Abstract: Embodiments of the present disclosure generally relate to aircraft engines and, more particularly, to detecting defects in aircraft engines using visual neuromorphic sensors. In some embodiments, an event associated with a portion of an aircraft engine may be identified based on a change on a visual data characteristic from a visual neuromorphic sensor. In response to identifying the event associated with the portion of the aircraft engine, synchronous data from a synchronous data collection sensor coupled to the aircraft engine may be retrieved for a predetermined period of time, and a defect associated with the aircraft engine detected based on the identified event and the synchronous data. Other embodiments may be disclosed or claimed.

Abstract: Systems and methods are provided for performing image processing. An exemplary method includes: receiving neuromorphic camera data from at least one neuromorphic camera; producing, using the neuromorphic camera data, a plurality of unaligned event data, where the unaligned event data are unaligned in time; aligning the unaligned event data in time producing aligned event data; generating, using a model and the aligned event data, a plurality of aligned event volumes; and aggregating the aligned event volumes to produce an image.

Abstract: Embodiments of the present disclosure generally relate to aircraft engines and, more particularly, to detecting material failures associated with aircraft engines using neuromorphic sensors. In some embodiments, an event associated with a material failure in the portion of an aircraft engine is identified based on a change in a data characteristic from a neuromorphic sensor. In response to identifying the event associated with the material failure of the aircraft engine, at least one other sensor coupled to the aircraft engine is activated, where the activated sensor(s) is configured to collect data associated with the aircraft engine synchronously. Other embodiments may be disclosed or claimed.

Abstract: Embodiments of the present disclosure generally relate to aircraft engines and, more particularly, to aircraft engine inlet monitoring using neuromorphic sensors. In some embodiments, an event associated with debris in the inlet of an aircraft engine may be identified based on a change in a visual data characteristic from a visual neuromorphic sensor. In response to identifying the event, one or more other neuromorphic sensors coupled to the aircraft may be activated. Other embodiments may be disclosed or claimed.

Abstract: Embodiments of the present disclosure generally relate to aircraft engines and, more particularly to data acquisition for aircraft engines using neuromorphic sensors. In some embodiments, an event associated with the aircraft engine may be identified based on a change in a data characteristic measured from a neuromorphic sensor and, in response to identifying the event associated with the aircraft engine, at least one other sensor coupled to the aircraft engine may be activated. Other embodiments may be disclosed or claimed.

Abstract: A laser powder bed fusion (LPBF) system includes a build plate, a build station piston configured to adjust the height of the build plate as a part is built on top of the build plate, and a powder chamber configured to contain loose build powder, wherein the powder chamber surrounds the build plate. The LPBF system also includes a laser system configured to direct a laser beam onto the loose build powder to form a melt pool, which forms a layer of the part as the melt pool solidifies. As each layer of the part is formed, the build station piston lowers the build plate and part by a predetermined distance corresponding to a desired thickness of a next layer of the part.

Abstract: A neuromorphic foreign object debris (FOD) detection system includes a FOD processing system and a neuromorphic sensor. The FOD processing system includes a FOD controller including a trained artificial intelligence machine learning (AIML) model representing an area of interest. The neuromorphic sensor has a field of view (FOV) containing the area of interest and is configured to output pixel data in response to FOD appearing in the FOV. The FOD controller detects the FOD is present in the area of interest in response to receiving the pixel data, and generates an alert signal indicating the presence of the FOD.

Abstract: Embodiments include techniques for developing a model framework for remote unit monitoring condition-based maintenance. The techniques include collecting data associated with unplanned service requests, and generating one or more models from the collected data. The techniques also include predicting unplanned service requests based at least in part on the one or more models, and transmitting an output of the prediction of the unplanned service request.

Abstract: A method for generating an application specific pose estimation responsive to a camera pose, a scene pose and an object pose includes generating an observed image (Iobs) of an object and a scene using a camera and generating a current pose estimate having an object image render and a scene image render. The method further includes generating a final rendered output (Irndr) by combining the object image render and the scene image render and processing the final rendered output (Irndr) and the observed image (Iobs) to generate a final pose estimate, wherein the final pose estimate matches the observed image (Iobs).

Abstract: Systems and methods are provided for performing image processing. An exemplary method includes: receiving neuromorphic camera data from at least one neuromorphic camera; producing, using the neuromorphic camera data, a plurality of unaligned event data, where the unaligned event data are unaligned in time; aligning the unaligned event data in time producing aligned event data; generating, using a model and the aligned event data, a plurality of aligned event volumes; and aggregating the aligned event volumes to produce an image.

Abstract: A wearable device includes neuromorphic event cameras. A processor receives data streams from the event cameras and makes application specific predictions/determinations. The event cameras may be outward facing to make determinations about the environment or a specific task, inward facing to monitor the state of the user, or both. The processor may be configured as a trained neural network to receive the data streams and produce output based on predefined sets of training data. Sensors other than event cameras may supply data to the processor and neural network, including other cameras via a feature recognition process.

Abstract: Embodiments regard implementing operations that provide a virtual sensor. A method includes receiving, from a neuromorphic sensor, a time series of delta images, receiving auxiliary data indicating (i) an orientation, location, direction, or speed of the neuromorphic sensor, or (ii) metadata of the neuromorphic sensor, the auxiliary data associated with about a same time as the time series of images, operating, based on an image of the time series of delta images and auxiliary data of the auxiliary data associated with the image as occurring at about a same time, a first machine learning (ML) model resulting in a low-resolution image, and operating, based on the low-resolution image, a second ML model resulting in a high-resolution image, the high-resolution image of a type different than that produced by the neuromorphic sensor.

Abstract: Methods of inspecting fan blades of a gas turbine engine for abnormalities are disclosed. One method utilizes a neural network to determine whether a photographic image of at least one fan blade depicts an abnormality of the at least one fan blade. Another method uses a dimensional reduction technique on one or more first photographic images that depict at least one first fan blade to obtain basis vectors, utilizes the basis vectors to obtain a reconstructed version of a second photographic image that depicts at least one second fan blade, and determines whether the at least one second fan blade includes an abnormality based on a difference between the second photographic image and the reconstructed version of the second photographic image. Another method determines whether a fan blade includes an abnormality based on contrast differences between adjacent areas of an image.

Abstract: Systems and methods for implementing neuromorphic sensor technology in aircraft cabin environments including smart devices to improve and/or expand the use and functionality of the smart devices. In some embodiments, neuromorphic sensors and smart devices are communicatively coupled to a controller operative to analyze object movements and instruct smart devices to operate according to preprogrammed instructions. Non-limiting examples include detecting passenger movements in passenger seating areas and lavatories associated with smart devices, detecting movements of smart devices configured to traverse aircraft aisles, etc. The neuromorphic sensors are incapable of capturing visually recognizable passenger identities and therefore satisfy privacy concerns.

Abstract: Systems and methods are provided for producing aligned and aggregated aligned time surfaces. An exemplary method includes: receiving camera event data from a first location; computing, using the camera event data, a plurality of unaligned time surfaces; determining a plurality of optical flows corresponding to the unaligned time surfaces; determining, using the optical flows, an accumulated optical flow for mapping displacement between any two time periods; producing, using the optical flows, for a plurality of consecutive mapped time surfaces; performing, using the accumulated optical flow, warping of the mapped time surfaces producing corresponding aligned time surfaces; and aggregating the aligned time surfaces to produce an image having lowered blurring of object boundaries.

Abstract: A method of monitoring a conveyance apparatus within a conveyance system including: detecting a vibratory signature at a primary location during a commissioning phase using a primary sensing apparatus located at the primary location; detecting a vibratory signature at a secondary location during the commissioning phase using a secondary sensing apparatus located at the secondary location; determining a transfer algorithm in response to the vibratory signature at the primary location during the commissioning phase and the vibratory signature at the secondary location during the commissioning phase; detecting a vibratory signature at the primary location during a normal operation phase using the primary sensing apparatus located at the primary location; and converting the vibratory signature at the primary location during the normal operation phase to a vibratory signature at the secondary location during the normal operation phase using the transfer algorithm.

Abstract: Systems and methods are provided for updating data in a computer network. An exemplary method includes: receiving input data from at least one device; performing an extraction operation on the input data to extract at least one feature; producing at least one feature vector based on the at least one feature; performing a similarity analysis between the at least one feature vector and a plurality of other feature vectors from a plurality of autoencoders; selecting a first autoencoder from the plurality of autoencoders demonstrating significant similarity with at least one feature vector; determining whether the input data exhibits a recurring drift or a new drift; and training a new autoencoder using at least a portion of the input data.

Abstract: According to an embodiment of the present disclosure, a method of training a machine learning model is provided. Input data is received from at least one remote device. A classifier is evaluated by determining a classification accuracy of the input data. A training data matrix of the input data is applied to a selected context autoencoder of a knowledge bank of autoencoders including at least one context autoencoder and the training data matrix is determined to be out of context for the selected autoencoder. The training data matrix is applied to each other context autoencoder of the at least one autoencoder and the training data matrix is determined to be out of context for each other context autoencoder. A new context autoencoder is constructed.

Abstract: According to an embodiment of the present disclosure, a method of training a machine learning model is provided. Input data is received from at least one remote device. A classifier is evaluated by determining a classification accuracy of the input data. A training data matrix of the input data is applied to a selected context autoencoder of a knowledge bank of autoencoders including at least one context autoencoder and the training data matrix is determined to be out of context for the selected autoencoder. The training data matrix is applied to each other context autoencoder of the at least one autoencoder and the training data matrix is determined to be out of context for each other context autoencoder. A new context autoencoder is constructed.

Abstract: A method of monitoring a conveyance apparatus within a conveyance system including: detecting a vibratory signature at a primary location during a commissioning phase using a primary sensing apparatus located at the primary location; detecting a vibratory signature at a secondary location during the commissioning phase using a secondary sensing apparatus located at the secondary location; determining a transfer algorithm in response to the vibratory signature at the primary location during the commissioning phase and the vibratory signature at the secondary location during the commissioning phase; detecting a vibratory signature at the primary location during a normal operation phase using the primary sensing apparatus located at the primary location; and converting the vibratory signature at the primary location during the normal operation phase to a vibratory signature at the secondary location during the normal operation phase using the transfer algorithm.