WO2020167291A1 - Système et procédé informatisés configurés avec des contraintes hiérarchiques pour assurer le fonctionnement sûr d'une machine autonome - Google Patents

Système et procédé informatisés configurés avec des contraintes hiérarchiques pour assurer le fonctionnement sûr d'une machine autonome Download PDF

Info

Publication number
WO2020167291A1
WO2020167291A1 PCT/US2019/017616 US2019017616W WO2020167291A1 WO 2020167291 A1 WO2020167291 A1 WO 2020167291A1 US 2019017616 W US2019017616 W US 2019017616W WO 2020167291 A1 WO2020167291 A1 WO 2020167291A1
Authority
WO
WIPO (PCT)
Prior art keywords
controller
control
learning
safety
hierarchical
Prior art date
Application number
PCT/US2019/017616
Other languages
English (en)
Inventor
Martin SEHR
Eugen SOLOWJOW
Chengtao Wen
Juan L. Aparicio Ojea
Heiko Claussen
Original Assignee
Siemens Aktiengesellschaft
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Siemens Aktiengesellschaft filed Critical Siemens Aktiengesellschaft
Priority to PCT/US2019/017616 priority Critical patent/WO2020167291A1/fr
Publication of WO2020167291A1 publication Critical patent/WO2020167291A1/fr

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1674—Programme controls characterised by safety, monitoring, diagnostic
    • B25J9/1676—Avoiding collision or forbidden zones
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1628—Programme controls characterised by the control loop
    • B25J9/163—Programme controls characterised by the control loop learning, adaptive, model based, rule based expert control
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0055—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots with safety arrangements
    • G05D1/0077—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots with safety arrangements using redundant signals or controls
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0088—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
    • G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/30—Nc systems
    • G05B2219/40—Robotics, robotics mapping to robotics vision
    • G05B2219/40496—Hierarchical, learning, recognition level controls adaptation, servo level

Definitions

  • Disclosed embodiments relate generally to the field of autonomous machines, and, more particularly, to system and method as may involve an
  • autonomous machine for carrying out an autonomous industrial process, and, even more particularly, to computerized system and method involving a control framework including a safe set control designed to ensure a safe operation of the autonomous machine in the environment of the autonomous industrial process.
  • autonomous machines such as industrial robots, etc.
  • AI artificial intelligence
  • robots have proven effective at performing repetitive tasks with little or no human intervention.
  • One challenge that must be solved to cost-effectively and reliably realize such tasks involves avoiding harmful situations for humans and/or for a given robot and, at the same time, preserving productivity.
  • learning control algorithms that may utilize AI may not be commonly deployed in industrial settings involving a substantial level of safety and/or reliability restrictions. This is predominately due to the uncertainty generally associated with running such algorithms on new data, e.g., data not available during training, etc.
  • a disclosed embodiment is directed to a computerized method.
  • the method includes: operatively coupling a learning-based controller to receive one or more signals in connection with an autonomous industrial process; operatively coupling a safety-preserving controller to receive one or more signals indicative of a potentially dangerous condition in an environment of the autonomous industrial process; arranging in a hierarchical-based processor a hierarchical control framework including a safe set control for the environment of the autonomous industrial process; jointly processing in the hierarchical-based processor a respective learning control output signal from the learning-based controller and a safety-preserving control output signal from the safety-preserving controller, where the jointly processing of the respective control output signals from the learning-based controller and from the safety-preserving controller is subject to the safe set control of the hierarchical control framework; generating in the hierarchical- based processor a control signal based on the safe set control of the hierarchical control framework; and transmitting the control signal to an autonomous machine operating in the environment of the autonomous industrial process to perform the autonomous industrial process, wherein in response to the transmitted control signal, the autonomous machine executes one or more actions configured to ensure a safe
  • a further disclosed embodiment is directed to a computerized system.
  • a learning-based controller may be coupled to receive one or more signals in connection with an autonomous industrial process.
  • a safety-preserving controller may be coupled to receive one or more signals indicative of a potentially dangerous condition in an environment of the autonomous industrial process.
  • a hierarchical-based processor may be arranged with a hierarchical control framework including a safe set control for the environment of the autonomous industrial process.
  • the hierarchical-based processor may be arranged to receive a respective learning control output signal from the learning-based controller and a safety-preserving control output signal from the safety-preserving controller.
  • the hierarchical-based processor may be arranged to perform a joint processing of the received control output signals from the learning-based controller and from the safety preserving controller subject to the safe set control of the hierarchical control framework.
  • the hierarchical-based processor may be configured to generate a control signal based on the safe set control of the hierarchical control framework.
  • the autonomous machine executes one or more actions configured to ensure a safe operation in the environment in connection with the autonomous industrial process.
  • FIG. 1 illustrates a block diagram of one non-limiting embodiment of a
  • FIG. 2 illustrates a block diagram of another non-limiting embodiment of the disclosed computerized system.
  • the present inventors have recognized that lack of systematically verifiably safe and robust operations may be one basic inhibitor in extending learning- based methodology, such as may involve machine learning and AI algorithms, to industrial automation applications involving a substantial level of safety and/or reliability restrictions.
  • Manufacturing processes for instance, may typically demand substantially high precision, safety, robustness, and reproducibility from control algorithms deployed in any given control solution in an industrial setting involving a substantial level of safety and/or reliability restrictions.
  • non- learning-based control methodologies such as feedback-based control methodologies
  • this may be relatively more challenging to achieve when machine learning and AI algorithms are part of such control solutions. As noted above, this may be due to the uncertainty generally associated when running such algorithms on new data, e.g., data not available during training, etc.
  • Disclosed embodiments formulate an innovative approach for cost-effectively and reliably realizing learning-based control algorithms in industrial automation settings involving a substantial level of safety and/or reliability restrictions.
  • disclosed embodiment may conceptually use methodologies gracefully blending or fusing the deterministic certainty obtained from a feedback-based control with the powerful versatility of learning algorithms to synergistically obtain reliable closed-loop properties and the versatility of modern learning algorithms for processing vast quantities of data while appropriately ensuring a safe operation in an industrial automation settings involving a substantial level of safety and/or reliability restrictions.
  • FIG. 1 illustrates a block diagram of one non-limiting embodiment of a
  • disclosed computerized system 10 as may be used in connection with an autonomous machine 30 for carrying out an autonomous process 40, such as without limitation, an autonomous industrial process.
  • a learning-based controller 12 may be
  • the one or more signals received by learning-based controller 12 may comprise content-rich signals in connection with autonomous industrial process 40, such as may include at least one stream of a video signal, multimedia signals, etc.
  • a safety-preserving controller 14 may be coupled to receive to receive one or more signals indicative of a potentially dangerous condition in the environment of autonomous industrial process 40.
  • the one or more signals received by safety-preserving controller 14 may comprise signal -content targeted to detect the potentially dangerous condition in the environment of the autonomous industrial process.
  • safety-preserving controller 14 may be a feedback-based controller, where, for example, fast and robust convergence may be guaranteed to a minimal error condition, whereas in a learning-based controller, such as learning-based controller 12, fast convergence may be difficult to guarantee in a few iterations.
  • outputs generated by safety-preserving controller 14 may be optionally supplied to learning-based controller 12 by way of interconnect line 27.
  • a hierarchical-based processor 16 may be arranged with a hierarchical control framework including a safe set control 17 (FIG. 2) suitable for the environment of autonomous industrial process 40.
  • hierarchical-based processor 16 may be arranged to receive a respective learning control output signal 18 from learning-based controller 12 and a safety-preserving control output signal 20 from safety-preserving controller 14.
  • hierarchical-based processor 16 may be arranged to perform a joint processing of the received control output signals from learning-based controller 12 and from safety-preserving controller 14 subject to the safe set control of the hierarchical control framework. In one non-limiting embodiment, hierarchical-based processor 16 may be configured to generate a control signal based on the safe set control of the hierarchical control framework. [0025] In one non-limiting embodiment, hierarchical-based processor 16 may be configured to transmit, such as over a network 22, the control signal to autonomous machine 30. In response to the transmitted control signal, autonomous machine 30 may execute one or more actions configured to ensure a safe operation in the environment in connection with autonomous industrial process 40.
  • the joint processing of the received control output signals from learning-based controller 12 and from safety-preserving controller 14 comprises configuring hierarchical-based processor 16 to constrain the respective learning control output signal from learning-based controller 12 based on the safety-preserving control output signal from safety preserving controller 14.
  • the constraint of the respective learning control output signal from the learning-based controller is configured to meet the safe set control of the hierarchical control framework.
  • safe set control 17 may be a predefined, fixed safe set control
  • the joint processing in hierarchical-based processor 16 of the control signal outputs 18, 20 from learning-based controller 12 and from safety-preserving controller 14 may involve configuring hierarchical-based processor 16 to project (schematically represented by line 19) the control output signal (schematically represented by cross 21) from learning-based controller 12 onto the predefined, fixed safe set control 17 toward the control output signal (schematically represented by cross 25) from safety-preserving controller 20.
  • star symbol 23 illustrates an example control signal resulting from such a projection that meets safe set control 17. It will be appreciated that the configuration of the safe set control illustrated in FIG.
  • the safe set control may be an updatable safe set control, and the methodology may involve configuring safety-preserving controller 14 to perform a periodic updating of such updatable safe set control.
  • disclosed embodiments may be used for robotic loading of a vehicle for transporting cargo, such as a truck, etc.
  • cargo such as a truck, etc.
  • robots in this type of application would have to be operated at reduced speeds, at least when humans share the environment with the robots and the magnitude of the inertial mass of the robot arm is sufficiently high to potentially create an unsafe condition.
  • an imaging-based system involving an imaging sensor, such as a camera, to monitor such an entrance, without limitation, the system can detect if people are present in any area of concern using, for example, a deep neural network algorithm using a content-rich signal provided by the imaging sensor. If no people are in the area, the robot can then operate safely at higher speeds, thus becoming a more efficient tool.
  • an imaging-based system may not detect a person. Additionally, one cannot guarantee that the neural network will consistently detect a person under dynamically changing environmental conditions, such as involving low lighting conditions, fog, scratches on the camera lens, certain forms of movement like crawling, etc.
  • the safety-preserving controller may be arranged to monitor signals from the second sensor layer to limit robot speed to a safe range based on an
  • the safe set control may be designed to ensure safe operation.
  • the emergency stop may be executed over a relatively short period of time but without generating high forces that could damage cargo presently handled by the robot. This ensures safe operation in the unlikely but not fully dismissible event that a person is missed by the imaging-based system.
  • a robot may be designed to quickly pick up and handle goods from random locations. Based on imaging-based
  • a learning-based controller may formulate a control strategy that results in electric currents for actuating one or more axes of a robotic arm to perform such pick up operations.
  • the learning-based controller may formulate a control strategy that involves a higher level of current to generate a higher magnitude of acceleration and thus enable the robot to arrive in time at the farther away location of the object.
  • the learning-based algorithm can potentially request a current level that is not safe for the robot motors and/or conducting wires, which can result in robot damage or even a fire.
  • the safety preserving controller may be arranged to monitor current levels and limit such levels to a safe range based on an appropriate safe set control.
  • safety may be achieved by way of a safety-preserving controller, such as a feedback-based controller, with access to safety-critical process data, which may serve as the basis to appropriately constrain control actions of a learning- based controller to just safe control actions.
  • a safety-preserving controller such as a feedback-based controller
  • Disclosed embodiments may be applied in a wide range of applications involving autonomous machines, including without limitation locomotion (e.g., wheeled, legged, flying, swimming, and crawling robots), manipulation (both arms and hands), tracking, navigation, mapping, process engineering applications, where one or more smart machines autonomously control a chemical process and make decisions about, temperatures, pressures and flow rates, etc.
  • locomotion e.g., wheeled, legged, flying, swimming, and crawling robots
  • manipulation both arms and hands
  • tracking navigation, mapping
  • process engineering applications where one or more smart machines autonomously control a chemical process and make decisions about, temperatures, pressures and flow rates, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Automation & Control Theory (AREA)
  • General Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Robotics (AREA)
  • Mechanical Engineering (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Game Theory and Decision Science (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Business, Economics & Management (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Electromagnetism (AREA)
  • Manipulator (AREA)

Abstract

L'invention concerne un système et un procédé informatisés. Un dispositif de commande basé sur l'apprentissage (12) peut être couplé pour recevoir un ou plusieurs signaux en relation avec un processus industriel autonome (40). Un dispositif de commande de préservation de sécurité (14) peut être couplé pour recevoir un ou plusieurs signaux indiquant une condition potentiellement dangereuse dans un environnement du processus industriel autonome. Un processeur à base hiérarchique (16) peut être agencé avec une commande d'ensemble sûre pour l'environnement du processus industriel autonome. Un processeur à base hiérarchique (16) peut être configuré pour générer un signal de commande sur la base de la commande d'ensemble sûre. Le signal de commande peut être transmis à une machine autonome (30) fonctionnant dans l'environnement du processus industriel autonome pour effectuer le processus industriel autonome, et en réponse au signal de commande transmis, la machine autonome peut exécuter une ou plusieurs actions configurées pour assurer un fonctionnement sûr dans l'environnement en relation avec le processus industriel autonome.
PCT/US2019/017616 2019-02-12 2019-02-12 Système et procédé informatisés configurés avec des contraintes hiérarchiques pour assurer le fonctionnement sûr d'une machine autonome WO2020167291A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/US2019/017616 WO2020167291A1 (fr) 2019-02-12 2019-02-12 Système et procédé informatisés configurés avec des contraintes hiérarchiques pour assurer le fonctionnement sûr d'une machine autonome

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/US2019/017616 WO2020167291A1 (fr) 2019-02-12 2019-02-12 Système et procédé informatisés configurés avec des contraintes hiérarchiques pour assurer le fonctionnement sûr d'une machine autonome

Publications (1)

Publication Number Publication Date
WO2020167291A1 true WO2020167291A1 (fr) 2020-08-20

Family

ID=65724499

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2019/017616 WO2020167291A1 (fr) 2019-02-12 2019-02-12 Système et procédé informatisés configurés avec des contraintes hiérarchiques pour assurer le fonctionnement sûr d'une machine autonome

Country Status (1)

Country Link
WO (1) WO2020167291A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113110069A (zh) * 2021-05-24 2021-07-13 武汉大学 一种基于磁悬浮平面电机迭代神经网络鲁棒控制方法

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017197170A1 (fr) * 2016-05-12 2017-11-16 The Regents Of The University Of California Commande sécurisée d'une entité autonome en présence d'agents intelligents
US20180247160A1 (en) * 2017-02-27 2018-08-30 Mohsen Rohani Planning system and method for controlling operation of an autonomous vehicle to navigate a planned path

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017197170A1 (fr) * 2016-05-12 2017-11-16 The Regents Of The University Of California Commande sécurisée d'une entité autonome en présence d'agents intelligents
US20180247160A1 (en) * 2017-02-27 2018-08-30 Mohsen Rohani Planning system and method for controlling operation of an autonomous vehicle to navigate a planned path

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
CHANGLIU LIU ET AL: "Robot Safe Interaction System for Intelligent Industrial Co-Robots", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 12 August 2018 (2018-08-12), XP081256209 *
CHANGLIU LIU ET AL: "SERoCS: Safe and Efficient Robot Collaborative Systems for Next Generation Intelligent Industrial Co-Robots", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 21 September 2018 (2018-09-21), XP081194080 *
MAEDA Y ET AL: "Hierarchical control for autonomous mobile robots with behavior-decision fuzzy algorithm", PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION NICE, MAY 12 - 14, 1992; [PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION], LOS ALAMITOS, IEEE COMP. SOC. PRESS, US, vol. CONF. 8, 12 May 1992 (1992-05-12), pages 117 - 122, XP010027947, ISBN: 978-0-8186-2720-0, DOI: 10.1109/ROBOT.1992.220325 *
NICOLA M. CERIANI; ANDREA M. ZANCHETTIN; PAOLO ROCCO; ANDREAS STOLT; ANDERS ROBERTSSON: "Reactive Task Adaptation Based on Hierarchical Constraints Classification for Safe Industrial Robots", IEEE/ASME TRANSACTIONS ON MECHATRONICS, vol. 20, no. 6, December 2015 (2015-12-01), XP011587897, DOI: doi:10.1109/TMECH.2015.2415462

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113110069A (zh) * 2021-05-24 2021-07-13 武汉大学 一种基于磁悬浮平面电机迭代神经网络鲁棒控制方法

Similar Documents

Publication Publication Date Title
Bi et al. Safety assurance mechanisms of collaborative robotic systems in manufacturing
Zanchettin et al. Safety in human-robot collaborative manufacturing environments: Metrics and control
Liu et al. Algorithmic safety measures for intelligent industrial co-robots
Chaimowicz et al. An architecture for tightly coupled multi-robot cooperation
Haddadin et al. Towards the robotic co-worker
Avanzini et al. Constrained model predictive control for mobile robotic manipulators
Abi-Farraj et al. A learning-based shared control architecture for interactive task execution
Pereira et al. Safety control of robots under computed torque control using reachable sets
Annem et al. Towards remote teleoperation of a semi-autonomous mobile manipulator system in machine tending tasks
Parusel et al. Modular state-based behavior control for safe human-robot interaction: A lightweight control architecture for a lightweight robot
US10335946B2 (en) Compositional impedance programming for robots
Ding et al. Collaborative behavior design of industrial robots for multiple human-robot collaboration
Polverini et al. A constraint-based programming approach for robotic assembly skills implementation
Magrini et al. Human-robot coexistence and contact handling with redundant robots
Kuts et al. COLLABORATIVE WORK BETWEEN HUMAN AND INDUSTRIAL ROBOT IN MANUFACTURING BY ADVANCED SAFETY MONITORING SYSTEM.
Mikhel et al. Collision driven multi scenario approach for human collaboration with industrial robot
WO2020167291A1 (fr) Système et procédé informatisés configurés avec des contraintes hiérarchiques pour assurer le fonctionnement sûr d'une machine autonome
Mukhtar et al. A control structure for ambidextrous robot arm based on Multiple Adaptive Neuro‐Fuzzy Inference System
Shu et al. A versatile humanoid robot platform for dexterous manipulation and human–robot collaboration
Selvaggio et al. Towards a self-collision aware teleoperation framework for compound robots
Lee et al. Intermediate desired value approach for continuous transition among multiple tasks of robots
Han et al. A continuous task transition algorithm for operational space control framework
Lee et al. Reliable software architecture design with EtherCAT for a rescue robot
Brending et al. Certifiable software architecture for human robot collaboration in industrial production environments
Van Dijk et al. Path-constrained motion planning for robotics based on kinematic constraints

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19710179

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19710179

Country of ref document: EP

Kind code of ref document: A1