Airbus Defence and Space

Computer implemented method for training a feed-forward deep learning model intended to the detection and instance segmentation of objects, comprising the steps of: inserting (S1) a gradient reversal layer (GRL) after a hidden layer of the deep learning model, inserting (S2) a domain classifier layer (DCL) after the gradient reversal layer (GRL); introducing (S3) a set of input values (I) of training data into an input layer of the deep learning model; obtaining (S4) a set of output values (O) in response to said input values (I); calculating (S5) a loss function of the last hidden layer of the deep learning model, calculating (S6) the gradient of the loss function; propagating backwards the loss function and repeating steps (S5) and (S6) for each hidden layers and adjusting (S7) the weight of each hidden layer, so as to minimize the gradient of the loss function.

Provided is a method (200) for supporting positioning services in a communication system, and a corresponding apparatus (400). The method comprises generating (210), by at least one transmitter (110, 120), at least one reference signal; allocating (220), by the at least one transmitter (110, 120), the at least one reference signal over a bandwidth, wherein the at least one reference signal is muted in frequency domain according to a frequency muting pattern; and transmitting (230), by the at least one transmitter (110, 120), the allocated at least one reference signal to a receiver (110, 120).

The present invention provides a method for calibrating a detector device (2) under test, comprising a) emitting a number of n first laser pulses (P2-P4) by a first set (S1) of corresponding n laser sources (3a-d) of a plurality of laser sources (3a-d) in a first time slot (t1) and a number of m second laser pulses (P1) by a second set (S2) of m laser sources (3a-d) of the plurality of laser sources (3a-d) in a second time slot (t2) different from the first time slot (t1), wherein n ≥ 1, m ≥ 1, and n + m ≥ 3, wherein a ratio of n and m is defined as a predetermined pulse ratio n/m, wherein the first laser pulses (P2-P4) and the second laser pulses (P1) have a predetermined pulse energy; b) combining the first set (S1) of n laser pulses and the second set (S2) of m laser pulses to obtain a first combined laser pulse (PC1) in the first time slot (t1) and a second combined laser pulse (PC2) in the second time slot (t2); c) directing the first and second combined laser pulses (PC1, PC2) onto a detector (21) of the detector device (2); d) detecting the first combined laser pulse (PC1) in the first time slot (t1) and the second combined laser pulse (PC2) in the second time slot (t2) by the detector device (2) to provide corresponding first and second electric signals (61, 62); e) repeating steps a) to d) by regrouping the plurality of laser sources (3a-d) into a different first set (S1) of n laser sources (3a-d) and a different second set (S2) of m laser sources (3a-d) through all permutations (A-D) of the plurality of laser sources (3a-d) by keeping the predetermined pulse ratio n/m; and f) averaging each of the detected first and second electric signals (62) of all permutations (A-D) to obtain an averaged first electric signal (61') and an averaged second electric signal (62') indicative of the response of the detector device (2) for the predetermined pulse ratio n/m. Further, the present invention provides a corresponding calibration device for calibrating a detector device (2) under test.

Systems and methods for capturing a test mass (50) within an electrode housing (12) of a spacecraft (1) with multiple actuation and sensing arrangements (22) are described. The method includes determining a position and/or attitude of the test mass (50) with respect to the actuation and sensing arrangements (22) and determining a control voltage for moving the test mass (50) into a desired position and/or attitude between the actuation and sensing arrangements (22) by electrostatic forces generated by the actuation and sensing arrangements (22), wherein electrostatic forces required for moving the test mass and the control voltage for moving the test mass (50) into the desired position are determined based on the detected position and/or attitude of the test mass (50) between the actuation and sensing arrangements (22).

Zusammenfassung wird geladen …

Computer implemented method for assessing the severity of damages caused to aircrafts, comprising the steps of: creating (S0) with an artificial intelligence model and a damage evaluation dataset, a map of maximum admissible damages, receiving (S2) an identification of a damaged aircraft component; receiving (S3) the defining parameters of the damage detected; selecting (S4) the spots of the map of maximum admissible damages which are closest to the damage detected; comparing (S5) the size of the detected damage with the maximum admissible damage sizes of the spots and providing (S6) as output an indication of the allowability or non-allowability of the damage, depending on the result of said comparison.

The present invention provides a locking apparatus for a service access in an external structure of an aircraft. In addition, the invention provides a service access having a locking apparatus of this kind, and an aircraft having a service access of this kind. The present invention furthermore provides a method for producing a locking apparatus for a service access in an external structure of an aircraft.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …