect autonomous travel without manned operation in a given travel area; and a vehicle location detector that detects a location of the utility vehicle. During the autonomous travel, the circuitry determines, based on road condition data where the travel area is divided into regions with different predetermined road condition levels, to which of the road condition levels a road condition of a region ahead of the location of the utility vehicle in a travel direction belongs, and the circuitry controls the travel structure such that a given travel parameter is within a corresponding one of permissible ranges predetermined respectively in association with the road condition levels.

Abstract: A vehicle lane marking detection system includes a 3D sensor, a driver assist component and an electronic controller. The 3D sensor is installed to a vehicle and is configured to scan physical objects around the vehicle outputting a plurality of data points each corresponding to a surface point of a physical feature. Each data point being defined by distance, direction, intensity and vertical location relative to the vehicle. The electronic controller is connected to the 3D sensor and the driver assist component. The electronic controller evaluates a point cloud defined by the data points identifying lane markings based on the intensity of the data points. The data points having intensities greater than a predetermined level are determined to correspond to lane marking and are provided to the driver assist component with the lane markings for use thereby.

Abstract: An encoding method includes that a first device determines an I-frame position of a first image stream and an I-frame position of a second image stream based on a first resolution, a second resolution, an I-frame time interval, and a frame rate; the first device encodes the first image stream based on the I-frame position of the first image stream to obtain a first compressed stream; the first device encodes the second image stream based on the I-frame position of the second image stream to obtain a second compressed stream; and the first device sends the first compressed stream and the second compressed stream to the second device.

Abstract: Aspects of the present disclosure relate generally to systems and methods for assessing validity of a map using image data collected by a laser sensor along a vehicle path. The method may compile image data received from the laser sensor. The map subject to assessment may define an area prohibiting entry by a vehicle.

Abstract: A mobility augmentation system monitors data representative of a user's motor intent and augments the user's mobility based on the monitored motor intent data. A machine-learned model is trained to identify an intended movement based on the monitored motor intent data. The machine-learned model may be trained based on generalized or specific motor intent data (e.g., user-specific motor intent data). A machine-learned model initially trained on generalized motor intent data may be re-trained on user-specific motor intent data such that the machine-learned model is optimized to the movements of the user. The system uses the machine-learned model to identify a difference between the user's monitored movement and target movement signals. Based on the identified difference, the system determines actuation signals to augment the user's movement. The actuation signals determined can be an adjustment to a currently applied actuation such that the system optimizes the actuation strategy during application.

Abstract: The described aspects and implementations enable fast and accurate object identification in autonomous vehicle (AV) applications by combining radar data with camera images. In one implementation, disclosed is a method and a system to perform the method that includes obtaining a radar image of a first hypothetical object in an environment of the AV, obtaining a camera image of a second hypothetical object in the environment of the AV, and processing the radar image and the camera image using one or more machine-learning models MLMs to obtain a prediction measure representing a likelihood that the first hypothetical object and the second hypothetical object correspond to a same object in the environment of the AV.

Abstract: According to the present invention there is provide a method a method for matching features in an image with landmark representations in a landmark map, the method comprising: providing at least one image; extracting from the image one or more features; providing a landmark map comprising a list of landmark representations, wherein each landmark representation is a representation of a respective physical landmark; creating a plurality of policies, each policy comprising at least one matching decision of the feature and the landmark representation, each matching decision having a stage cost; assigning a policy cost to each of the policies, wherein the policy cost is a function of the stage costs of the matching decisions the policy is comprised of; selecting from the collection of the policies the policy with the desired policy cost. There is further provided an assembly having a processor which is configured to carry out said method.

Abstract: A driver assistance system include a camera mounted on a host vehicle and having an external field of view of the host vehicle and a controller configured to process image data captured by the camera. The controller may analyze the image data to determine whether a lane splitting vehicle approaching or close to the host vehicle is present and control a steering device of the host vehicle so that the host vehicle laterally moves away from a lane line close to the lane splitting vehicle when the lane splitting vehicle is present.

Abstract: In a network of autonomous or semi-autonomous vehicles, an alert may be triggered when one of the vehicles switches from autonomous to manual mode. The alert may be communicated to nearby autonomous vehicles so that drivers of those vehicles may become aware of a potentially unpredictable manual driver nearby. Drivers of autonomous vehicles who may have become disengaged (e.g., sleeping, reading, talking, etc.) during autonomous driving may become re-engaged upon noticing the alert. A re-engaged driver may choose to switch his/her own vehicle from autonomous to manual mode in order to appropriately react to an unpredictable nearby manual driver. In additional or alternative embodiments, the alert may be triggered or intensified when indications of impairment of a nearby driver or malfunction of a nearby vehicle are detected.

Abstract: An autonomous vehicle (AV) includes features that allows the AV to perform safe driving operations. An example method includes detecting that a motorcycle is operating on a roadway on which the autonomous vehicle is located. The method further includes classifying a behavior state of the motorcycle based on a location of the motorcycle relative to a split zone that extends between and into two adjacent lanes of the roadway. The behavior state indicates whether the motorcycle is lane splitting. The method further includes determining, based on the behavior state of the motorcycle, a lane permission parameter that controls whether a given trajectory for the autonomous vehicle can extend into one of the two adjacent lanes. The method further includes causing the autonomous vehicle to operate in accordance with a trajectory that satisfies the lane permission parameter based on transmitting instructions related to the trajectory to subsystems of the autonomous vehicle.

Abstract: An apparatus for steering adjustment detects a gaze direction of a driver of a vehicle from a facial image in which a face of the driver is represented. During autonomous driving in which steering of the vehicle is autonomously controlled, when an operation by the driver on a steering wheel is detected, determines whether the gaze direction of the driver corresponds to the steering direction in the operation of the driver. The apparatus adjusts a reaction force with respect to the operation of the steering wheel smaller when the gaze direction corresponds to the steering direction than when the gaze direction does not correspond to the steering direction.

Abstract: A method and apparatus for modeling the environment proximate an autonomous system. The method and apparatus accesses vision data, assigns semantic labels to points in the vision data, processes points that are identified as being a drivable surface (ground) and performs an optimization over the identified points to form a surface model. The model is subsequently used for detecting objects, planning, and mapping.

Abstract: To provide a vehicle control system that is capable of planning a trajectory that can ensure more visibility and enables safe traveling when an invisible range of a sensor exists. A vehicle control system that plans a target trajectory of a vehicle based on recognition information from an external environment sensor, the vehicle control system including a recognizing unit that recognizes an object at a periphery of the vehicle based on the recognition information; and a trajectory planning unit that plans the target trajectory such that an actual detection range of the external environment sensor becomes wide when the recognizing unit recognizes the object.

Abstract: In Multi-Policy Decision-Making (MPDM), many computationally-expensive forward simulations are performed in order to predict the performance of a set of candidate policies. In risk-aware formulations of MPDM, only the worst outcomes affect the decision making process, and efficiently finding these influential outcomes becomes the core challenge. Recently, stochastic gradient optimization algorithms, using a heuristic function, were shown to be significantly superior to random sampling. In this disclosure, it was shown that accurate gradients can be computed-even through a complex forward simulation—using approaches similar to those in dep networks. The proposed approach finds influential outcomes more reliably, and is faster than earlier methods, allowing one to evaluate more policies while simultaneously eliminating the need to design an easily-differentiable heuristic function.

Abstract: A method includes, at a mobile chassis: detecting presence of a plant in a crop row based on an image of the crop bed captured by a LIDAR sensor integrated into the mobile chassis; estimating a first distance from the plant to a first weeding tool aligned to the crop row and coupled to a front of a rail mounted to the mobile chassis; estimating a second distance from the plant to a second weeding tool aligned to the crop row and coupled to a rear of the rail; driving the first weeding tool according to a first pathway to locate the first weeding tool on a first side of the plant at a first time; and, driving the second weeding tool according to a second pathway to locate the second weeding tool on a second side of the plant at a second time succeeding the first time.

Abstract: A method and system are disclosed for estimating and using a trailer angle of a trailer relative to a vehicle connected to thereto. The method includes receiving image data from at least one first camera disposed on a vehicle and from at least one second camera disposed on a trailer coupled to the vehicle, and identifying matched point pairs by matching points in the image data from the at least one first camera with points in the image data from the at least one second camera, the points matched not being points of a representation of the vehicle or a representations of the trailer in the image data. The trailer angle of the trailer relative to the vehicle is estimated based upon the matched point pairs.

Abstract: Autonomous vehicles are capable of executing missions that abide by on-street rules or regulations, while also being able to seamlessly transition to and from “zones,” including off-street zones, with their our set(s) of rules or regulations. An on-board memory stores roadgraph information. An on-board computer is operative to execute commanded driving missions using the roadgraph information, including missions with one or more zones, each zone being defined by a sub-roadgraph with its own set of zone-specific driving rules and parameters. A mission may be coordinated with one or more payload operations, including zone with “free drive paths” as in a warehouse facility with loading and unloading zones to pick up payloads and place them down, or zone staging or entry points to one or more points of payload acquisition or placement. The vehicle may be a warehousing vehicle such as a forklift.

Abstract: A lane keeping controller, a vehicle system including the same, and a method thereof are provided. The lane keeping controller includes a processor that monitors a risk level of a vehicle in real time, upon a lane keeping control, calculates a target lateral movement distance based on a line component, integrates an offset from a predetermined offset threshold to the vehicle, when an offset between a target route and the vehicle departs from the predetermined offset threshold, and corrects the target lateral movement distance based on the integrated value to calculate a final target lateral movement distance and a storage storing data and an algorithm run by the processor.

Abstract: Techniques for performing computer vision operations for a vehicle using image data of an environment of the vehicle are described herein. In some cases, a vehicle (e.g., an autonomous vehicle) can determine a predicted position (including a predicted depth) and/or a predicted trajectory for an object in the vehicle environment based on data generated by projecting a map object described by the map data for a vehicle environment to image data of the vehicle environment.

Abstract: A method is provided automatically creating map geometry from data from various sources gathered within a geographical area using data aggregation and conflation with statistical analysis. Methods may include: receiving observation data associated with a geographic area; rasterizing objects within the observation data onto corresponding channels in a raster image corresponding to the geographic area having a given resolution; determining a distribution of locations of the objects parameterized by values in different channels in the raster image from the observation data; extracting analytic geometries from the raster image; generating map geometry based on the extracted analytic geometries; and updating a map database with the generated map geometry.

Abstract: The robot teaching method includes a pre-registration step, robot operation step, and teaching step. The pre-registration step is for specifying a relative self-position of a measuring device with respect to surrounding environment by measuring the surrounding environment using the measuring device, and registering an environment teaching point that is a teaching point of the robot specified using the relative self-position. The robot operation step for automatically operating the robot so that the relative self-position of the robot with respect to the surrounding environment become equal to the environment teaching point in a state where the measuring device is attached to the robot. The teaching step for registering a detection value of a position and a posture of the robot measured by an internal sensor as teaching information in a state where the relative self-position of the robot with respect to the surrounding environment become equal to the environment teaching point.

Abstract: Some embodiments of a method disclosed herein may include: receiving a predicted driving route, sensor ranges of sensors on an autonomous vehicle (AV), and sensor field-of-view (FOV) data; determining whether minimum sensor visibility requirements are met along the predicted driving route; predicting blind areas along the predicted driving route, wherein the predicted blind areas are determined to have potentially diminished sensor visibility; and displaying an augmented reality (AR) visualization of the blind areas using an AR display device.

Abstract: Methods, systems, and apparatuses related to autonomous vehicle object detection are described. An autonomous vehicle can capture an image corresponding to an unknown object disposed within a sight line of the autonomous vehicle. Processing resources available to a plurality of memory devices associated with the autonomous vehicle can be reallocated in response to capturing the image and an operation involving the image corresponding to the unknown object to classify the unknown object can be performed using the reallocated processing resources.

Abstract: According to the present invention, disclosed are a device and a method of generating an object image, recognizing an object, and learning an environment of a mobile robot which perform a deep learning algorithm which allows a robot to create a map and load environment information acquired during the autonomous movement while the autonomous mobile robot is being charged and may be used for an application which finds out a location by finally recognizing objects such as furniture using a method of checking a location of the recognized objects to mark the location on the map.

Abstract: Examples of the present disclosure provide a computer-implemented system, comprising instructions for performing operations including: estimating locations in a region of a map where features associated with water, snow or ice are likely to be formed under certain adverse weather conditions; simulating one of the adverse weather conditions; generating the features associated with the simulated one of the adverse weather conditions at the locations; determining a response of a perception stack of an autonomous vehicle (AV) to the adverse weather conditions observed at the locations; determining a reaction of the AV to the features generated in the region, the reaction of the AV being a function of a configuration of the AV; in response to the reaction, updating the configuration; repeating the determining the reaction and the updating the configuration until a desired reaction is obtained; and exporting a final configuration corresponding to the desired reaction to a physical AV.

Abstract: The present disclosure is directed to mobile correctional facility robots and systems and methods for coordinating mobile correctional facility robots to perform various tasks in a correctional facility. The mobile correctional facility robots can be used to perform many of the tasks traditionally assigned to correctional facility guards to help reduce the number of guards needed in any given correctional facility. When cooperation is employed among multiple mobile correctional facility robots to execute tasks, a central controller can be used to coordinate the efforts of the multiple robots to improve the performance of the overall system of robots as compared to the performance of the robots when working in uncoordinated effort to execute the tasks.

Abstract: A vehicle onboard apparatus mounted in a vehicle includes an in-vehicle processing unit that performs image recognition on an input image from a camera mounted in the vehicle. The in-vehicle processing unit performs the image recognition using an image recognition model corresponding to position information of the vehicle and additional reference information different from the position information.

Abstract: A method of localizing a robot includes receiving odometry information plotting locations of the robot and sensor data of the environment about the robot. The method also includes obtaining a series of odometry information members, each including a respective odometry measurement at a respective time. The method also includes obtaining a series of sensor data members, each including a respective sensor measurement at the respective time. The method also includes, for each sensor data member of the series of sensor data members, (i) determining a localization of the robot at the respective time based on the respective sensor data, and (ii) determining an offset of the localization relative to the odometry measurement at the respective time. The method also includes determining whether a variance of the offsets determined for the localizations exceeds a threshold variance. When the variance among the offsets exceeds the threshold variance, a signal is generated.

Abstract: A method for assisting safer riving applied in a vehicle-mounted electronic device obtains RGB images of a scene in front of a vehicle when engine is running, processes the RGB images by a trained depth estimation model, obtains depth images and converts the depth images to three-dimensional (3D) point cloud maps. A curvature of the driving path of the vehicle is calculated, and 3D regions of interest of the vehicle are extracted from the 3D point cloud maps according to a size of the vehicle and the curvature or deviation from straight ahead. The 3D regions of interest are analyzed for presence of obstacles. When no obstacles are present, the vehicle is controlled to continue driving, when presence of at least one obstacle is determined, an alarm is issued.

Abstract: Provided is a monitoring system configured such that, even when there is a deviation of a shooting region of a camera, if the deviation is within an acceptable level, an accurate vehicle detection using a recognition model can be performed without readjusting parameters or re-training the model. Based on an image of a no-entry zone and a platform zone, a processor performs a first operation for detecting a person in the no-entry zone and a second operation for detecting a vehicle in the no-entry zone; and based on detection results, determines whether an alert needs to be issued. When there is a deviation of the shooting region of a camera, the processor performs a conversion operation for converting an image of a current shooting region to an image that would be captured by the camera with an original shooting region and uses the converted image for the two detection operations.

Abstract: The present invention is a method of characterizing a route (T) travelled by a user. The method comprises a step of locating and timestamping measurement during the ride (MES), then determining a modified discrete Fréchet distance (DFM) between the measured route and previous routes. Finally, the modified discrete Fréchet distance (DFM) is used to characterize the route to be characterized (T).

Abstract: The techniques and systems herein enable track association based on azimuth extension and compactness errors. Specifically, first and second tracks comprising respective locations and footprints of respective objects are received. An azimuth distance is determined based on an azimuth extension error that corresponds to azimuth spread between the first and second tracks with respect to a host vehicle. A position distance is also determined based on a compactness error that corresponds to footprint difference between the first and second tracks. Based on the azimuth and position distances, it is established whether the first object and the second object are a common object. By doing so, the system can better determine if the tracks are of the common object when the tracks are extended (e.g., not point targets) and/or partially observed (e.g., the track is not of an entire object).

Abstract: Various embodiments for systems and methods for cooperative driving of connected autonomous vehicles using responsibility-sensitive safety (RSS) rules are disclosed herein. The CAV system integrates proposed RSS rules with CAV's motion planning algorithm to enable cooperative driving of CAVs. The CAV system further integrates a deadlock detection and resolution system for resolving traffic deadlocks between CAVs. The CAV system reduces redundant calculation of dependency graphs.

Abstract: A container transportation system using autonomous driving includes: a container to be transported; an autonomous vehicle docked with or undocked from the container and autonomously travelling to transport the container to a transport destination; and a management server controlling travelling of the autonomous vehicle, wherein the autonomous vehicle comprises a container coupling part coupled to the container, and the container comprises: a container body; a plurality of height adjustment pillars coupled to respective corners of a lower surface of the container body and adjustable in length to lift or lower the container body from or to the ground; and a vehicle coupling part formed on the lower surface of the container body and coupled to the container coupling part.

Abstract: A method for controlling an ego vehicle in an environment includes associating, by a velocity model, one or more objects within the environment with a respective velocity instance label. The method also includes selectively, by a recurrent network of the taillight recognition system, focusing on a selected region of the sequence of images according to a spatial attention model for a vehicle taillight recognition task. The method further includes concatenating the selected region with the respective velocity instance label of each object of the one or more objects within the environment to generate a concatenated region label. The method still further planning a trajectory of the ego vehicle based on inferring, at a classifier of the taillight recognition system, an intent of each object of the one or more objects according to a respective taillight state of each object, as determined based on the concatenated region label.

Abstract: A system and method for real world autonomous vehicle trajectory simulation may include: receiving training data from a data collection system; obtaining ground truth data corresponding to the training data; performing a training phase to train a plurality of trajectory prediction models; and performing a simulation or operational phase to generate a vicinal scenario for each simulated vehicle in an iteration of a simulation. Vicinal scenarios may correspond to different locations, traffic patterns, or environmental conditions being simulated. Vehicle intention data corresponding to a data representation of various types of simulated vehicle or driver intentions.

Abstract: Provided are a method, system and device for global path planning for an unmanned vehicle in an off-road environment. The method includes: obtaining satellite elevation data and a satellite remote sensing image of a current off-road environment; constructing a digital elevation model (DEM); determining slope and land surface relief of each grid in the current off-road environment; performing gray processing on the satellite remote sensing image to obtain grayscale values of the grids; determining traversal costs of the grids corresponding to different ground types; constructing a global grid map based on the slope and the land surface relief of each grid, as well as the traversal costs corresponding to the different ground types; determining a rugged terrain potential field and path costs; and searching for paths using a Bresenham's line algorithm and Theta* algorithm based on the rugged terrain potential field and the path costs, to generate a global path.

Abstract: Provided is a method for driving behavior modeling based on spatio-temporal information fusion, relating to the field of driving behavior simulations. The method includes: constructing a driving behavior model, where the driving behavior model includes a spatial information encoding network, a temporal information encoding network, a feature fusion network, and a feature decoding network, with the feature fusion network being connected to both the spatial information encoding network and the temporal information encoding network, and the feature decoding network being connected to the feature fusion network; determining a future trajectory sequence of a target main vehicle at future time points based on the trained driving behavior model according to spatial information and temporal information of the target main vehicle, where the target main vehicle is controlled to travel according to the future trajectory sequence at the future time points.

Abstract: A method for autonomously operating a following vehicle in a vehicle train together with a vehicle driving ahead of the following vehicle with respect to a direction of travel includes acquiring first route information with a front-mounted sensor. The method also includes operating the following vehicle on the basis of first route information and acquiring second route information with a rear-mounted sensor. The method further includes transmitting the second route information to the following vehicle and, when a substitute criterion is present, operating the following vehicle on the basis of the second route information.

Abstract: Disclosed is a method for designing a packaging plant, wherein a measuring vehicle is moved within an area in which the production plant is to be erected or modified, and a position of the measuring vehicle relative to the area is detected. The measuring vehicle is positioned at a plurality of positions within this area and at this position records respective images and/or data of the area with a first image capturing device, wherein at least one geometric property of the area and/or the packaging plant is detected.

Abstract: To select a lane in a multi-lane road segment for a vehicle travelling on the road segment, a system identifies, in multiple lanes and in a region ahead of the vehicle, another vehicle defining a target; the system applies an optical flow technique to track the target during a period of time, to generate an estimate of how fast traffic moves; and the system applies the estimate to machine learning (ML) model to generate a recommendation which one of the plurality of lanes the vehicle is to choose.

Abstract: An obstacle detecting device includes: an image converting portion for converting, into a circular cylindrical image, an image captured by a camera installed on a vehicle; a detection subject candidate image detecting portion for detecting a detection subject candidate image through pattern matching; an optical flow calculating portion for calculating an optical flow; an outlier removing portion for removing an optical flow that is not a detection subject; a TTC calculating portion for calculating a TTC (TTCX, TTCY); a tracking portion for generating a region of the detection subject on the circular cylindrical image by tracking the detection subject candidate; and a collision evaluating portion for evaluating whether or not there is the risk of a collision, wherein the optical flow calculating portion calculates the optical flow based on the detection subject candidate image and the region.

Abstract: Provided are methods for managing traffic light detections, which can include: deriving a first state of a traffic light at an intersection a vehicle is approaching, according to first detection data acquired by a first traffic light detection (TLD) system; deriving a second state of the traffic light at the intersection, according to second detection data acquired by a second TLD system that is independent from the first TLD system; determining traffic light information at the intersection based on at least one of (i) the first state or (ii) a result of checking whether the first state is same as the second state; and causing the vehicle to operate in accordance with the determined traffic light information at the intersection. Systems and computer program products are also provided.