Abstract: A computing system including a processor configured to train a synthetic aperture radar (SAR) classifier neural network. The SAR classifier neural network is trained at least in part by, at a SAR encoder, receiving training SAR range profiles that are tagged with respective first training labels, and, at an image encoder, receiving training two-dimensional images that are tagged with respective second training labels. Training the SAR classifier neural network further includes, at a shared encoder, computing shared latent representations based on the SAR encoder outputs and the image encoder outputs, and, at a classifier, computing respective classification labels based on the shared latent representations. Training the SAR classifier neural network further includes computing a value of a loss function based on the plurality of first training labels, the plurality of second training labels, and the plurality of classification labels and performing backpropagation based on the value of the loss function.

Abstract: A computing system including a processor configured to train a synthetic aperture radar (SAR) classifier neural network. The SAR classifier neural network is trained at least in part by, at a SAR encoder, receiving training SAR range profiles that are tagged with respective first training labels, and, at an image encoder, receiving training two-dimensional images that are tagged with respective second training labels. Training the SAR classifier neural network further includes, at a shared encoder, computing shared latent representations based on the SAR encoder outputs and the image encoder outputs, and, at a classifier, computing respective classification labels based on the shared latent representations. Training the SAR classifier neural network further includes computing a value of a loss function based on the plurality of first training labels, the plurality of second training labels, and the plurality of classification labels and performing backpropagation based on the value of the loss function.

Abstract: A computing system including a processor configured to train a synthetic aperture radar (SAR) classifier neural network. The SAR classifier neural network is trained at least in part by, at a SAR encoder, receiving training SAR range profiles that are tagged with respective first training labels, and, at an image encoder, receiving training two-dimensional images that are tagged with respective second training labels. Training the SAR classifier neural network further includes, at a shared encoder, computing shared latent representations based on the SAR encoder outputs and the image encoder outputs, and, at a classifier, computing respective classification labels based on the shared latent representations. Training the SAR classifier neural network further includes computing a value of a loss function based on the plurality of first training labels, the plurality of second training labels, and the plurality of classification labels and performing backpropagation based on the value of the loss function.

Abstract: A method of identifying a target from synthetic aperture radar (SAR) data without incurring the computational load associated with generating an SAR image. The method includes receiving SAR data collected by a radar system including RF phase history data associated with reflected RF pulses from a target in a scene, but excluding an SAR image. Range profile data is determined from the SAR data by converting the RF phase history data into a structured temporal array that can be applied as input to a classifier incorporating a recurrent neural network, such as a recurrent neural network made up of long short-term memory (LSTM) cells that are configured to recognize temporal or spatial characteristics associated with a target, and provide an identification of a target based on the recognized temporal or spatial characteristic.

Abstract: A method of identifying a target from synthetic aperture radar (SAR) data without incurring the computational load associated with generating an SAR image. The method includes receiving SAR data collected by a radar system including RF phase history data associated with reflected RF pulses from a target in a scene, but excluding an SAR image. Range profile data is determined from the SAR data by converting the RF phase history data into a structured temporal array that can be applied as input to a classifier incorporating a recurrent neural network, such as a recurrent neural network made up of long short-term memory (LSTM) cells that are configured to recognize temporal or spatial characteristics associated with a target, and provide an identification of a target based on the recognized temporal or spatial characteristic.

Abstract: The disclosed system for cooperative engagement generally includes a control system in communication with an actuation system, an effector system, a sensor system and a communications interface. Disclosed features and specifications may be variously controlled, adapted or otherwise optionally modified to realize improved distributed designation and/or engagement function. Exemplary embodiments of the present invention generally provide cooperative designation and engagement of targets for air-, land-, sea- or space-based weapon systems.

Abstract: A system and method for decentralized cooperative control of a team of agents for geographic and other search tasks. The approach is behavior-based and uses probability particle approach to the search problem. Agents are attracted to probability distributions in the form of virtual particles of probability that represent hypotheses about the existence of objects of interest in a geographic area or a data-space. Reliance on dependable, high-bandwidth communication is reduced by modeling the movements of other team members and the objects of interest between periodic update messages.

Abstract: An apparatus configured to obtain, process, and relay data to a user in a coherent and useful manner. An active fiducial is equipped with an interface for receiving and transmitting data. The fiducial may transmit its position using a satellite-based position sensing device such as a GPS. Active fiducials may also be equipped with battery power pack regenerated with solar cells. Similarly, the fiducials can be equipped with at least one video camera or other device having a focal plane array and a computer software system, configured to recognize shapes. The fiducials may also be equipped with inductive coils or other means for sensing metal containing compounds. The active fiducials may be equipped with a gas chromatograph. The active fiducials may use a variety of propulsion means including motor driven tracks, motor driven wheels, propellers, or other device or a combination of devices.

Abstract: The present invention relates to computing reachable areas given a first point. More specifically, the present invention relates to the computation of an intersection between a first surface and a second surface for determining a set of points that are reachable from a first point. Using the present invention, a user can determine either (1) a locus of target sites that can be struck by a ballistic projectile from a given launch site, or (2) a locus of launch sites that can be used to hit a given target site. The disclosed system and method employs graphics hardware to determine an intersection between a first surface defined by trajectory paths, and a second surface defined by terrain.

Abstract: Described is a method for event localization within a distributed sensor array using a plurality of sensor nodes. The method includes an act of receiving a signal in at least one detecting node. The signal originates from an external disturbance and has a local signal such that the local signal reflects a relative proximity (i.e., between the detecting nodes) to the external disturbance. The method also includes acts of exchanging information regarding the signal between the detecting node and nearby nodes; and localizing the external disturbance based on its relative proximity. Through receiving a signal that reflects a relative proximity to the external disturbance and exchanging that information between nearby nodes, the sensor array localizes the external disturbance.

Abstract: The disclosed system for cooperative engagement generally includes a control system in communication with an actuation system, an effector system, a sensor system and a comunications interface. Disclosed features and specifications may be variously controlled, adapted or otherwise optionally modified to realize improved distributed designation and/or engagement function. Exemplary embodiments of the present invention generally provide cooperative designation and engagement of targets for air-, land-, sea- or space-based weapon systems.

Abstract: Moving a target mobile node to arrange a number of mobile nodes includes determining a first direction. A first relative position for each of the neighboring mobile nodes of the target mobile node is established. The target mobile node moves according to a first alignment procedure to align the mobile nodes in the first direction. A second direction substantially orthogonal to the first direction is determined. A second relative position for each of the neighboring mobile nodes is established. The target mobile node moves according to a second alignment procedure to align the mobile nodes in the second direction.

Abstract: A system and method for decentralized cooperative control of a team of agents for geographic and other search tasks. The approach is behavior-based and uses probability particle approach to the search problem. Agents are attracted to probability distributions in the form of virtual particles of probability that represent hypotheses about the existence of objects of interest in a geographic area or a data-space. Reliance on dependable, high-bandwidth communication is reduced by modeling the movements of other team members and the objects of interest between periodic update messages.

Abstract: A filter driver (usable with a system having a bus, a host connected to the bus and one or more devices connected to the bus) that supplants first device objects (DOs) with second DOs. Such a filter driver includes: an intercept code portion to intercept a set of data identifying one or more first DOs, respectively; a determination code portion to determine addresses of second DOs corresponding to the first DOs identified by the data set, respectively; and a change code portion to change the data set such that members thereof identify the second DOs rather than the first DOs.

Abstract: A method and apparatus for computing properties of a physical environment is provided, using a plurality of agents forming a distributed network embedded within the environment. The method comprises determining an initiating agent 200, transmitting a signal including a cumulative cost value to neighboring agents 202, and processing the signal at each neighboring agent to augment the cumulative cost value with local information 204. If multiple signals are received, determining which has the best cumulative cost value for generating a new signal 206, then treating the neighboring agent as an initiating agent 208 and transmitting the new signal to neighboring agents 208 and retaining the best augmented cost value in memory 210. Methods further include determining paths using shortest path computations, using dual gradients for aligning agents on a path between two reference agents, and discovering and converging agents on choke points.