Abstract: Systems and methods of predicting a grade of a road upon which a vehicle is traveling are disclosed. An autonomous vehicle system can receive sensor data from a sensor measuring a response from at least one mechanical component of the autonomous vehicle as the autonomous vehicle navigates a road; detect a speed of the autonomous vehicle; determine a predicted grade of the road based on the sensor data and the speed; and navigate the autonomous vehicle based on the predicted grade of the road.

Abstract: Systems and methos for learning safe driving paths in autonomous driving adversity conditions can include acquiring, by a computer system, vehicle sensor data for a plurality of driving events associated with one or more autonomous driving adversity conditions. The vehicle sensor data can include, for each driving event, corresponding image data depicting surroundings of the vehicle and corresponding inertial measurement data. The computer system can acquire, for each driving event of the plurality of driving events, corresponding vehicle positioning data indicative of a corresponding trajectory followed by the vehicle, and train, using the vehicle sensor data and the vehicle positioning data, a machine learning model to predict navigation trajectories during the autonomous driving adversity conditions.

Abstract: A method comprises monitoring, by a processor, using a sensor of a first vehicle, data associated with a retroreflective feature near a road being driven by the first vehicle; vectorizing, by the processor, the data associated with the retroreflective feature; generating, by the processor, a digital map including vectorized data associated with the retroreflective feature and a location associated with the retroreflective feature; receiving, by the processor, data associated with the retroreflective feature from a second vehicle; and executing, by the processor, a localization protocol to identify a location of the second vehicle using the digital map.

Abstract: Systems and techniques performing autonomous navigation are illustrated. One embodiment includes a method for navigation. The method inputs a set of sensor data obtained from a plurality of sensors into at least one convolutional neural network (CNN). The at least one CNN generates a plurality of key-value pairs where each the key-value pair corresponds to an individual sensor from the plurality of sensors; and a value included in the key-value pair is determined based upon a subset of sensor data obtained from the individual sensor. The method inputs at least one navigation query and the plurality of key-value pairs into a Cross-Attention Transformer (CAT). The method obtains, from the CAT, a set of weighted sums, wherein each weighted sum corresponds to: a certain key-value pair; and a certain sensor from the plurality of sensors. The method updates a model depicting a 3D environment based on the set of weighted sums.

Abstract: An autonomous vehicle can include a perception sensor and one or more processors. The processors can be configured to automatically control the autonomous vehicle to an energy supply station to resupply the autonomous vehicle; detect an arrival of the autonomous vehicle at the energy supply station; detect a calibration target; collect data from the perception sensor corresponding with the calibration target; and calibrate the perception sensor based on the collected data.

Abstract: An autonomous vehicle comprises one or more processors. The processors can be configured to receive, from a sensor of the autonomous vehicle, an image of an environment outside of the autonomous vehicle. The processors can detect potential unknown objects based on the image. The processors can compare the detection based on the image to a set of data points of a LiDAR scan to determine if there are unknown objects on a roadway.

Abstract: Disclosed herein include systems and methods for calibrating perception sensors of an automated vehicle using stationary calibration targets and preconfigured geographic information for the calibration targets. A controller of the automated vehicle references map data indicating locations of each calibration target, which is then recorded by the automated vehicle's perception sensors. The automated vehicle is configured to use a corrected geographical position system (GPS) process (e.g., RTK, PPK) to determine positions and orientations of each sensor and/or the automated vehicle. The controller uses these values to generate accurate calibrations of the automated vehicle and the sensors.

Abstract: An autonomous vehicle can include a horn; a sensor(s) configured to capture images; and one or more processors. The one or more processors can be configured to receive a sequence of images from the sensor(s), the sequence of images captured by the sensor(s) as the autonomous vehicle was moving; execute a machine learning model using the sequence of images as input to detect a human inside a second vehicle or in the surrounding environment of the autonomous vehicle depicted within the sequence of images; determine, based on the detection of the human inside the second vehicle or in the surrounding environment of the autonomous vehicle within the sequence of images, the human is depicted performing a defined arm gesture within the sequence of images; and activate the horn responsive to the determination that the human is depicted performing the defined arm gesture within the sequence of images.

Abstract: Systems and methods for generating safe driving paths in autonomous driving adversity conditions can include receiving, by a computer system including one or more processors, sensor data of a vehicle that is traveling. The sensor data can be indicative of one or more autonomous driving adversity conditions, and can include image data depicting surroundings of the vehicle. The method can include executing, by the computer system, a trained machine learning model to predict, using the sensor data, a trajectory to be followed by the vehicle through the one or more autonomous driving adversity conditions, and providing, by the computer system, an indication of the predicted trajectory to an autonomous driving system of the vehicle.

Abstract: Aspects of this technical solution can include determining, by one or more processors of a liquid collection system and based on respective amounts of liquid detected by a plurality of sensors and a velocity of a vehicle, a direction of movement of the liquid. The plurality of sensors can each be associated with respective ones of a plurality of receptacles. The plurality of sensors can be configured to detect respective amounts of liquid collected by respective openings of each of the plurality of receptacles associated with corresponding directions of movement of a liquid. The plurality of receptacles can be configured to collect at least a portion of the liquid via the respective openings.

Abstract: A system and method for performing distributed recognition divides processing steps between a device, having lower processing power, and a remotely located server, having significantly more processing power. Images captured by the device are processed at the device by applying a first set of image processing steps that includes applying a first detection. First processed images having at least one detected human is transmitted to the server, whereas a second set of image processing steps are applied to determine a stored entry matching the detected human of the first processed image.

Abstract: System is configured to receive article information corresponding to articles to be transported by computer-controlled vehicles, the articles comprising a first article and a second article, each having a maximum article dimension. System is also configured to assign travel routes about a grid comprising grid cells for the vehicles to travel thereon. The travel route of a first vehicle carrying the first article includes a turning maneuver at a first grid cell. System is configured to control the turning maneuver of the first vehicle in the first grid cell such that there is no contact between the first article carried on the first vehicle with a second article carried on a second vehicle present in a second grid cell that is adjacent to the first grid cell when the first vehicle is undertaking the turning maneuver.

Abstract: Inspection robots with a payload engagement device are described. An example inspection robot may have a housing, a drive module, having at least one wheel and a motor, where the drive module is operatively coupled to the housing. The example inspection robot may also have a payload coupled to the drive module, where the payload includes a sensor mounted to the payload, and a payload engagement device operationally coupled to the drive module and the payload, where the payload engagement device applies a selected downward force on the payload.

Abstract: An oversized representation of at least a portion of a robot is filtered (e.g., voxels are set as unoccupied for any objects that reside completely within the oversized representation) from a representation of an operational environment, which provides a digital model of the operational environment which can, for example, be used for motion planning for the robot. The oversized representation exceeds a physical dimension of at least a portion (e.g. appendage) of the robot, to advantageously account for cables and other features that are attached to, and extending beyond, the outer dimensions of the robot. The specific dimensions of the oversized representation can be based on a variety of factors, for example a geometry of the cable, orientation or position of the robot appendage, orientation or position of the cable with respect to the robot appendage, velocity of the appendage, slack in the cable, etc., which may be modeled.

Abstract: Methods and systems are described herein for determining three-dimensional locations of objects within identified portions of images. An image processing system may receive an image and an identification of location within an image. The image may be input into a machine learning model to detect one or more objects within the identified location. Multiple images may then be used to generate location estimations of those objects. Based on the location estimations, an accurate three-dimensional location may be calculated.

Abstract: A seating system includes a seat, with structures for supporting a user thereon, and actuators. Each actuator includes a deformable shell with an enclosed internal cavity containing a fluid dielectric contained, a pair of electrodes disposed on opposing sides of the deformable shell. The actuators are integrated into the structures and configured for providing at least one function, such as haptic feedback, seat adjustment, alert notification, vibratory signal, user input receiving, and massage function. A portion of the plurality of actuators may be enclosed within an encapsulating shell to form an encapsulated sheet of actuators. Each actuator may be a part of a button-on-demand system, wherein the actuator is normally in a collapsed position such that a user-facing surface of the encapsulating shell is substantially flat and, when activated by a user, the actuator is configured to expand such that the encapsulating shell is raised to form a button.

Abstract: System for use in directing a sorting operation comprises a server, a receptacle logical destination identification (ID) assignment engine, a plurality of receptacles, and an automated transport device for transporting and depositing articles into the receptacles. The system is configured to: assign a first logical destination ID to a first receptacle; and direct an automated transport device to transport and deposit a first article into the first receptacle associated with the first logical destination ID. The system is further configured to: determine that the first receptacle is full; re-assign the first logical destination ID to a second receptacle; and re-direct the automated transport device to transport and deposit the first article into the second receptacle associated with the first logical destination ID.

Abstract: System for guiding an instrument within a vascular network of a patient are disclosed. In some embodiments, the system receives a medical image from a medical imaging device and identifies a distal tip and a direction the instrument in the image. The system may then determine a waypoint for the distal tip of the instrument based at least in part on the position and direction of the distal tip of the instrument. The system may then generate a trajectory command for moving the instrument through the vascular network from the current position to the waypoint. The system may operate in a closed loop. The system may provide the trajectory command to a robotic medical system configured to move the instrument according to the command.

Abstract: A method of providing information around road based events through autonomous vehicles. The method describes how a system can maintain or upload information around a digital map indicating locations of events detected. The system can react to updated events to affect route planning, biasing and lane selection. The system can transmit updated map information to various road users including other autonomous vehicles.

Abstract: A machine vision-based method and system for locating an object within a scene are provided. The method includes uniformly illuminating a target surface of the object within the scene with light having an intensity within a relatively narrow range of wavelengths such that the light overwhelms the intensity of ambient light within the narrow range to obtain reflected, backscattered illumination. The method also includes sensing brightness of the surface due to a diffuse component of the backscattered illumination to obtain brightness information. Backscattered illumination from the target surface is inspected to obtain geometric information. Rotationally and positionally invariant surface albedo of the object is computed based on the brightness and geometric information. The surface albedo and the geometric information may then be used by a matching algorithm.