Abstract: Systems and methods are directed to automated delivery systems. In one example, a vehicle is provided including a drive system, a passenger cabin; and a delivery service pod provided relative to the passenger cabin. The delivery service pod includes an access unit configured to allow for loading and unloading of a plurality of delivery crates into the delivery service pod. The delivery service pod further includes a conveyor unit comprising multiple delivery crate holding positions, the delivery crate holding positions being defined by neighboring sidewalls spaced apart within the delivery service pod such that a respective delivery crate of the plurality of delivery crates can be positioned between neighboring sidewalls, wherein the conveyor unit is configured to be rotated to align each of the delivery crate holding positions with the access unit.

Abstract: Various examples are directed to systems and methods for navigating an autonomous vehicle. Trip plan data may describe a plurality of candidate vehicle start points, a plurality of candidate waypoints, and a plurality of candidate vehicle end points. A plurality of candidate routes may be determined between an algorithm start point and an algorithm end point. Each candidate route of the plurality of candidate routes may include at least one of the plurality of candidate waypoints and at least one of the plurality of candidate vehicle end points. A best rate may be determined using the plurality of candidate routes. The best route may include a first candidate vehicle start point, a first candidate waypoint, and a first candidate vehicle end point. The autonomous vehicle may be controlled along the best route from the first candidate vehicle start point towards the first candidate vehicle end point.

Abstract: In one aspect, a method for conducting a drop test of an article with one or more target parameters can include dropping the article and a drop carriage of a drop tower from an initial height with respect to a base of the drop tower for an initial drop test. The article can be coupled to the drop carriage. The method can include detecting accelerometer data with respect to the article for an initial impact between the drop carriage and the base of the drop tower; determining a constant energy balance curve; determining, based on the constant energy balance curve and a target pulse duration, a target complex stiffness and/or a target total weight; adjusting, based on the target complex stiffness or the target total weight, the complex stiffness and/or the total weight for a subsequent drop test; and conducting the subsequent drop test.

Abstract: Systems and methods for controlling autonomous vehicles are provided. In one example embodiment, a computing system can obtain data indicative of a service assignment associated with an autonomous vehicle. The service assignment is indicative of a destination location. The computing system can determine a checklist associated with the destination location. The computing system can provide, for display via a display device, data indicative of a user interface. The user interface can present the checklist associated with the destination location. The computing system can obtain data indicative of user input associated with the checklist. The computing system can determine that the checklist has been completed based at least in part on the user input associated with the checklist. The computing system can, in response to determining that the checklist has been completed, cause the autonomous vehicle to initiate a motion control to travel to the destination location.

Abstract: The present disclosure is directed to a modular autonomous-vehicle interior. In particular, an autonomous vehicle can comprise an interior comprising a plurality of different and distinct regions. Each of the regions can comprise one or more mechanical interfaces associated therewith. The autonomous vehicle can also include, for each of one or more of the regions, at least one modular interior unit comprising an interior element and a base mechanically interfaced with the interior via the mechanical interface(s) associated with the region.

Abstract: Systems and methods for training machine-learned models are provided. A method can include receiving a rasterized image associated with a training object and generating a predicted trajectory of the training object by inputting the rasterized image into a first machine-learned model. The method can include converting the predicted trajectory into a rasterized trajectory that spatially corresponds to the rasterized image. The method can include utilizing a second machine-learned model to determine an accuracy of the predicted trajectory based on the rasterized trajectory. The method can include determining an overall loss for the first machine-learned model based on the accuracy of the predictive trajectory as determined by the second machine-learned model. The method can include training the first machine-learned model by minimizing the overall loss for the first machine-learned model.

Abstract: Systems, methods, tangible non-transitory computer-readable media, and devices for autonomous vehicle operation are provided. For example, a computing system can receive object data that includes portions of sensor data. The computing system can determine, in a first stage of a multiple stage classification using hardware components, one or more first stage characteristics of the portions of sensor data based on a first machine-learned model. In a second stage of the multiple stage classification, the computing system can determine second stage characteristics of the portions of sensor data based on a second machine-learned model. The computing system can generate an object output based on the first stage characteristics and the second stage characteristics. The object output can include indications associated with detection of objects in the portions of sensor data.

Abstract: A georeferenced trajectory system for vehicles receives trajectory data generated by a plurality of vehicle sensors and scaffolds of previously generated maps and aligns geometry data for a geographic region and trajectory data from the received data from different map builds. A scaffold of a geographic region to be mapped during an initial map build is generated, and the trajectory data from respective map builds is aligned with the scaffold of previously generated maps to generate a map of the geographic region. The resulting map expands the coverage of the existing map such that old and new map data is in a common consistent reference frame whereby the map may be built incrementally by merging or expanding local scaffolds and filling in the merged or expanded scaffold while ensuring global consistency.

Abstract: Various examples are directed to systems and methods for controlling an autonomous vehicle. For example, a navigator system at an autonomous vehicle may generate a plurality of local routes beginning at a vehicle location and extending to a plurality of local route end points. The navigator system may access general route cost data, the general route cost data describing general route costs from the plurality of local route end points to a trip end point. The navigator system may select the first local route of the plurality of routes based at least in part on the general route cost data. A vehicle autonomy system at the autonomous vehicle may begin to control the autonomous vehicle along the first local route.

Abstract: Systems, devices, products, apparatuses, and/or methods for generating a driving path for an autonomous vehicle on a roadway by determining one or more prior probability distributions of one or more motion paths for one or more objects that have previously moved in a geographic location and/or for controlling travel of an autonomous vehicle on a roadway by predicting movement of a detected object according to one or more prior probability distributions of one or more motion paths for one or more objects that have previously moved in a geographic location.

Abstract: Systems and methods for power and thermal management of autonomous vehicles are provided. In one example embodiment, a computing system includes processor(s) and one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the processor(s) cause the computing system to perform operations. The operations include obtaining data associated with an autonomous vehicle. The operations include identifying one or more vehicle parameters associated with the autonomous vehicle based at least in part on the data associated with the autonomous vehicle. The operations include determining a modification to one or more operating characteristics of one or more systems onboard the autonomous vehicle based at least in part on the one or more vehicle parameters.

Abstract: Systems and methods are provided for generating data indicative of a friction associated with a driving surface, and for using friction data as part of controlling autonomous vehicle operations. In one example, a computing system can detect an event including at least one of an acceleration, a deceleration, or a stop associated with an autonomous vehicle and obtain, in response to detecting the event, operational data associated with the autonomous vehicle during the event. The computing system can determine, based at least in part on the operational data, data indicative of a friction associated with a surface upon which the autonomous vehicle is traveling during the event. The computing system can control the autonomous vehicle based at least in part on the data indicative of the friction associated with the surface.

Abstract: Systems and methods for dictating motion for bi-directional vehicles is provided. The method includes obtaining passenger and map data. The passenger data identifies an orientation of a passenger and the map data identifies route attributes for one or more route segments. The method includes determining one or more motion constraints for a bi-directional vehicle and map constraints for a routing the bi-directional vehicle based on the passenger data and the map data. The motion constraints can identify a vehicle orientation with which the bi-directional vehicle can travel. The map constraints can identify one or more route segments restricted from travel by the bi-directional vehicle. The method includes generating a constrained route based on the motion and map constraint(s). The constrained route can include permitted route segments and movements for the bi-directional vehicle. The method can include initiating the motion of the bi-directional vehicle based on the constrained route.

Abstract: Systems, methods, tangible non-transitory computer-readable media, and devices associated with depth estimation are provided. For example, a feature representation associated with stereo images including a first and second plurality of points can be accessed. Sparse disparity estimates associated with disparities between the first and second plurality of points can be determined. The sparse disparity estimates can be based on machine-learned models that estimate disparities based on comparisons of the first plurality of points to the second plurality of points. Confidence ranges associated with the disparities between the first and second plurality of points can be determined based on the sparse disparity estimates and the machine-learned models. A disparity map for the stereo images can be generated based on using the confidence ranges and machine-learned models to prune the disparities outside the confidence ranges.

Abstract: The present disclosure is directed to a system for calibrating cameras with a fixed focal point. In particular, a camera calibration system comprising one or more computing devices can project a plurality of fiducial markers on a target surface using the plurality of collimators. The camera calibration system can capture, using the camera, a plurality of images of the target surface with the camera, wherein the camera is rotated between each captured image in the plurality of images. The camera calibration system can compare the plurality of images with a ground truth projection. The camera calibration system can generate calibration data based on the comparison of the plurality of images with the ground truth projection. The camera calibration system can store the calibration data for use in rectifying the camera.

Abstract: Aspects of the present disclosure involve a vehicle computer system comprising a computer-readable storage medium storing a set of instructions, and a method for online light detection and ranging (Lidar) intensity normalization. Consistent with some embodiments, the method may include accumulating point data output by a channel of a Lidar unit during operation of an autonomous or semi-autonomous vehicle. The accumulated point data includes raw intensity values that correspond to a particular surface type. The method further includes calculating a median intensity value based on the raw intensity values and generating an intensity normalization multiplier for the channel based on the median intensity value. The intensity normalization multiplier, when applied to the median intensity value, results in a reflectivity value that corresponds to the particular surface type.

Abstract: Systems, methods, tangible non-transitory computer-readable media, and devices associated with trajectory prediction are provided. For example, trajectory data and goal path data can be accessed. The trajectory data can be associated with an object's predicted trajectory. The predicted trajectory can include waypoints associated with waypoint position uncertainty distributions that can be based on an expectation maximization technique. The goal path data can be associated with a goal path and include locations the object is predicted to travel. Solution waypoints for the object can be determined based on application of optimization techniques to the waypoints and waypoint position uncertainty distributions. The optimization techniques can include operations to maximize the probability of each of the solution waypoints. Stitched trajectory data can be generated based on the solution waypoints. The stitched trajectory data can be associated with portions of the solution waypoints and the goal path.

Abstract: Disclosed herein are methods and systems for performing instance segmentation that can provide improved estimation of object boundaries. Implementations can include a machine-learned segmentation model trained to estimate an initial object boundary based on a truncated signed distance function (TSDF) generated by the model. The model can also generate outputs for optimizing the TSDF over a series of iterations to produce a final TSDF that can be used to determine the segmentation mask.

Abstract: A collision-avoidance system for use with an autonomous-capable vehicle can continuously receive image frames captured of the roadway to determine drivable space in a forward direction of the vehicle. The system can determine, for each image frame, whether individual regions of the image frame depict drivable space. The system can do so using machine-learned image recognition algorithms such as convolutional neural networks generated using extensive training data. Using such techniques, the system can label regions of the image frames as corresponding to drivable space or non-drivable space. By analyzing the labeled image frames, the system can determine whether the vehicle is likely to impact a region of non-drivable space. And, in response to such a determination, the system can generate control signals that override other control systems or human operator input to control the brakes, the steering, or other sub-systems of the vehicle to avoid the collision.

Abstract: Various examples are directed to systems and methods for routing an autonomous vehicle. A vehicle autonomy system may generate first route data describing a first route for the autonomous vehicle to a first target location and control the autonomous vehicle using the first route data. The vehicle autonomy system may determine that the autonomous vehicle is within a threshold of the first target location and select a second target location associated with at least a second stopping location. The vehicle autonomy system may generate second route data describing a route extension of the first route from the first target location to the second target location and control the autonomous vehicle using the second route data.