US 11,947,863 B2
Intelligent audio analytic apparatus (IAAA) and method for space system
Samarjit Das, Sewickley, PA (US); and Joseph Szurley, Upper Saint Clair, PA (US)
Assigned to Robert Bosch GmbH, Stuttgart (DE)
Appl. No. 16/976,540
Filed by Robert Bosch GmbH, Stuttgart (DE)
PCT Filed Feb. 26, 2019, PCT No. PCT/EP2019/054660
§ 371(c)(1), (2) Date Aug. 28, 2020,
PCT Pub. No. WO2019/166397, PCT Pub. Date Sep. 6, 2019.
Claims priority of provisional application 62/636,498, filed on Feb. 28, 2018.
Prior Publication US 2020/0409653 A1, Dec. 31, 2020
Int. Cl. G06F 17/00 (2019.01); G05B 23/02 (2006.01); G06F 3/16 (2006.01); G06N 3/08 (2023.01); G06Q 10/20 (2023.01)
CPC G06F 3/165 (2013.01) [G05B 23/0229 (2013.01); G06N 3/08 (2013.01); G06Q 10/20 (2013.01)] 12 Claims
OG exemplary drawing
 
1. A space system equipment monitoring system comprising:
a plurality of sensing assemblies, each sensing assembly being configured for incorporation in, onto or proximate a respective space system equipment of a plurality of space system equipment arranged at various different locations, each of the plurality of sensing assemblies being configured to output sensor signals indicative of sounds and/or vibrations from the respective space system equipment; and
an intelligent audio analytic (IAA) device configured to receive the sensor signals from the plurality of sensing assemblies, to process the sensor signals to detect anomalies from operation of the plurality of space system equipment, and to output detection information to an output device to alert an operator of detection of any anomalies, the IAA device including programmed instructions stored on a non-transitory computer readable storage medium and a processor configured to execute the programmed instructions, the programmed instructions including an analytics algorithm for execution by the processor to continuously process the sensor signals to detect the anomalies,
wherein the analytics algorithm includes an autoencoder neural network, the autoencoder network being configured to receive the respective sensor signals as input and to output respective reconstructed sensor signals based on the input,
wherein the processor is configured to (i) generate respective reconstruction error signals that are indicative of a difference between the input and the output of the autoencoder neural network for the respective sensor signals, (ii) detect anomalies based on a magnitude of each respective reconstruction error signal, and (iii) perform spatio-temporal analysis of load distribution impacts on the plurality of space system equipment by aggregating results from the plurality of space system equipment arranged at the various different locations.