Siemens Energy, Inc.

Siemens Aktiengesellschaft

Siemens Corporation

University System of Maryland

Siemens Audiologische Technik Gmbh

Siemens Medical Instruments Pte. Ltd.

Heiko Claussen

Upul P DeSilva

Justinian Rosca

Nancy H Ulerich

Juan L Aparicio Ojea

Eugen Solowjow

Zainul Momin

Michelle Xiaohong Yan

Chengtao Wen

Patrick Ronald Flohr

Radu Victor Balan

Joshua S McConkey

Martin Sehr

Karthik Ragunathan

Tong Fang

Henning Puder

Ti-chiun Chang

Edward David Thompson

Richard Gary McDaniel

Volkmar Sterzing

Livio Dalloro

Hans-Gerd Brummel

Ingo Thon

Michael T Loiacono

Richard Hobart Bunce

Nazif Cihan Tas

Christian P Schultz

Christine P Spiegelberg

Josep Soler Garrido

Tao Cui

Johannes Frank

Phani Ram Kumar Kuruganty

Swen Elpelt

Yash Shahapurkar

Yenming Mark Lai

Steven A Russell

Andreas Macher

Thomas Wiese

Xiaowen Yu

Carolyn Joiner

Norman Drews

Renè Fischer

Ines Ugalde Diaz

Wei Xi Xia

Günter Struck

Changtao Wen

Upul P Desilva

A computer-implemented method for building an object detection module uses mesh representations of objects belonging to specified object classes of interest to render images by a physics-based simulator. Each rendered image captures a simulated environment containing objects belonging to multiple object classes of interest placed in a bin or on a table. The rendered images are generated by randomizing a set of parameters by the simulator to render a range of simulated environments. The randomized parameters include environmental and sensor-based parameters. A label is generated for each rendered image, which includes a two-dimensional representation indicative of location and object classes of objects in that rendered image frame. Each rendered image and the respective label constitute a data sample of a synthetic training dataset. A deep learning model is trained using the synthetic training dataset to output object classes from an input image of a real-world physical environment.

Synthetic dataset creation for object detection and classification with deep learning

Autonomous operations, such as robotic grasping and manipulation, in unknown or dynamic environments present various technical challenges. For example, three-dimensional (3D) reconstruction of a given object often focuses on the geometry of the object without considering how the 3D model of the object is used in solving or performing a robot operation task. As described herein, in accordance with various embodiments, models are generated of objects and/or physical environments based on tasks that autonomous machines perform with the objects or within the physical environments. Thus, in some cases, a given object or environment may be modeled differently depending on the task that is performed using the model. Further, portions of an object or environment may be modeled with varying resolutions depending on the task associated with the model.

Task-oriented 3D reconstruction for autonomous robotic operations

An extension device for one or more automation devices in an industrial system is provided. Industrial data processing units capable of performing data processing based on one or more artificial neural networks are provided. To enable and/or accelerate one or more computations in an industrial system, thereby simplifying integration of artificial intelligence into the industrial system, and to simplify data exchange between an extension device capable of processing data using artificial intelligence and an automation device, one or more results of the one or more computations are obtained. The results indicate one or more states of the industrial system. The one or more results are provided via a process state model shared with the automation device to monitor and/or control the industrial system.

Extension device for an automation device

A controller system includes a CPU module, one or more technology modules, and a backplane bus. The CPU module comprises a processor executing a control program. The technology modules include an artificial intelligence (AI) accelerator processor configured to (a) receive input data values related to one or more machine learning models, and (b) apply the machine learning models to the input data values to generate one or more output data values. The backplane bus connects the CPU module and the technology modules. The technology modules transfer the output data values to the processor over the backplane bus and the processor uses output data values during execution of the control program.

Programmable logic controller-based modular acceleration module for artificial intelligence

A system controller for visual servoing includes a technology module with dedicated hardware acceleration for deep neural network that retrieves a desired configuration of a workpiece object being manipulated by a robotic device and receives visual feedback information from one or more sensors on or near the robotic device that includes a current configuration of the workpiece object. The hardware accelerator executes a machine learning model trained to process the visual feedback information and determine a configuration error based on a difference between the current configuration of the workpiece object and the desired configuration of the workpiece object. A servo control module adapts a servo control signal to the robotic device for manipulation of the workpiece object in response to the configuration error.

Machine learning enabled visual servoing with dedicated hardware acceleration

Computerized system and method are provided. A robotic manipulator () is arranged to grasp objects (). A gripper () is attached to robotic manipulator (), which includes an imaging sensor (). During motion of robotic manipulator (), imaging sensor () is arranged to capture images providing different views of objects in the environment of the robotic manipulator. A processor () is configured to find, based on the different views, candidate grasp locations and trajectories to perform a grasp of a respective object in the environment of the robotic manipulator. Processor () is configured to calculate respective values indicative of grasp quality for the candidate grasp locations, and, based on the calculated respective values indicative of grasp quality for the candidate grasp locations, processor () is configured to select a grasp location likely to result in a successful grasp of the respective object.

Computerized system and method using different image views to find grasp locations and trajectories for robotic pick up

Over the past several decades, rapid advances in semiconductors, automation, and control systems have resulted in the adoption of programmable logic controllers (PLCs) in an immense variety of environments. Machine learning techniques help train replacement PLCs when a legacy PLC must be replaced, e.g., due to aging or failure. The techniques facilitate the efficient adoption and correct operation of replacement PLCs in the industrial environment.

Method for replacing legacy programmable logic controllers

A method for estimating a time of flight of an acoustic signal in a hot gas flow path having background noise wherein a plurality of transceivers each generate an acoustic signal. The method includes providing an acoustic signal of interest that travels in a direction opposite a reference acoustic signal on a first acoustic path between a pair of transceivers. The method also includes identifying at least one minimum peak height in the acoustic signal of interest having a peak height that is greater than the background noise. Further, the method includes obtaining time of flight information of other acoustic paths having substantially the same path length as the first path to provide a range of time of flights wherein a time of flight that falls within the time of flight range and is associated with a minimum peak height forms the estimated time of flight.

Method for estimating time of flight for acoustic pyrometry

Systems and methods for volumetrically fabricating 3D objects. The system includes at least a controller operably connected to a build volume and one or more energy sources. The build volume defines a volumetric buildable size of the object to be fabricated and includes media for fabricating the object therein. The energy source(s) is configured to emit at least a first and second energy beam therefrom. The controller is configured to direct the first and second emitted energy beams towards an intersecting point in the build volume to begin fabricating the media therein. The energy from each of the first and second beams is not sufficient to fabricate the object out of the media. The object is fabricated once multiple beams intersect, as collectively, the energy of the intersecting beams is sufficient to fabricate the object.

Systems and methods of volumetric 3D printing

A method of calibrating transceiver positions inside an acoustic pyrometry measuring vessel that contains a plurality of transceivers, includes determining () a speed of sound in the acoustic pyrometry measuring vessel from a temperature and gas composition of a gas inside the acoustic pyrometry measuring vessel, acquiring () time-of-flight (TOFs) &#x394;timeasurements from a plurality of pairs i,j of transceivers inside the acoustic pyrometry measuring vessel, estimating () a radius of the acoustic pyrometry measuring vessel from an average of the acquired TOF measurements, and using () an estimated radius of the acoustic pyrometry measuring vessel to estimate errors &#x394;&#x3b8;j of displacement angles of the transceivers.

Fully automated calibration for acoustic pyrometry

A profile map of parameter values in a region is determined based on average values along linear paths through the region. In one example, a temperature map of a region of a gas turbine engine is created based on time-of-flight measurements from acoustic transceiver pairs arranged circumferentially around the region. A speed of sound for each transceiver pair is determined based on the time-of-flight measurements. An average temperature along each path is estimated from the time of flight, and the profile map is computed from the average temperatures. The profile map may be computed using a polynomial approximation technique, a grid optimization technique or a basis function technique.

Parameter distribution mapping in a gas turbine engine

A gas turbine engine including a hybrid waveguide for placement adjacent a flow path for hot combustion gases in the engine is presented. The waveguide includes first and second opposing ends and a first section formed of a first material extending from a transition region to the first end and a second section formed of a second material extending from the transition region to the second end. The first material is a metal or a ceramic and the second material is an electrically isolating material having a lower thermal conductivity than the material of the first section. A membrane may be used to isolate a cavity of the waveguide from combustion gas.

Acoustic measurement system incorporating a hybrid waveguide providing thermal isolation

Techniques for monitoring health of an igniter () in a gas turbine engine () include a sensor () mounted to detect signals in a combustor () and a processor () in electrical communication with the sensor. The processor is configured to receive () from the sensor first data that indicates signal changes in each of a plurality of time bins during operation of the igniter in the combustor. Information characteristics of the first data are determined (). A value is determined () of a similarity measure that indicates similarity of the information characteristics to target information characteristics. Based on the value of the similarity measure, a condition of the igniter is determined () and a device is operated () based on the conditions, for example the condition is presented on a display ().

Method and system for igniter health monitoring in a gas turbine engine

A method for determining waveguide temperature for at least one waveguide of a transceiver utilized for generating a temperature map. The transceiver generates an acoustic signal that travels through a measurement space in a hot gas flow path defined by a wall such as in a combustor. The method includes calculating a total time of flight for the acoustic signal and subtracting a waveguide travel time from the total time of flight to obtain a measurement space travel time. A temperature map is calculated based on the measurement space travel time. An estimated wall temperature is obtained from the temperature map. An estimated waveguide temperature is then calculated based on the estimated wall temperature wherein the estimated waveguide temperature is determined without the use of a temperature sensing device.

Method for determining waveguide temperature for acoustic transceiver used in a gas turbine engine

A method for measuring air mass flow into a compressor section of a gas turbine engine is provided. Air is introduced into a chamber upstream from an inlet of the compressor section. The chamber includes filter packages with at least some of the filter packages including a flow sensor and a filter structure, the filter structure filtering the air. The flow sensors measure the velocity of the air flowing through the corresponding filter package. A controller uses the air flow sensor signal data to characterize a two dimensional flow field through the chamber. Additionally, a plurality of temperature, humidity, and static pressure sensors are disposed throughout the filter packages. The controller uses the temperature, humidity, and static pressure sensor signal data to characterize a two dimensional density field through the chamber. The controller combines the flow field with the density field to calculate a two dimensional air mass flow field.

System for measuring air mass flow into a gas turbine

Growing community of inventors

North Brunswick, NJ, United States of America

Heiko Claussen