Abstract: Datasets for autonomous driving systems and multi-modal scenes may be automatically labeled using previously trained models as priors to mitigate the limitations of conventional manual data labeling. Properly versioned models, including model weights and knowledge of the dataset on which the model was previously trained, may be used to run an inference operation on unlabeled data, thus automatically labeling the dataset. The newly labeled dataset may then be used to train new models, including sparse data sets, in a semi-supervised or weakly-supervised fashion.

Abstract: A vehicular warning system control system can include a processor and a memory in communication with the processor. The memory can include a warning system control module having instructions that, when executed by the processor, cause the processor to detect, using sensor data having information about a gaze of each eye of a driver of a vehicle, an abnormality of a gaze of the driver. The instructions further cause the processor to modify, using the sensor data, a signal emitted by the vehicle when the abnormality is detected.

Abstract: Systems and methods of using a common control scheme to autonomously controlling a vehicle during semi-autonomous and fully autonomous driving modes are provided. In particular, embodiments of the presently disclosed technology incorporate reference tracking for driving input and vehicle state into this common control scheme. In some embodiments, this common control scheme may be implemented using Model Predictive Control (MPC).

Abstract: System, methods, and other embodiments described herein relate to emulating vehicle dynamics. In one embodiment, a method for emulating vehicle dynamics in a vehicle having a plurality of wheels and equipped with all-wheel steering, includes receiving emulation settings that indicate one or more environment parameters and/or vehicle parameters, detecting driver inputs including at least steering input and throttle input, executing a simulation model that receives the driver inputs and emulation settings, simulates the vehicle operating based on the driver inputs and the emulation settings, and outputs one or more simulated states of the vehicle based on the simulated operation of the vehicle, determining one or more actuation commands for each wheel of the vehicle to cause the vehicle to emulate the one or more simulated states, and executing the one or more actuation commands, wherein the actuation commands include at least wheel angle commands and torque commands.

Abstract: In accordance with one embodiment of the present disclosure, a method includes receiving an input image having an object and a background, intrinsically decomposing the object and the background into an input image data having a set of features, augmenting the input image data with a 2.5D differentiable renderer for each feature of the set of features to create a set of augmented images, and compiling the input image and the set of augmented images into a training data set for training a downstream task network.

Abstract: A method for vehicle prediction, planning, and control is described. The method includes separately encoding traffic state information at an intersection into corresponding traffic state latent spaces. The method also includes aggregating the corresponding traffic state latent spaces to form a generalized traffic geometry latent space. The method further includes interpreting the generalized traffic geometry latent space to form a traffic flow map including current and future vehicle trajectories. The method also includes decoding the generalized traffic geometry latent space to predict a vehicle behavior according to the traffic flow map based on the current and future vehicle trajectories.

Abstract: A cluster-supporting catalyst including porous carrier particles having acid sites, and catalyst metal clusters supported within the pores of the porous carrier particles. The catalyst metal clusters are obtained by supporting catalyst metal clusters having a positive charge, which is formed in a dispersion liquid containing a dispersion medium and the porous carrier particles dispersed in the dispersion medium, on the acid sites within the pores of the porous carrier particles through an electrostatic interaction.

Abstract: A method for performing a task by a robotic device includes mapping a group of task image pixel descriptors associated with a first group of pixels in a task image of a task environment to a group of teaching image pixel descriptors associated with a second group of pixels in a teaching image based on positioning the robotic device within the task environment. The method also includes determining a relative transform between the task image and the teaching image based on mapping the plurality of task image pixel descriptors. The relative transform indicates a change in one or more of points of 3D space between the task image and the teaching image. The method also includes performing the task associated with the set of parameterized behaviors based on updating one or more parameters of a set of parameterized behaviors associated with the teaching image based on determining the relative transform.

Abstract: A method for processing text prompts includes identifying a set of elements in a text prompt that form a basis for a generative output. The method also includes identifying an element of the set of elements that satisfy a refinement condition. The method further includes updating the element based on the element satisfying the refinement condition. The method still further includes generating the generative output in accordance with updating the element.

Abstract: A method for fusing neural radiance fields (NeRFs) is described. The method includes re-rendering a first NeRF and a second NeRF at different viewpoints to form synthesized images from the first NeRF and the second NeRF. The method also includes inferring a transformation between a re-rendered first NeRF and a re-rendered second NeRF based on the synthesized images from the first NeRF and the second NeRF. The method further includes blending the re-rendered first NeRF and the re-rendered second NeRF based on the inferred transformation to fuse the first NeRF and the second NeRF.

Abstract: Systems, methods, and other embodiments described herein relate to augmenting image embeddings using derived geometries for estimating scaled depth. In one embodiment, a method includes generating a geometric viewing vector using pixel coordinates and intrinsic parameters about a camera for an image captured about a scene. The method also includes deriving geometric embeddings from the geometric viewing vector associated with the image for the camera. The method also includes computing a representation by augmenting image embeddings with the geometric embeddings, the image embeddings associated with visual characteristics about the image. The method also includes estimating a scaled depth of the image from the representation.

Abstract: Systems, methods, and other embodiments described herein relate to estimating scaled depth maps by sampling variational representations of an image using a learning model. In one embodiment, a method includes encoding data embeddings by a learning model to form conditioned latent representations using attention networks, the data embeddings including features about an image from a camera and calibration information about the camera. The method also includes computing a probability distribution of the conditioned latent representations by factoring scale priors. The method also includes sampling the probability distribution to generate variations for the data embeddings. The method also includes estimating scaled depth maps of a scene from the variations at different coordinates using the attention networks.

Abstract: Systems, methods, and other embodiments described herein relate to augmenting an image frame during training that enhances scene geometries and transformation capabilities for depth prediction. In one embodiment, a method includes generating rays with camera intrinsics to form a grid for an image frame. The method also includes injecting noise, by an encoder during training of a learning model, to individually perturb pixels within pixel boundaries for the rays, the pixel boundaries defined by the grid. The method also includes removing a subset of the rays randomly by the encoder and extract features from the rays. The method also includes comparing scaled depth estimates to a ground truth for a grid resolution using the features and adjust the learning model from the comparison.

Abstract: A performance enhancement system for an electric vehicle includes a battery pack, a pump system, and a control module. The battery pack contains electrolyte fluid, and the pump system is operable to redistribute the electrolyte fluid within the battery pack. The control module may be configured to identify a vehicle usage event. The vehicle usage event may be a payload event, a pitching event, a rolling event, and/or a yawing event. In response to identifying the vehicle usage event, the control module may be configured to operate the pump system to redistribute the electrolyte fluid within the battery pack to change a static center of mass of the vehicle, a dynamic center of mass of the vehicle, and/or a moment of inertia of the vehicle.

Abstract: A method of motion planning and maneuvering of an ego vehicle is described. The method includes computing a set of potential vehicle maneuvers during a current motion planning cycle of the ego vehicle. The method also includes identifying, for the set of potential vehicle maneuvers, space corridor compatible vehicle maneuvers from a set of previously computed vehicle maneuvers during a previous motion planning cycle. The method further includes prioritizing a currently executed vehicle maneuver when the currently executed vehicle maneuver is a space corridor compatible vehicle maneuver with the set of potential vehicle maneuvers.

Abstract: A system is provided for implementing steering wheel controls, namely a center indicator on a decouplable steering wheel that is visible to the driver. The center indication controls provide a center indicator for a temporarily decouplable steering wheel, where the center indicator is visibly generated on the steering wheel. A system includes center indicator devices that are disposed along a perimeter on a steering wheel. A controller determines a position on the steering wheel that corresponds to a center position, and selects one of the plurality of center indicator devices at the determined center position on the steering wheel. The selected center indicator device produces feedback to the driver to indicate the center position on the steering wheel. Feedback from the center indicator allows drivers to be better aware of the steering wheel's position and to make safe and/or accurate maneuvers until the steering wheel is properly recoupled.

Abstract: Systems and methods are provided for improving a localization estimate of a vehicle by leveraging localization estimates of surrounding vehicles. A vehicle may estimate its own location relative to a global reference frame. The vehicle may identify nearby vehicles. The vehicle may estimate the location of the nearby vehicles. The vehicle and nearby vehicles may generate and exchange localization packets containing information about the vehicles and the location estimates. The vehicle may refine its localization estimate based on received localization packets.

Abstract: Methods and devices for picking up containers using a lifting-container device are disclosed. The container-lifting device may include a lifting fork, a friction contact surface, and a passive lifting mechanism. The passive lifting mechanism is coupled to the lifting fork and the friction contact surface on the two ends. The friction contact surface is configured to lift a target container. The lifting fork is configured to slide underneath the target container when the container is lifted. The passive lifting mechanism may transfer a fraction of a forward momentum of the container-lifting device into an upward momentum, and the friction contact surface may lift the target container by applying an upward friction force to one or more sides of the target container as a result of the upward momentum.

Abstract: In one embodiment, a signal light state detection system includes one or more processors, an a non-transitory memory module storing computer-readable instructions. The computer-readable instructions are configured to cause the one or more processors to receive a first image of a vehicle and receiving a second image of the vehicle, wherein the second image is later in time from the first image, and generate a warped image from the first image and the second image, wherein the warped image has individual pixels of one of the first image and the second image that are shifted to locations of corresponding pixels of the other of the first image and the second image. The one or more processors further generate a difference image from the warped image and one of the first image and the second image, and determine, using a classifier module, a probability of a state of vehicle signal lights.

Abstract: A method to select between tactical driving decisions of a controlled ego vehicle to reach a target destination is described. The method includes determining upcoming tactical driving decisions of the controlled ego vehicle to reach the target destination according to a mission plan. The method also includes ranking upcoming tactical driving maneuvers associated with each of the upcoming tactical driving decisions. The method further includes selecting a tactical driving maneuver prior to a critical intersection according to the ranking of the upcoming tactical driving maneuvers to reach the target destination according to the mission plan.