Abstract: Systems and methods for forming radio frequency beams in communication systems are provided. Signals from one or more devices are received at a base station and are processed using a vector based deep learning (VBDL) model or network. The VBDL model can receive and process vector and/or spatial information related to or part of the received signals. An optimal beamforming vector for a received signal is determined by the VBDL model, without reference to a codebook. The VBDL model can incorporate parameters that are pruned during training to provide efficient operation of the model.

Abstract: Microwave radiometers (MRs) and methods for detecting microwave emissions using an electro-optical receiver that incorporates a photonic integrated circuit are provided. The electro-optical receiver includes an electro-optic modulator that modulates received radio frequency signals onto an optical carrier signal supplied by a pump laser. The resulting upconverted signal, containing the full spectrum of the radio frequency signals, is divided into channels by an optical filter. Each of the channels is connected to a corresponding photodetector, which produces an electrical output having an amplitude that is proportional to the amplitude of the received optical signal. The components included in the photonic integrated circuit can be formed on a single substrate. In addition, the optical filter can filter the received full spectrum optical signal into a large number of channels (e.g. greater than 50).

Abstract: Imaging systems and methods that implement a deep learning network are disclosed. The deep learning network utilizes pose information associated with at least some identified objects. The network is pruned, to reduce the amount of information processed and to optimize runtime processing when the network is deployed. In operation, the network identifies objects, and propagates pose information for at least some of the objects or components of identified objects. The network can be deployed as part of a processing system of an imaging system included as part of a remote platform.

Abstract: A multiple functional instrument is provided. The instrument includes an optical autocovariance function interferometer that can feature multiple fields of view to detect winds in the atmosphere. The instrument can include an infrared camera to detect atmospheric temperatures and the presence of clouds, and a detector assembly that detects the polarization of light returned to the interferometer. Data collected by the instrument can be provided to a deep and reinforcement learning algorithm for real-time prediction of clear air turbulence and other wind-based aviation safety phenomena. Moreover, predicted and actual conditions can be correlated and used to train a deep learning algorithm to enable more accurate predictions. The instrument can be carried by an aircraft or other platform and operated to detect clear air turbulence or other atmospheric phenomena, and to provide instructions regarding flight parameters including wind-aided navigation in order to minimize the effect of predicted turbulence.

Abstract: A multiple functional instrument is provided. The instrument includes an optical autocovariance function interferometer that can feature multiple fields of view to detect winds in the atmosphere. The instrument can include an infrared camera to detect atmospheric temperatures and the presence of clouds, and a detector assembly that detects the polarization of light returned to the interferometer. Data collected by the instrument can be provided to a deep and reinforcement learning algorithm for real-time prediction of clear air turbulence and other wind-based aviation safety phenomena. Moreover, predicted and actual conditions can be correlated and used to train a deep learning algorithm to enable more accurate predictions. The instrument can be carried by an aircraft or other platform and operated to detect clear air turbulence or other atmospheric phenomena, and to provide instructions regarding flight parameters including wind-aided navigation in order to minimize the effect of predicted turbulence.

Abstract: Methods and systems for remotely detecting gases and emissions of gases are provided. Data is collected from a scene using a sensor system. The data is initially optionally processed as 1D data to remove noise, and is then assigned a confidence value by processing the 1D data using a neural network. The confidence value is related to a likelihood that an emission has been detected at a particular location. The processed 1D data, including the confidence value, is gridded into 2D space. The 2D data is then processed using a neural network to assign a 2D confidence value. The 2D data can be fused with RGB data to produce a map of emission source locations. The data identifying emissions can also be processed using a neural network to determine and output emission rate data.

Abstract: Phase correction systems and methods capable of operating in a deployed antenna system are provided. The phase correction system includes a signal generator and a signal coupler. The signal coupler injects a signal at an end of a signal line adjacent an antenna element. Changes in an effective length of the signal line can be detected at a controller that monitors characteristics of the injected signal after it has passed through the signal line. The system can adapt to detected changes in the electrical length by controlling an adjustable phase shifter provided in line with the signal line or by applying suitable post-processing.

Abstract: Methods and systems for remotely detecting gases and emissions of gases are provided. Data is collected from a scene using a sensor system. The data is initially optionally processed as 1D data to remove noise, and is then assigned a confidence value by processing the 1D data using a neural network. The confidence value is related to a likelihood that an emission has been detected at a particular location. The processed 1D data, including the confidence value, is gridded into 2D space. The 2D data is then processed using a neural network to assign a 2D confidence value. The 2D data can be fused with RGB data to produce a map of emission source locations. The data identifying emissions can also be processed using a neural network to determine and output emission rate data.