According to one aspect, a method is provided to determine whether an autonomous system is ready to be deployed or is otherwise ready for use, scene-based metrics, or metrics based on instances of scenarios. Scene-based metrics are mapped, or otherwise translated, to distance-based metrics such that substantially standard distance-based metrics may be used to gate the readiness of an autonomy system for deployment.

Included is a surface cleaning service system including: one or more robotic surface cleaning devices, each including: a chassis; a set of wheels; one or more motors to drive the wheels; one or more processors; one or more sensors; and a network interface card, wherein the one or more processors of each of the one or more robotic surface cleaning devices determine respective usage data. A control system or the one or more processors of each of the one or more robotic surface cleaning devices is configured to associate each usage data with a particular corresponding robotic surface cleaning device of the one or more robotic surface cleaning devices.

In accordance with one embodiment of the present disclosure, method includes obtaining multi-level environment data corresponding to a plurality of driving environment levels, encoding the multi-level environment data at each level, extracting features from the multi-level environment data at each encoded level, fusing the extracted features from each encoded level with a spatial-temporal attention framework to generate a fused information embedding, and decoding the fused information embedding to predict driving environment information at one or more driving environment levels.

A method includes generating a parameter of a trajectory associated with a scenario using a path planner. The parameter is generated based on a training dataset. The method includes comparing the parameter of the trajectory against a validation parameter associated with a validation dataset. The validation parameter is based on human-based vehicle driving trajectory data associated with scenarios that satisfy a level of similarity with the scenario. The method further includes determining a level of similarity between the parameter associated with the scenario and the validation parameter associated with the scenarios, and, subsequent to determining that the level of similarity fails to satisfy a similarity threshold, the method concludes with providing training data associated with the scenario to the training dataset so that a subsequent parameter of a subsequent trajectory generated by the path planner and associated with the scenario satisfies the level of similarity against the validation parameter.

The technology relates to controlling a vehicle in an autonomous driving mode, the method. For instance, a vehicle may be maneuvered in the autonomous driving mode using pre-stored map information identifying traffic flow directions. Data may be received from a perception system of the vehicle identifying objects in an external environment of the vehicle related to a traffic redirection not identified the map information. The received data may be used to identify one or more corridors of a traffic redirection. One of the one or more corridors may be selected based on a direction of traffic flow through the selected corridor. The vehicle may then be controlled in the autonomous driving mode to enter and follow the selected one of the one or more corridors based on the determined direction of flow of traffic through each of the one or more corridors.

A method for recognizing at least one object in at least one input image. In the method, a template image of the object is processed by a first convolutional neural network (CNN) to form at least one template feature map; the input image is processed by a second CNN to form at least one input feature map; the at least one template feature map is compared to the at least one input feature map; it is evaluated from the result of the comparison whether and possibly at which position the object is contained in the input image, the convolutional neural networks each containing multiple convolutional layers, and at least one of the convolutional layers being at least partially formed from at least two filters, which are convertible into one another by a scaling operation.

A controller of a machine determines jointly a sequence of control inputs defining a state trajectory of the machine and a desired knowledge of the environment by solving a multivariable constrained optimization of a model of dynamics of the machine relating the state trajectory with the sequence of control inputs subject to a constraint on admissible values of the states and the control inputs defined based on the desired knowledge of the surrounding environment represented by the state of the environment and the uncertainty of the state of the environment determined from the measurements of the environment. In such a manner, the controller performs joint but imbalance optimization of the control inputs and the sensing instructions to the sensor for learning the environment.

A controller of a machine determines jointly a sequence of control inputs defining a state trajectory of the machine and a desired knowledge of the environment by solving a multivariable constrained optimization of a model of dynamics of the machine relating the state trajectory with the sequence of control inputs subject to a constraint on admissible values of the states and the control inputs defined based on the desired knowledge of the surrounding environment represented by the state of the environment and the uncertainty of the state of the environment determined from the measurements of the environment. In such a manner, the controller performs joint but imbalance optimization of the control inputs and the sensing instructions to the sensor for learning the environment.

Techniques for navigating semi-autonomous mobile robots are described. A semi-autonomous mobile robot moves within an environment to complete a task. A navigation server communicates with the robot and provides the robot information. The robot includes a navigation map of the environment, interaction information, and a security level. To complete the task, the robot transmits a route reservation request to the navigation server, the route reservation request including a priority for the task, a timeslot, and a route. The navigation server grants the route reservation if the task priority is higher than the task priorities of conflicting route reservation requests from other robots. As the robot moves within the environment, the robot detects an object and attempts to classify the detected object as belonging to an object category. The robot retrieves an interaction profile for the object, and interacts with the object according to the retrieved interaction profile.

Techniques for navigating semi-autonomous mobile robots are described. A semi-autonomous mobile robot moves within an environment to complete a task. A navigation server communicates with the robot and provides the robot information. The robot includes a navigation map of the environment, interaction information, and a security level. To complete the task, the robot transmits a route reservation request to the navigation server, the route reservation request including a priority for the task, a timeslot, and a route. The navigation server grants the route reservation if the task priority is higher than the task priorities of conflicting route reservation requests from other robots. As the robot moves within the environment, the robot detects an object and attempts to classify the detected object as belonging to an object category. The robot retrieves an interaction profile for the object, and interacts with the object according to the retrieved interaction profile.