A method of training a machine learning, ML algorithm to control a watercraft is described. The watercraft is a submarine or a submersible submerged in water. The method is implemented, at least in part, by a computer, comprising a processor and a memory, aboard the watercraft. The method comprises: obtaining training data including respective sets of sensor signals, related to respective deterrents, and corresponding actions of a set of communicatively isolated watercraft, including a first watercraft; and training the ML algorithm comprising determining relationships between the respective sets of sensor signals and the corresponding actions of the watercraft of the set thereof. A method of controlling a watercraft by a trained ML algorithm is also described.

A HAPS platform may execute a neural network (a HAPSNN) as it monitors air traffic; the neural network enables it to classify, predict, and resolve events in its airspace of coverage in real time as well as learn from new events that have never before been seen or detected. The HAPSNN-equipped HAPS platform may provide surveillance of nearly 100% of air traffic in its airspace of coverage, and the HAPSNN may process data received from a drone to facilitate safe and efficient drone operation within an airspace.

One example method includes receiving datasets based on one or more instances of sensor data that are received from one or more sensors. The sensor data is associated with an aggregate fragility level that indicates how fragile one or more packages being transported by a movable edge node in an edge environment are. Features that are based on the datasets are extracted. Based on the extracted features, events that indicate anomalous driving patterns for the movable edge node are determined. In response to determining the events, an alarm based on a predetermined threshold that is based on the aggregate fragility level is generated.

An example includes a method for training an agent to control an aircraft. The method includes: selecting, by the agent, first actions for the aircraft to perform within a first environment respectively during first time intervals based on first states of the first environment during the first time intervals, updating the agent based on first rewards that correspond respectively to the first states, selecting, by the agent, second actions for the aircraft to perform within a second environment respectively during second time intervals based on second states of the second environment during the second time intervals, and updating the agent based on second rewards that correspond respectively to the second states. At least one first rule of the first environment is different from at least one rule of the second environment.

A system and a method of controlling a drone based on pressure are provided. The method includes: installing a pressure sensor under a bearing surface of a platform body; obtaining a pressure distribution map via the pressure sensor; obtaining a centroid on the bearing surface by inputting the pressure distribution map to a machine learning model; calculating a vector from a reference centroid on the bearing surface to the centroid; and controlling a flight of the drone according to the vector.

The method can include: determining sensor information with an aircraft sensor suite: based on the sensor information, determining a flight command using a set of models: validating the flight command S130; and facilitating execution of a validated flight command. The method can optionally include generating a trained model. However, the method S100 can additionally or alternatively include any other suitable elements. The method can function to facilitate aircraft control based on autonomously generated flight commands. The method can additionally or alternatively function to achieve human-in-the-loop autonomous aircraft control, and/or can function to generate a trained neural network based on validation of autonomously generated aircraft flight commands.

A marine vessel management system, comprising: receiving input data comprising at least radar input data indicative of a first field of view and imagery input data indicative of a second field of view being at least partially overlapping with said first field of view. Processing the input data to determine data indicative of reflecting object(s) within an overlapping portion of said first field of view. Determining respective locations(s) within said second field of view, where said reflecting object(s) are identified, and obtaining radar meta-data of said reflecting object(s); processing said input imagery data said respective locations in an overlapping portion of said second field of view. Determining image data piece(s) corresponding with section(s) of said imagery data associated with said reflecting object(s). Using said radar meta-data for generating label data and generating output data comprising said image data section(s) and said label data.

A computer-implemented method of training a machine learning, ML algorithm to control spin-stabilized steerable projectiles is described. The method comprises: obtaining training data including respective policies and corresponding trajectories of a set of spin-stabilized steerable projectiles including a first projectile, wherein each policy relates to steering a projectile of the set thereof towards a target and wherein each corresponding trajectory comprises a series of states in a state space of the projectile (S2001); and training the ML algorithm comprising determining relationships between the respective policies and corresponding trajectories of the projectiles of the set thereof based on respective results of comparing the trajectories and the targets (S2002).

The present application provides an unmanned aerial vehicle (UAV)-assisted federated learning resource allocation method for an UAV-assisted federated learning wireless network scenario, which takes into account the effect of altitude of the UAV on the coverage range in order to achieve an equilibrium between the total energy consumption of the user and federated learning performance. The method simultaneously considers the total energy consumption of the user and the federated learning performance, defines the total cost function of the system. The total cost function consists of weighting of the total energy consumption of the user and the inverse of the number of users participating in federated learning, and forms the optimization problem with a minimization of the total cost function.

A morphing wing (140) of the present invention includes a link mechanism configured to be deployed in a first direction and retracted in a second direction opposite to the first direction, a plurality of front wing covers (180) mounted on a front side which is one side of the link mechanism perpendicular to the first direction, and a plurality of flight feathers (160) mounted on a rear side which is the other side of the link mechanism perpendicular to the first direction, wherein the front wing covers (180) and the flight feathers (160) are streamlined from the front side toward the rear side, and when the link mechanism is retracted, the flight feathers (160) are retracted inside the adjacent flight feathers (160).