Abstract: Systems and methods are provided for detecting objects of interest. A computing system can input sensor data to one or more first machine-learned models associated with detecting objects external to an autonomous vehicle. The computing system can obtain as an output of the first machine-learned models, data indicative of one or more detected objects. The computing system can determine data indicative of at least one uncertainty associated with the one or more detected objects and input the data indicative of the one or more detected objects and the data indicative of the at least one uncertainty to one or more second machine-learned models. The computing system can obtain as an output of the second machine-learned models, data indicative of at least one prediction associated with the one or more detected objects. The at least one prediction can be based at least in part on the detected objects and the uncertainty.

Abstract: Systems and methods are provided for detecting objects of interest. A computing system can input sensor data to one or more first machine-learned models associated with detecting objects external to an autonomous vehicle. The computing system can obtain as an output of the first machine-learned models, data indicative of one or more detected objects. The computing system can determine data indicative of at least one uncertainty associated with the one or more detected objects and input the data indicative of the one or more detected objects and the data indicative of the at least one uncertainty to one or more second machine-learned models. The computing system can obtain as an output of the second machine-learned models, data indicative of at least one prediction associated with the one or more detected objects. The at least one prediction can be based at least in part on the detected objects and the uncertainty.

Abstract: Systems and methods are provided for detecting objects of interest. A computing system can input sensor data to one or more first machine-learned models associated with detecting objects external to an autonomous vehicle. The computing system can obtain as an output of the first machine-learned models, data indicative of one or more detected objects. The computing system can determine data indicative of at least one uncertainty associated with the one or more detected objects and input the data indicative of the one or more detected objects and the data indicative of the at least one uncertainty to one or more second machine-learned models. The computing system can obtain as an output of the second machine-learned models, data indicative of at least one prediction associated with the one or more detected objects. The at least one prediction can be based at least in part on the detected objects and the uncertainty.

Abstract: A vehicle navigation system may include a data pipeline that proposes routing graph modifications from automatically ingested data sources and that provides enough context to suggest an expiration time to remove the routing graph modifications once imposed. The system and method receives from at least one data source routing graph modification data including geographic location data identifying a location to which a routing graph modification applies. The routing graph modification data is associated with one or more roadway elements in a routing graph of a navigation constraints system, and the routing graph modification data associated with the one or more roadway elements is classified as a routing graph modification. The routing graph modification is added to or, if expired, removed from the navigation constraints system.

Abstract: Systems and methods are provided for detecting objects of interest. A computing system can input sensor data to one or more first machine-learned models associated with detecting objects external to an autonomous vehicle. The computing system can obtain as an output of the first machine-learned models, data indicative of one or more detected objects. The computing system can determine data indicative of at least one uncertainty associated with the one or more detected objects and input the data indicative of the one or more detected objects and the data indicative of the at least one uncertainty to one or more second machine-learned models. The computing system can obtain as an output of the second machine-learned models, data indicative of at least one prediction associated with the one or more detected objects. The at least one prediction can be based at least in part on the detected objects and the uncertainty.

Abstract: Systems and methods for determining object motion and controlling autonomous vehicles are provided. In one example embodiment, a computing system can be configured to perform operations. The operations can include obtaining data indicative of state(s) of a first object and a second object within a surrounding environment of an autonomous vehicle. The operations can include determining a first predicted motion trajectory of the first object based at least in part on the state data. The operations can include determining a second predicted motion trajectory of the second object based at least in part on the state data and the first predicted motion trajectory of the first object. The operations can include determining a motion plan for the autonomous vehicle based at least in part on the second predicted motion trajectory of the second object and initiating a motion control in accordance with at least a portion of the motion plan.

Abstract: In accordance with the present invention, the number of bits used to code a frame of video is controlled by controlling the frame size. In a preprocessing stage (40) of a video coding circuit (10), each frame of video is processed by a low pass anti-aliasing filter (43,44) in the horizontal and vertical directions. The cutoff frequencies in the horizontal and vertical directions are determined by a desired frame reduction factor in the corresponding direction. Then to reduce the sampling rate in each dimension, each frame of video is processed horizontally and vertically by a shift varying filter (45,46). When the sampling rate in both the horizontal and vertical directions is reduced by M/N, the number of code bits decreases linearly with M/N.

Abstract: A technique for the sub-band decomposition and reconstruction of video signals is disclosed. In an illustrative embodiment of the present invention, each field of a video signal is decomposed into sub-bands utilizing a polyphase filter bank unit including an infinite impulse response allpass filter with coefficients that are powers of two. To reconstruct the original image, an FIR is utilized which approximates an IIR allpass filter This eliminates the need for a full field memory in the reconstruction process.