Abstract: A controller for an autonomous vehicle may include: one or more processors configured to: determine a maneuver planned for the vehicle based on a safety driving model and based on a first message from a network component external to the vehicle, the first message including a respective assessment for each proposed maneuver of at least two maneuvers proposed for the vehicle, and provide an in-vehicle instruction to perform the maneuver planned for the vehicle.

Abstract: Autonomous driving system methods and devices which trigger vehicular actions based on the monitoring of one or more occupants of a vehicle are presented. The methods, and corresponding devices, may include identifying a plurality of features in a plurality of subsets of image data detailing the one or more occupants; tracking changes over time of the plurality of features over the plurality of subsets of image data; determining a state, from a plurality of states, of the one or more occupants based on the tracked changes; and triggering the vehicular action based on the determined state.

Abstract: Deep neural networks (DNNs) with budding ensemble architectures may be trained using diversity loss. A DNN may include a backbone and a plurality of heads. The backbone includes one or more layers. A layer in the backbone may generate an intermediate tensor. The plurality of heads may include one or more pairs of heads. A pair of heads includes a first head and a second head duplicated from the first head. The second head may include the same tensor operations as the first head but different internal parameters. The intermediate tensor generated by a backbone layer may be input into both the first head and the second head. The first head may compute a first detection tensor, and the second head may compute a second detection tensor. A similarity between the first detection tensor and the second detection tensor may be used as a diversity loss for training the DNN.

Abstract: Autonomous driving system methods and devices which trigger vehicular actions based on the monitoring of one or more occupants of a vehicle are presented. The methods, and corresponding devices, may include identifying a plurality of features in a plurality of subsets of image data detailing the one or more occupants; tracking changes over time of the plurality of features over the plurality of subsets of image data; determining a state, from a plurality of states, of the one or more occupants based on the tracked changes; and triggering the vehicular action based on the determined state.

Abstract: Autonomous driving system methods and devices which trigger vehicular actions based on the monitoring of one or more occupants of a vehicle are presented. The methods, and corresponding devices, may include identifying a plurality of features in a plurality of subsets of image data detailing the one or more occupants; tracking changes over time of the plurality of features over the plurality of subsets of image data; determining a state, from a plurality of states, of the one or more occupants based on the tracked changes; and triggering the vehicular action based on the determined state.

Abstract: A controller for an autonomous vehicle may include: one or more processors configured to: determine a maneuver planned for the vehicle based on a safety driving model and based on a first message from a network component external to the vehicle, the first message including a respective assessment for each proposed maneuver of at least two maneuvers proposed for the vehicle, and provide an in-vehicle instruction to perform the maneuver planned for the vehicle.

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed that calibrate error aligned uncertainty for regression and continuous structured prediction tasks/optimizations. An example apparatus includes a prediction model, at least one memory, instructions, and processor circuitry to at least one of execute or instantiate the instructions to calculate a count of samples corresponding to an accuracy-certainty classification category, calculate a trainable uncertainty calibration loss value based on the calculated count, calculate a final differentiable loss value based on the trainable uncertainty calibration loss value, and calibrate the prediction model with the final differentiable loss value.

Abstract: A controller for an autonomous vehicle may include: one or more processors configured to: determine a maneuver planned for the vehicle based on a safety driving model and based on a first message from a network component external to the vehicle, the first message including a respective assessment for each proposed maneuver of at least two maneuvers proposed for the vehicle, and provide an in-vehicle instruction to perform the maneuver planned for the vehicle.

Abstract: For example, an apparatus may include an input to receive Machine Learning (ML) model information corresponding to an ML model to process input information; and a processor to construct a multi-model ML architecture including a plurality of ML model variants based on the ML model, wherein the processor is configured to determine the plurality of ML model variants based on an attribution-based diversity metric corresponding to a model group including a first ML model variant and a second ML model variant, wherein the attribution-based diversity metric corresponding to the model group is based on a diversity between a first attribution scheme and a second attribution scheme, the first attribution scheme representing first portions of the input information attributing to an output of the first ML model variant, the second attribution scheme representing second portions of the input information attributing to an output of the second ML model variant.

Abstract: Disclosed herein are systems, devices, and methods for an uncertainty-aware robot system that may perform a personalized risk analysis of an observed person. The uncertainty-aware robot system determines a recognized behavior for the observed person based on sensor information indicative of a gait of the observed person. The uncertainty-aware robot system determines an uncertainty score for the recognized behavior based on a comparison of the recognized behavior to potentially expected behaviors associated with the observed person and the environment. The uncertainty-aware robot system generates a remedial action instruction based on the uncertainty score.

Abstract: Disclosed herein are systems, devices, and methods for improving the safety of a robot. The safety system may determine a safety envelope of a robot based on a planned movement of the robot and based on state information about a load carried by a robot. The state information may include a dynamic status of the load. The safety system may also determine a safety risk based on a detected object with respect to the safety envelope. The safety system may also generate a mitigating action to the planned movement if the safety risk exceeds a threshold value.

Abstract: A computing device, including: a memory configured to store computer-readable instructions; and unintended human motion detection processing circuitry configured to execute the computer-readable instructions to cause the computing device to: interpret a human action; receive autonomous mobile robot (AMR) sensor data from an AMR sensor; and detect whether the human action is intended or unintended, wherein the detection is based on a predicted human action, a current human emotional or physical state, the interpreted human action, and the AMR sensor data.

Abstract: Disclosed herein are systems, devices, and methods of a robot safety system for improving the safety of human-robot collaborations. The robot safety system may include a processor that receives a collaboration policy for task collaboration between a robot and a collaborator. The task collaboration may include a movement plan of the robot to move a manipulator to a collaboration position. The robot safety system may also determine a new collaboration position based on the collaboration policy. The robot safety system may also update the collaboration position of the movement plan to the new collaboration position.

Abstract: Disclosed herein are systems, devices, and methods of a sanitation system for intelligently adapting the cleaning operations of a sanitation robot. The sanitation system detects a plurality of objects in a workspace and then, for each detected object of the plurality of objects, determines a sanitation score for the detected object based on observations of the detected object over a time period. The sanitation system generates a cleaning schedule based on the sanitation score for each detected object, wherein the cleaning schedule comprises instructions for a sanitation robot to clean the plurality of objects.

Abstract: Devices and methods for a vehicle are provided in this disclosure. A device for controlling an active seat of a vehicle may include a processor and a memory. The memory may be configured to store a transfer function. The processor may be configured to predict an acceleration of the active seat of the vehicle based on a first sensor data and the transfer function. The first sensor data may include information indicating an acceleration of a vibration source for the vehicle. The processor may be further configured to generate a control signal to control a movement of the active seat at a first instance of time based on the predicted acceleration.

Abstract: Devices and methods for an occupant of a vehicle are provided in this disclosure. A device for identifying an occupant of a vehicle may include a processor. The processor may be configured to determine an identity for the occupant based on a first sensor data including information indicating a detection of a facial attribute of the occupant. The processor may further be configured to estimate a behavior for the occupant based on a second sensor data including information indicating a detection of a body attribute of the occupant. The processor may further be configured to determine a spoofing attempt result indicating whether the occupant attempts a spoofing based on the determined identity and the estimated behavior.

Abstract: Devices and methods for a road user are provided in this disclosure. An apparatus for determining detection of a road user may include a memory configured to store a plurality of data items received from a plurality of further road users. Each of the plurality of data items may include a detection information indicating whether an object has or has not been detected by one of the plurality of further road users. Furthermore, the apparatus may include a processor that is configured to determine a detection result indicating whether the road user has been detected by the one of the plurality of further road users based on the detection information.

Abstract: Techniques are disclosed to detect, inform, and automatically correct typical awareness-related human driver mistakes. This may include those that are caused by a misunderstanding of the current situation, a lack of focus or attention, and/or overconfidence in any currently-engaged assistance features. The disclosure is directed to the prediction of vehicle maneuvers using driver and external environment modeling. The consequence of executing a predicted maneuver is categorized based upon its risk or danger posed to the driving environment, and the vehicle may execute various actions based upon the categorization of a predicted riving maneuver to mitigate or eliminate that risk.

Abstract: A vehicle data relation device includes an internal audio/image data analyzer, configured to receive first data representing at least one of audio from within the vehicle or an image from within the vehicle; identify within the first data second data representing an audio indicator or an image indicator, wherein the audio indicator is human speech associated with a significance of an object external to the vehicle, and wherein the image indicator is an action of a human within the vehicle associated with a significance of an object external to the vehicle; an external image analyzer, configured to receive third data representing an image of a vicinity external to the vehicle; identify within the third data an object corresponding to at least one of the audio indicator or the video indicator; and an object data generator, configured to generate data corresponding to the object.

Abstract: Disclosed embodiments prioritize gaps in V2X coverage and then selectively route traffic based on the prioritized gaps. Some embodiments combine historical vehicle presence along a route with predicted prospective vehicle traffic along the route to generate a map of regions that have a high confidence of a need for V2X coverage. This high confidence map is compared to a historical V2X coverage in those regions. From this comparison, a set of high priority V2X gaps is identified. Vehicles are then selectively routed either around or into the gaps.