Abstract: An autonomous vehicle is configured to detect an active turn signal indicator on another vehicle. An image-capture device of the autonomous vehicle captures an image of a field of view of the autonomous vehicle. The autonomous vehicle captures the image with a short exposure to emphasize objects having brightness above a threshold. Additionally, a bounding area for a second vehicle located within the image is determined. The autonomous vehicle identifies a group of pixels within the bounding area based on a first color of the group of pixels. The autonomous vehicle also calculates an oscillation of an intensity of the group of pixels. Based on the oscillation of the intensity, the autonomous vehicle determines a likelihood that the second vehicle has a first active turn signal. Additionally, the autonomous vehicle is controlled based at least on the likelihood that the second vehicle has a first active turn signal.

Abstract: In some implementations, an image classification system of an autonomous or semi-autonomous vehicle is capable of improving multi-object classification by reducing repeated incorrect classification of objects that are considered rarely occurring objects. The system can include a common instance classifier that is trained to identify and recognize general objects (e.g., commonly occurring objects and rarely occurring objects) as belonging to specified object categories, and a rare instance classifier that is trained to compute one or more rarity scores representing likelihoods that an input image is correctly classified by the common instance classifier. The output of the rare instance classifier can be used to adjust the classification output of the common instance classifier such that the likelihood of input images being incorrectly classified is reduced.

Abstract: An autonomous vehicle is configured to detect an active turn signal indicator on another vehicle. An image-capture device of the autonomous vehicle captures an image of a field of view of the autonomous vehicle. The autonomous vehicle captures the image with a short exposure to emphasize objects having brightness above a threshold. Additionally, a bounding area for a second vehicle located within the image is determined. The autonomous vehicle identifies a group of pixels within the bounding area based on a first color of the group of pixels. The autonomous vehicle also calculates an oscillation of an intensity of the group of pixels. Based on the oscillation of the intensity, the autonomous vehicle determines a likelihood that the second vehicle has a first active turn signal. Additionally, the autonomous vehicle is controlled based at least on the likelihood that the second vehicle has a first active turn signal.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating a future occupancy prediction for a region of an environment. In one aspect, a method comprises: receiving sensor data generated by a sensor system of a vehicle that characterizes an environment in a vicinity of the vehicle as of a current time point, wherein the sensor data comprises a plurality of sensor samples characterizing the environment that were each captured at different time points; processing a network input comprising the sensor data using a neural network to generate an occupancy prediction output for a region of the environment, wherein: the occupancy prediction output characterizes, for one or more future intervals of time after the current time point, a respective likelihood that the region of the environment will be occupied by an agent in the environment during the future interval of time.

Abstract: Systems, methods, devices, and other techniques for planning a trajectory of a vehicle. A computing system can implement a trajectory planning neural network configured to, at each time step of multiple time steps: obtain a first neural network input and a second neural network input. The first neural network input can characterize a set of waypoints indicated by the waypoint data, and the second neural network input can characterize (a) environmental data that represents a current state of an environment of the vehicle and (b) navigation data that represents a planned navigation route for the vehicle. The trajectory planning neural network may process the first neural network input and the second neural network input to generate a set of output scores, where each output score in the set of output scores corresponds to a different location of a set of possible locations in a vicinity of the vehicle.

Abstract: A neural network system for identifying positions of objects in an input image can include an object detector neural network, a memory interface subsystem, and an external memory. The object detector neural network is configured to, at each time step of multiple successive time steps, (i) receive a first neural network input that represents the input image and a second neural network input that identifies a first set of positions of the input image that have each been classified as showing a respective object of the set of objects, and (ii) process the first and second inputs to generate a set of output scores that each represents a respective likelihood that an object that is not one of the objects shown at any of the positions in the first set of positions is shown at a respective position of the input image that corresponds to the output score.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating object predictions using a neural network. One of the methods includes receiving respective projections of a plurality of channels of input sensor data, wherein each channel of input sensor data represents different respective characteristics of electromagnetic radiation reflected off of one or more objects. Each of the projections of the plurality of channels of input sensor data are provided to a neural network subsystem trained to receive projections of input sensor data as input and to provide an object prediction as an output. At the output of the neural network subsystem, an object prediction that predicts a region of space that is likely to be occupied by an object is received.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for selecting locations in an environment of a vehicle where objects are likely centered and determining properties of those objects. One of the methods includes receiving an input characterizing an environment external to a vehicle. For each of a plurality of locations in the environment, a respective first object score that represents a likelihood that a center of an object is located at the location is determined. Based on the first object scores, one or more locations from the plurality of locations are selected as locations in the environment at which respective objects are likely centered. Object properties of the objects that are likely centered at the selected locations are also determined.

Abstract: A computing device of a first vehicle may receive a first image and a second image of a second vehicle having flashing light signals. The computing device may determine, in the first image and the second image, an image region that bounds the second vehicle such that the image region substantially encompasses the second vehicle. The computing device may determine a polar grid that partitions the image region in the first image and the second image into polar bins, and identify portions of image data exhibiting a change in color and a change in brightness between the first image and the second image. The computing device may determine a type of the flashing light signals and a type of the second vehicle; and accordingly provide instructions to control the first vehicle.

Abstract: In some implementations, an autonomous or semi-autonomous vehicle is capable of using a collision prediction system to determine a confidence that any objects detected within a vicinity of the vehicle are on a trajectory that will collide with the vehicle. Laser obstacle points derived from recent sensor readings of one or more sensors of a vehicle are initially obtained. The laser obstacle points are projected into a pose coordinate system to generate an occupancy grid of a vicinity of the vehicle. A confidence that any objects represented by the laser obstacle points are on a trajectory that will collide with the vehicle is determined by applying a particle filter to the occupancy grid.

Abstract: A computing device of a first vehicle may receive a first image and a second image of a second vehicle having flashing light signals. The computing device may determine, in the first image and the second image, an image region that bounds the second vehicle such that the image region substantially encompasses the second vehicle. The computing device may determine a polar grid that partitions the image region in the first image and the second image into polar bins, and identify portions of image data exhibiting a change in color and a change in brightness between the first image and the second image. The computing device may determine a type of the flashing light signals and a type of the second vehicle; and accordingly provide instructions to control the first vehicle.

Abstract: Systems, methods, devices, and other techniques for planning a trajectory of a vehicle. A computing system can implement a trajectory planning neural network configured to, at each time step of multiple time steps: obtain a first neural network input and a second neural network input. The first neural network input can characterize a set of waypoints indicated by the waypoint data, and the second neural network input can characterize (a) environmental data that represents a current state of an environment of the vehicle and (b) navigation data that represents a planned navigation route for the vehicle. The trajectory planning neural network may process the first neural network input and the second neural network input to generate a set of output scores, where each output score in the set of output scores corresponds to a different location of a set of possible locations in a vicinity of the vehicle.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for selecting locations in an environment of a vehicle where objects are likely centered and determining properties of those objects. One of the methods includes receiving an input characterizing an environment external to a vehicle. For each of a plurality of locations in the environment, a respective first object score that represents a likelihood that a center of an object is located at the location is determined. Based on the first object scores, one or more locations from the plurality of locations are selected as locations in the environment at which respective objects are likely centered. Object properties of the objects that are likely centered at the selected locations are also determined.

Abstract: An autonomous vehicle is configured to detect an active turn signal indicator on another vehicle. An image-capture device of the autonomous vehicle captures an image of a field of view of the autonomous vehicle. The autonomous vehicle captures the image with a short exposure to emphasize objects having brightness above a threshold. Additionally, a bounding area for a second vehicle located within the image is determined. The autonomous vehicle identifies a group of pixels within the bounding area based on a first color of the group of pixels. The autonomous vehicle also calculates an oscillation of an intensity of the group of pixels. Based on the oscillation of the intensity, the autonomous vehicle determines a likelihood that the second vehicle has a first active turn signal. Additionally, the autonomous vehicle is controlled based at least on the likelihood that the second vehicle has a first active turn signal.

Abstract: Methods and systems for the use of detected objects for image processing are described. A computing device autonomously controlling a vehicle may receive images of the environment surrounding the vehicle from an image-capture device coupled to the vehicle. In order to process the images, the computing device may receive information indicating characteristics of objects in the images from one or more sources coupled to the vehicle. Examples of sources may include RADAR, LIDAR, a map, sensors, a global positioning system (GPS), or other cameras. The computing device may use the information indicating characteristics of the objects to process received images, including determining the approximate locations of objects within the images. Further, while processing the image, the computing device may use information from sources to determine portions of the image to focus upon that may allow the computing device to determine a control strategy based on portions of the image.

Abstract: A vehicle may receive one or more images of an environment of the vehicle. The vehicle may also receive a map of the environment. The vehicle may also match at least one feature in the one or more images with corresponding one or more features in the map. The vehicle may also identify a given area in the one or more images that corresponds to a a portion of the map that is within a threshold distance to the one or more features. The vehicle may also compress the one or more images to include a lower amount of details in areas of the one or more images other than the given area. The vehicle may also provide the compressed images to a remote system, and responsively receive operation instructions from the remote system.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining a location of a particular object relative to a vehicle. In one aspect, a method includes obtaining sensor data captured by one or more sensors of a vehicle. The sensor data is processed by a convolutional neural network to generate a sensor feature representation of the sensor data. Data is obtained which defines a particular spatial region in the sensor data that has been classified as including sensor data that characterizes the particular object. An object feature representation of the particular object is generated from a portion of the sensor feature representation corresponding to the particular spatial region. The object feature representation of the particular object is processed using a localization neural network to generate the location of the particular object relative to the vehicle.

Abstract: An autonomous vehicle is configured to detect an active turn signal indicator on another vehicle. An image-capture device of the autonomous vehicle captures an image of a field of view of the autonomous vehicle. The autonomous vehicle captures the image with a short exposure to emphasize objects having brightness above a threshold. Additionally, a bounding area for a second vehicle located within the image is determined. The autonomous vehicle identifies a group of pixels within the bounding area based on a first color of the group of pixels. The autonomous vehicle also calculates an oscillation of an intensity of the group of pixels. Based on the oscillation of the intensity, the autonomous vehicle determines a likelihood that the second vehicle has a first active turn signal. Additionally, the autonomous vehicle is controlled based at least on the likelihood that the second vehicle has a first active turn signal.

Abstract: Methods and systems for the use of detected objects for image processing are described. A computing device autonomously controlling a vehicle may receive images of the environment surrounding the vehicle from an image-capture device coupled to the vehicle. In order to process the images, the computing device may receive information indicating characteristics of objects in the images from one or more sources coupled to the vehicle. Examples of sources may include RADAR, LIDAR, a map, sensors, a global positioning system (GPS), or other cameras. The computing device may use the information indicating characteristics of the objects to process received images, including determining the approximate locations of objects within the images. Further, while processing the image, the computing device may use information from sources to determine portions of the image to focus upon that may allow the computing device to determine a control strategy based on portions of the image.

Abstract: In some implementations, an image classification system of an autonomous or semi-autonomous vehicle is capable of improving multi-object classification by reducing repeated incorrect classification of objects that are considered rarely occurring objects. The system can include a common instance classifier that is trained to identify and recognize general objects (e.g., commonly occurring objects and rarely occurring objects) as belonging to specified object categories, and a rare instance classifier that is trained to compute one or more rarity scores representing likelihoods that an input image is correctly classified by the common instance classifier. The output of the rare instance classifier can be used to adjust the classification output of the common instance classifier such that the likelihood of input images being incorrectly classified is reduced.