CN112365004A - Robot autonomous anomaly restoration skill learning method and system - Google Patents
Robot autonomous anomaly restoration skill learning method and system Download PDFInfo
- Publication number
- CN112365004A CN112365004A CN202011367295.8A CN202011367295A CN112365004A CN 112365004 A CN112365004 A CN 112365004A CN 202011367295 A CN202011367295 A CN 202011367295A CN 112365004 A CN112365004 A CN 112365004A
- Authority
- CN
- China
- Prior art keywords
- abnormal
- skill
- robot
- motor
- motion
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000002159 abnormal effect Effects 0.000 claims abstract description 128
- 238000012544 monitoring process Methods 0.000 claims abstract description 15
- 230000003044 adaptive effect Effects 0.000 claims abstract description 11
- 230000005856 abnormality Effects 0.000 claims description 27
- 238000006243 chemical reaction Methods 0.000 claims description 22
- 238000001514 detection method Methods 0.000 claims description 14
- 230000001052 transient effect Effects 0.000 claims description 14
- 230000006399 behavior Effects 0.000 claims description 13
- 238000009826 distribution Methods 0.000 claims description 13
- 230000002567 autonomic effect Effects 0.000 claims description 12
- 208000012661 Dyskinesia Diseases 0.000 claims description 6
- 238000012512 characterization method Methods 0.000 claims description 4
- 238000003745 diagnosis Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 230000001737 promoting effect Effects 0.000 abstract description 2
- 206010000117 Abnormal behaviour Diseases 0.000 description 13
- 238000010586 diagram Methods 0.000 description 8
- 230000002085 persistent effect Effects 0.000 description 6
- 230000000007 visual effect Effects 0.000 description 4
- 238000003860 storage Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Manipulator (AREA)
Abstract
The invention discloses a robot autonomous anomaly repairing skill learning method and a system, wherein the method comprises the following steps: pre-defining a human demonstration track of a robot when the robot executes a complex task; acquiring multi-modal sensing information of the robot when the robot executes the complex task, monitoring by using the multi-modal sensing information, and acquiring abnormal motor skills; and enabling an adaptive motion restoration strategy to carry out autonomous restoration on the abnormal motor skill based on the abnormal state type of the abnormal motor skill. In the embodiment of the invention, the abnormal types of the robot in the complex task execution can be distinguished, and the corresponding repair strategy can be formulated, which is beneficial to promoting the robot to realize longer-term autonomous operation.
Description
Technical Field
The invention relates to the technical field of robot skill learning, in particular to a robot autonomous anomaly repairing skill learning method and system.
Background
With the continuous promotion of the application range and the depth of the robot, the existing intelligent technology cannot meet the requirements, and the cooperative operation of the robot and the robot is the most effective solution. The man-machine cooperation means that the robot senses through multiple sensors and cooperates with people to complete various fine and complex operation tasks, and the method is widely applied to the fields of intelligent manufacturing, logistics storage, medical service and the like at present. However, in the human-computer cooperation environment, due to a program error of the robot, sensor noise, human misoperation and the like, abnormal events such as a collision between the robot and the environment, a collision between the robot and the human, an object operation failure and the like are caused, and a task failure is possibly caused, even a human or a robot body is damaged. Therefore, the research of the robot autonomous anomaly repair skill learning method is significant for different types of abnormal events.
Human-computer cooperation-oriented abnormal repair should exert human expectation on robot motion, and a human-computer cooperation idea of human-centered can be more embodied through a human-assisted robot abnormal repair strategy. However, conventionally, a rule is preset manually or a motion planning method of the robot itself is relied on to repair the robot abnormality, and the abnormality type and the characteristics of a human-computer cooperation system are not considered, so that the requirements of practical application cannot be met.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a robot autonomous anomaly repairing skill learning method and system, which can distinguish anomaly types encountered by a robot when the robot executes complex tasks and formulate a corresponding repairing strategy, and are beneficial to promoting the robot to realize longer-term autonomous operation.
In order to solve the above problems, the present invention provides a robot autonomic abnormality recovery skill learning method, including:
pre-defining a human demonstration track of a robot when the robot executes a complex task;
acquiring multi-modal sensing information of the robot when the robot executes the complex task, monitoring by using the multi-modal sensing information, and acquiring abnormal motor skills;
and enabling an adaptive motion restoration strategy to carry out autonomous restoration on the abnormal motor skill based on the abnormal state type of the abnormal motor skill.
Optionally, the pre-defining a human demonstration trajectory of the robot when performing the complex task includes:
n motor skills required by the robot when the robot executes a complex task are predefined;
and utilizing a Finite State Machine (FSM) to carry out serialized characterization on the types and the execution sequence of the N motor skills, and generating corresponding N motor skill sequences.
Optionally, the obtaining of the multi-modal sensing information of the robot during the execution of the complex task and monitoring by using the multi-modal sensing information include:
acquiring multi-modal sensing information of the robot when the robot executes the complex task, preprocessing the multi-modal sensing information, and extracting a low-dimensional feature vector;
establishing an abnormality detection model by using the N motor skill sequences, inputting the low-dimensional feature vector into the abnormality detection model for monitoring, and judging whether the robot is in an abnormal state;
if yes, inputting the low-dimensional feature vector into a multi-category abnormal classifier for diagnosis, and acquiring abnormal motor skills of the robot;
and if not, returning to obtain the multi-mode sensing information of the robot in the complex task executing process again.
Optionally, the enabling of the adaptive motion restoration strategy to autonomously restore the abnormal motor skill based on the abnormal state type of the abnormal motor skill includes:
enabling an instantaneous movement redo repair strategy to repair the abnormal movement skill again based on the fact that the abnormal state type of the abnormal movement skill is an instantaneous state;
and starting a continuous motion adjustment and repair strategy to perform continuous adjustment on the abnormal motor skill based on the fact that the abnormal state type of the abnormal motor skill is a continuous state.
Optionally, the transient exercise redo repair strategy includes:
recording a last motor skill relative to the abnormal motor skill based on the human demonstration track;
and updating the task scheduling directed graph by adopting polynomial probability distribution learning, and adding a conversion node for motion redoing between the abnormal motor skill and the previous motor skill to complete the parameter restoration of the abnormal motor skill.
Optionally, the continuous motion adjustment repair strategy includes:
recording a next motor skill relative to the abnormal motor skill based on the human demonstration trajectory;
and updating the task scheduling directed graph based on the dynamic motion primitive learning model, adding a motion adjustment conversion node between the abnormal motion skill and the next motion skill, generating a human demonstration repair behavior, and completing parameter adjustment of the next motion skill.
In addition, an embodiment of the present invention further provides a system for learning skills for repairing robot autonomic abnormality, where the system includes:
the motion track definition module is used for predefining a human demonstration track of the robot when the robot executes a complex task;
the abnormal skill monitoring module is used for acquiring multi-modal sensing information of the robot when the robot executes the complex task, monitoring by utilizing the multi-modal sensing information and acquiring abnormal motor skills;
and the abnormal skill repairing module is used for starting an adaptive motion repairing strategy to carry out autonomous repairing on the abnormal motor skill based on the abnormal state type of the abnormal motor skill.
Optionally, the abnormal skill restoration module is configured to enable an instantaneous movement redo restoration strategy to restore the abnormal motor skill based on that the abnormal state type of the abnormal motor skill is an instantaneous state; and starting a continuous motion adjustment and repair strategy to perform continuous adjustment on the abnormal motor skill based on the fact that the abnormal state type of the abnormal motor skill is a continuous state.
Optionally, the transient exercise redo repair strategy includes:
recording a last motor skill relative to the abnormal motor skill based on the human demonstration track;
and updating the task scheduling directed graph by adopting polynomial probability distribution learning, and adding a conversion node for motion redoing between the abnormal motor skill and the previous motor skill to complete the parameter restoration of the abnormal motor skill.
Optionally, the continuous motion adjustment repair strategy includes:
recording a next motor skill relative to the abnormal motor skill based on the human demonstration trajectory;
and updating the task scheduling directed graph based on the dynamic motion primitive learning model, adding a motion adjustment conversion node between the abnormal motion skill and the next motion skill, generating a human demonstration repair behavior, and completing parameter adjustment of the next motion skill.
In the embodiment of the invention, the abnormal types encountered by the robot when executing complex tasks can be effectively distinguished, and adaptive behavior repairing strategies are formulated aiming at different types of abnormal behaviors by relying on a polynomial probability distribution learning and dynamic motion primitive learning model, so that the repairing skills learned by the robot have certain expansibility and generalization, the robot is promoted to realize longer-term autonomous operation, and the requirement of practical application can be met.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a robot autonomous anomaly repair skill learning method in an embodiment of the present invention;
FIG. 2 is a schematic diagram of a process for implementing a motion redo repair technique in an embodiment of the present invention;
FIG. 3 is a schematic diagram of a process for implementing a motion adjustment repair technique in an embodiment of the present invention;
fig. 4 is a schematic structural composition diagram of the robot autonomous anomaly restoration skill learning system in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
Referring to fig. 1, fig. 1 is a schematic flowchart illustrating a robot autonomic abnormality repair skill learning method according to an embodiment of the present invention.
As shown in fig. 1, a method for learning a robot autonomic abnormality repair skill, the method comprising:
s101, predefining a human demonstration track of the robot when the robot executes a complex task;
the implementation process of the invention comprises the following steps: (1) n motor skills required by the robot when the robot executes a complex task are predefined; (2) and utilizing a Finite State Machine (FSM) to carry out serialized characterization on the types and the execution sequence of the N motor skills, and generating corresponding N motor skill sequences.
S102, acquiring multi-modal sensing information of the robot when the robot executes the complex task, monitoring by using the multi-modal sensing information, and acquiring abnormal motor skills;
the implementation process of the invention comprises the following steps: (1) acquiring multi-modal sensing information of the robot when the robot executes the complex task, preprocessing the multi-modal sensing information, and extracting a low-dimensional feature vector; (2) establishing an abnormality detection model by using the N motor skill sequences, inputting the low-dimensional feature vector into the abnormality detection model for monitoring, and judging whether the robot is in an abnormal state, wherein the corresponding judgment result is as follows: if the robot is in an abnormal state, inputting the low-dimensional feature vector into a multi-category abnormal classifier for diagnosis, and acquiring abnormal motor skills of the robot; and if the robot is in a normal state, returning to acquire the multi-modal sensing information of the robot in the process of executing the complex task again so as to realize detection of each motor skill of the robot in the process of executing the whole complex task.
S103, based on the abnormal state type of the abnormal motor skill, enabling an adaptive motor repair strategy to perform autonomous repair on the abnormal motor skill.
In the embodiment of the invention, based on the fact that the abnormal state type of the abnormal motor skill is a transient state, a transient motor redo repair strategy is started to repair the abnormal motor skill again; or, based on that the abnormal state type of the abnormal motor skill is a persistent state, a persistent motor adjustment repair strategy is started to perform continuous adjustment on the abnormal motor skill. The instantaneous movement redo repairing strategy mainly aims at solving instantaneous abnormalities such as human collision, object sliding and the like and completes repairing of current abnormal movement skills; the continuous motion adjustment and repair strategy mainly aims at solving continuous abnormalities such as tool collision, wall collision and the like and completing adjustment of the next continuous motion skill. In addition, the continuous motion adjustment and repair strategy is executed on the basis of the instantaneous motion redo repair strategy, that is, the continuous motion adjustment and repair strategy can be started only under the condition that the current abnormal motor skills cannot be repaired (the number of attempts is set to be not less than two in the embodiment of the invention) after the instantaneous motion redo repair strategy is repeatedly executed, so that the repair workload of the robot is prevented from being increased.
Specifically, with reference to fig. 2, a schematic process diagram for implementing a motor redo restoration technique is shown, where M represents a motor skill, I represents an abnormality detection model, and V represents a visual sensor, and the transient motor redo restoration strategy includes:
(1) recording a last motor skill relative to the abnormal motor skill based on the human demonstration track;
in practice, it is assumed that the current motor skills are given
(
Is a starting node,
As the target node), the last motor skill (i.e., the motor skill selected to be redone) is
(
Is a starting node,
Is a target node) and is performing the current motor skill BiThere is any transient abnormal behavior FxThis occurs.
(2) And updating the task scheduling directed graph by adopting polynomial probability distribution learning, and adding a conversion node for motion redoing between the abnormal motor skill and the previous motor skill to complete the parameter restoration of the abnormal motor skill.
In practice, by applying to the currentMotor skill BiAnd the last motor skill B*A conversion node R for motion redoing is added betweenR：
Learning of conversion parameters using polynomial probability distributions
Is calculated and finally can be calculated by the conversion node RRIs determined to redo the last motor skill B*The success rate of. Wherein the conversion parameter
Is that the human being is acting for abnormal behavior FxPost-occurrence redoing motor skills B*E.g. setting abnormal behaviour FxThe random sample vector of the transition motion probability distribution of
K is the movement from the robot to the abnormal behavior FxCurrent motor skill of the place BiTotal number of skills of, Ni(i 1, 2.., K.) is the number of successful redo of the corresponding i-th motor skill, at which time motor skill B is redone for N independent redoes*Which converts node RRThe probability mass function of (a) is:
in the formula: thetaiProbability of being selected for the ith motor skill, and θi∈[0，1]，
Specifically, with reference to the schematic process diagram of fig. 3, in which M represents a motor skill, I represents an abnormality detection model, and V represents a visual sensor, the continuous motor adjustment repair strategy includes:
(1) recording a next motor skill relative to the abnormal motor skill based on the human demonstration trajectory;
in practice, it is assumed that the current motor skills are given
(
Is a starting node,
Is a target node) and is performing the current motor skill BjThere is any persistent abnormal behavior FyTakes place while recording the next motor skill's adjustment node
(2) And updating the task scheduling directed graph based on the dynamic motion primitive learning model, adding a motion adjustment conversion node between the abnormal motion skill and the next motion skill, generating a human demonstration repair behavior, and completing parameter adjustment of the next motion skill.
In practice, by applying the current motor skill BjAdjusting the node with the next motor skill
A motion-adjusted conversion node R is added betweenA：
Generating a human demonstration repair behavior based on the dynamic motion primitive learning model
Recombining said human exemplary repair behavior BhTo determine the end pose point P of
Thereby adjusting the node of the next motor skill
Correction is performed.
In the embodiment of the invention, the abnormal types encountered by the robot when executing complex tasks can be effectively distinguished, and adaptive behavior repairing strategies are formulated aiming at different types of abnormal behaviors by relying on a polynomial probability distribution learning and dynamic motion primitive learning model, so that the repairing skills learned by the robot have certain expansibility and generalization, the robot is promoted to realize longer-term autonomous operation, and the requirement of practical application can be met.
Examples
Referring to fig. 4, fig. 4 is a schematic structural composition diagram of a robot autonomic abnormality repair skill learning system in an embodiment of the present invention.
As shown in fig. 4, a robotic autonomic abnormality repair skills learning system, the system comprising:
a motion track definition module 201, configured to pre-define a human demonstration track of the robot when performing a complex task;
the implementation process of the invention comprises the following steps: (1) n motor skills required by the robot when the robot executes a complex task are predefined; (2) and utilizing a Finite State Machine (FSM) to carry out serialized characterization on the types and the execution sequence of the N motor skills, and generating corresponding N motor skill sequences.
The abnormal skill monitoring module 202 is configured to acquire multi-modal sensing information of the robot when the robot executes the complex task, monitor the multi-modal sensing information, and acquire an abnormal motor skill;
the implementation process of the invention comprises the following steps: (1) acquiring multi-modal sensing information of the robot when the robot executes the complex task, preprocessing the multi-modal sensing information, and extracting a low-dimensional feature vector; (2) establishing an abnormality detection model by using the N motor skill sequences, inputting the low-dimensional feature vector into the abnormality detection model for monitoring, and judging whether the robot is in an abnormal state, wherein the corresponding judgment result is as follows: if the robot is in an abnormal state, inputting the low-dimensional feature vector into a multi-category abnormal classifier for diagnosis, and acquiring abnormal motor skills of the robot; and if the robot is in a normal state, returning to acquire the multi-modal sensing information of the robot in the process of executing the complex task again so as to realize detection of each motor skill of the robot in the process of executing the whole complex task.
And the abnormal skill repairing module 203 is used for enabling an adaptive motion repairing strategy to perform autonomous repairing on the abnormal motor skill based on the abnormal state type of the abnormal motor skill.
In the embodiment of the invention, based on the fact that the abnormal state type of the abnormal motor skill is a transient state, a transient motor redo repair strategy is started to repair the abnormal motor skill again; or, based on that the abnormal state type of the abnormal motor skill is a persistent state, a persistent motor adjustment repair strategy is started to perform continuous adjustment on the abnormal motor skill. The instantaneous movement redo repairing strategy mainly aims at solving instantaneous abnormalities such as human collision, object sliding and the like and completes repairing of current abnormal movement skills; the continuous motion adjustment and repair strategy mainly aims at solving continuous abnormalities such as tool collision, wall collision and the like and completing adjustment of the next continuous motion skill. In addition, the continuous motion adjustment and repair strategy is executed on the basis of the instantaneous motion redo repair strategy, that is, the continuous motion adjustment and repair strategy can be started only under the condition that the current abnormal motor skills cannot be repaired (the number of attempts is set to be not less than two in the embodiment of the invention) after the instantaneous motion redo repair strategy is repeatedly executed, so that the repair workload of the robot is prevented from being increased.
Specifically, with reference to fig. 2, a schematic process diagram for implementing a motor redo restoration technique is shown, where M represents a motor skill, I represents an abnormality detection model, and V represents a visual sensor, and the transient motor redo restoration strategy includes:
(1) recording a last motor skill relative to the abnormal motor skill based on the human demonstration track;
in practice, it is assumed that the current motor skills are given
(
Is a starting node,
As the target node), the last motor skill (i.e., the motor skill selected to be redone) is
(
Is a starting node,
Is a target node) and is performing the current motor skill BiThere is any transient abnormal behavior FxThis occurs.
(2) And updating the task scheduling directed graph by adopting polynomial probability distribution learning, and adding a conversion node for motion redoing between the abnormal motor skill and the previous motor skill to complete the parameter restoration of the abnormal motor skill.
In practice, by applying the current motor skill BiAnd the previous oneMotor skill B*A conversion node R for motion redoing is added betweenR：
Learning of conversion parameters using polynomial probability distributions
Is calculated and finally can be calculated by the conversion node RRIs determined to redo the last motor skill B*The success rate of. Wherein the conversion parameter
Is that the human being is acting for abnormal behavior FxPost-occurrence redoing motor skills B*E.g. setting abnormal behaviour FxThe random sample vector of the transition motion probability distribution of
K is the movement from the robot to the abnormal behavior FxCurrent motor skill of the place BiTotal number of skills of, Ni(i 1, 2.., K.) is the number of successful redo of the corresponding i-th motor skill, at which time motor skill B is redone for N independent redoes*Which converts node RRThe probability mass function of (a) is:
in the formula: thetaiProbability of being selected for the ith motor skill, and θi∈[0，1]，
Specifically, with reference to the schematic process diagram of fig. 3, in which M represents a motor skill, I represents an abnormality detection model, and V represents a visual sensor, the continuous motor adjustment repair strategy includes:
(1) recording a next motor skill relative to the abnormal motor skill based on the human demonstration trajectory;
in practice, it is assumed that the current motor skills are given
(
Is a starting node,
Is a target node) and is performing the current motor skill BjThere is any persistent abnormal behavior FyTakes place while recording the next motor skill's adjustment node
(2) And updating the task scheduling directed graph based on the dynamic motion primitive learning model, adding a motion adjustment conversion node between the abnormal motion skill and the next motion skill, generating a human demonstration repair behavior, and completing parameter adjustment of the next motion skill.
In practice, by applying the current motor skill BjAdjusting the node with the next motor skill
A motion-adjusted conversion node R is added betweenA：
Generating a human demonstration repair behavior based on the dynamic motion primitive learning model
Recombining said human exemplary repair behavior BhTo determine the end pose point P of
Thereby adjusting the node of the next motor skill
Correction is performed.
In the embodiment of the invention, the abnormal types encountered by the robot when executing complex tasks can be effectively distinguished, and adaptive behavior repairing strategies are formulated aiming at different types of abnormal behaviors by relying on a polynomial probability distribution learning and dynamic motion primitive learning model, so that the repairing skills learned by the robot have certain expansibility and generalization, the robot is promoted to realize longer-term autonomous operation, and the requirement of practical application can be met.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable storage medium, and the storage medium may include: a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic or optical disk, or the like.
The robot autonomous anomaly repairing skill learning method and system provided by the embodiment of the invention are described in detail, a specific embodiment is adopted in the text to explain the principle and the implementation mode of the invention, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A method of robotic autonomic anomaly repair skill learning, the method comprising:
pre-defining a human demonstration track of a robot when the robot executes a complex task;
acquiring multi-modal sensing information of the robot when the robot executes the complex task, monitoring by using the multi-modal sensing information, and acquiring abnormal motor skills;
and enabling an adaptive motion restoration strategy to carry out autonomous restoration on the abnormal motor skill based on the abnormal state type of the abnormal motor skill.
2. The method of robot autonomous anomaly repair skill learning according to claim 1, wherein said predefining human demonstration trajectories of robots in performing complex tasks comprises:
n motor skills required by the robot when the robot executes a complex task are predefined;
and utilizing a Finite State Machine (FSM) to carry out serialized characterization on the types and the execution sequence of the N motor skills, and generating corresponding N motor skill sequences.
3. The method for learning the skill for autonomously recovering the abnormality of the robot according to claim 2, wherein the acquiring multi-modal sensing information of the robot when the robot performs the complex task and monitoring by using the multi-modal sensing information includes:
acquiring multi-modal sensing information of the robot when the robot executes the complex task, preprocessing the multi-modal sensing information, and extracting a low-dimensional feature vector;
establishing an abnormality detection model by using the N motor skill sequences, inputting the low-dimensional feature vector into the abnormality detection model for monitoring, and judging whether the robot is in an abnormal state;
if yes, inputting the low-dimensional feature vector into a multi-category abnormal classifier for diagnosis, and acquiring abnormal motor skills of the robot;
and if not, returning to obtain the multi-mode sensing information of the robot in the complex task executing process again.
4. The method of claim 3, wherein enabling an adapted motion restoration strategy to autonomously restore the abnormal motor skill based on the abnormal state type of the abnormal motor skill comprises:
enabling an instantaneous movement redo repair strategy to repair the abnormal movement skill again based on the fact that the abnormal state type of the abnormal movement skill is an instantaneous state;
and starting a continuous motion adjustment and repair strategy to perform continuous adjustment on the abnormal motor skill based on the fact that the abnormal state type of the abnormal motor skill is a continuous state.
5. The method of robotic autonomic anomaly repair skills learning according to claim 4, wherein said transient motor redo repair strategy comprises:
recording a last motor skill relative to the abnormal motor skill based on the human demonstration track;
and updating the task scheduling directed graph by adopting polynomial probability distribution learning, and adding a conversion node for motion redoing between the abnormal motor skill and the previous motor skill to complete the parameter restoration of the abnormal motor skill.
6. The method of robotic autonomic anomaly repair skills learning according to claim 4, wherein said continuous motion adjustment repair strategy comprises:
recording a next motor skill relative to the abnormal motor skill based on the human demonstration trajectory;
and updating the task scheduling directed graph based on the dynamic motion primitive learning model, adding a motion adjustment conversion node between the abnormal motion skill and the next motion skill, generating a human demonstration repair behavior, and completing parameter adjustment of the next motion skill.
7. A robotic autonomic abnormality repair skills learning system, the system comprising:
the motion track definition module is used for predefining a human demonstration track of the robot when the robot executes a complex task;
the abnormal skill monitoring module is used for acquiring multi-modal sensing information of the robot when the robot executes the complex task, monitoring by utilizing the multi-modal sensing information and acquiring abnormal motor skills;
and the abnormal skill repairing module is used for starting an adaptive motion repairing strategy to carry out autonomous repairing on the abnormal motor skill based on the abnormal state type of the abnormal motor skill.
8. The system of claim 7, wherein the abnormal skill rehabilitation module is configured to initiate a transient motor rework rehabilitation strategy to rehabilitate the abnormal motor skill based on the abnormal state type of the abnormal motor skill being a transient state; and starting a continuous motion adjustment and repair strategy to perform continuous adjustment on the abnormal motor skill based on the fact that the abnormal state type of the abnormal motor skill is a continuous state.
9. The robotic autonomic anomaly repair skills learning system of claim 8, wherein said transient motor redo repair strategy comprises:
recording a last motor skill relative to the abnormal motor skill based on the human demonstration track;
and updating the task scheduling directed graph by adopting polynomial probability distribution learning, and adding a conversion node for motion redoing between the abnormal motor skill and the previous motor skill to complete the parameter restoration of the abnormal motor skill.
10. The robotic autonomic abnormality repair skills learning system according to claim 8, wherein the continuous motion adjustment repair strategy comprises:
recording a next motor skill relative to the abnormal motor skill based on the human demonstration trajectory;
and updating the task scheduling directed graph based on the dynamic motion primitive learning model, adding a motion adjustment conversion node between the abnormal motion skill and the next motion skill, generating a human demonstration repair behavior, and completing parameter adjustment of the next motion skill.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011367295.8A CN112365004B (en) | 2020-11-27 | 2020-11-27 | Robot autonomous anomaly repair skill learning method and system |
| PCT/CN2021/133685 WO2022111652A1 (en) | 2020-11-27 | 2021-11-26 | Robot autonomous abnormality repair skill learning method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011367295.8A CN112365004B (en) | 2020-11-27 | 2020-11-27 | Robot autonomous anomaly repair skill learning method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112365004A true CN112365004A (en) | 2021-02-12 |
| CN112365004B CN112365004B (en) | 2024-05-31 |
Family
ID=74536249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011367295.8A Active CN112365004B (en) | 2020-11-27 | 2020-11-27 | Robot autonomous anomaly repair skill learning method and system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112365004B (en) |
| WO (1) | WO2022111652A1 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114474150A (en) * | 2021-12-31 | 2022-05-13 | 苏州通标检测科技有限公司 | Stability testing method and system for seven-axis robot |
| WO2022111652A1 (en) * | 2020-11-27 | 2022-06-02 | 广东省科学院智能制造研究所 | Robot autonomous abnormality repair skill learning method and system |
| CN115631401A (en) * | 2022-12-22 | 2023-01-20 | 广东省科学院智能制造研究所 | Robot autonomous grabbing skill learning system and method based on visual perception |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107671887A (en) * | 2017-08-22 | 2018-02-09 | 广东美的智能机器人有限公司 | Robot self-test control method, robot and dispatch server |
| CN110370317A (en) * | 2019-07-24 | 2019-10-25 | 广东工业大学 | Robot restorative procedure and device |
| CN110633189A (en) * | 2019-09-20 | 2019-12-31 | 深圳市广通软件有限公司 | Intelligent operation and maintenance monitoring method and intelligent operation and maintenance monitoring system of IT system |
| US20200074349A1 (en) * | 2018-08-31 | 2020-03-05 | Cogitai, Inc. | Method and system for continual learning in an intelligent artificial agent |
| CN111761576A (en) * | 2020-06-15 | 2020-10-13 | 上海高仙自动化科技发展有限公司 | Health monitoring method and system, intelligent robot and readable storage medium |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108406767A (en) * | 2018-02-13 | 2018-08-17 | 华南理工大学 | Robot autonomous learning method towards man-machine collaboration |
| CN111444954A (en) * | 2020-03-24 | 2020-07-24 | 广东省智能制造研究所 | Robot autonomous assembly method based on multi-mode perception and learning |
| CN111890353B (en) * | 2020-06-24 | 2022-01-11 | 深圳市越疆科技有限公司 | Robot teaching track reproduction method and device and computer readable storage medium |
| CN112365004B (en) * | 2020-11-27 | 2024-05-31 | 广东省科学院智能制造研究所 | Robot autonomous anomaly repair skill learning method and system |
-
2020
- 2020-11-27 CN CN202011367295.8A patent/CN112365004B/en active Active
-
2021
- 2021-11-26 WO PCT/CN2021/133685 patent/WO2022111652A1/en unknown
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107671887A (en) * | 2017-08-22 | 2018-02-09 | 广东美的智能机器人有限公司 | Robot self-test control method, robot and dispatch server |
| US20200074349A1 (en) * | 2018-08-31 | 2020-03-05 | Cogitai, Inc. | Method and system for continual learning in an intelligent artificial agent |
| CN110370317A (en) * | 2019-07-24 | 2019-10-25 | 广东工业大学 | Robot restorative procedure and device |
| CN110633189A (en) * | 2019-09-20 | 2019-12-31 | 深圳市广通软件有限公司 | Intelligent operation and maintenance monitoring method and intelligent operation and maintenance monitoring system of IT system |
| CN111761576A (en) * | 2020-06-15 | 2020-10-13 | 上海高仙自动化科技发展有限公司 | Health monitoring method and system, intelligent robot and readable storage medium |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022111652A1 (en) * | 2020-11-27 | 2022-06-02 | 广东省科学院智能制造研究所 | Robot autonomous abnormality repair skill learning method and system |
| CN114474150A (en) * | 2021-12-31 | 2022-05-13 | 苏州通标检测科技有限公司 | Stability testing method and system for seven-axis robot |
| CN115631401A (en) * | 2022-12-22 | 2023-01-20 | 广东省科学院智能制造研究所 | Robot autonomous grabbing skill learning system and method based on visual perception |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2022111652A1 (en) | 2022-06-02 |
| CN112365004B (en) | 2024-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112365004A (en) | Robot autonomous anomaly restoration skill learning method and system | |
| Tsounis et al. | Deepgait: Planning and control of quadrupedal gaits using deep reinforcement learning | |
| CN108828944B (en) | Encoder fault diagnosis system and method based on improved PSO and SVM | |
| CN109159113B (en) | Robot operation method based on visual reasoning | |
| CN109492763B (en) | Automatic parking method based on reinforcement learning network training | |
| EP3812105B1 (en) | Artificial intelligence architecture for industrial welding | |
| Sadeghzadeh et al. | Self-learning visual servoing of robot manipulator using explanation-based fuzzy neural networks and Q-learning | |
| CN115781685B (en) | High-precision mechanical arm control method and system based on reinforcement learning | |
| CN115256395A (en) | Model uncertain robot safety control method based on control barrier function | |
| Yasutomi et al. | A peg-in-hole task strategy for holes in concrete | |
| CN114035550B (en) | Autonomous underwater robot actuating mechanism fault diagnosis method based on ESO | |
| CN113954069A (en) | Mechanical arm active fault-tolerant control method based on deep reinforcement learning | |
| CN111002302B (en) | Mechanical arm grabbing track planning method combining Gaussian mixture model and dynamic system | |
| CN116604532A (en) | Intelligent control method for upper limb rehabilitation robot | |
| CN112894808B (en) | Robot screwing valve system and method based on deep reinforcement learning | |
| Li et al. | A unified perspective on multiple shooting in differential dynamic programming | |
| CN116803635B (en) | Near-end strategy optimization training acceleration method based on Gaussian kernel loss function | |
| Young et al. | Formalized task characterization for human-robot autonomy allocation | |
| CN109352649A (en) | A kind of method for controlling robot and system based on deep learning | |
| Deng et al. | Learning human compliant behavior from demonstration for force-based robot manipulation | |
| CN111142378A (en) | Neural network optimization method of biped robot neural network controller | |
| Rojas et al. | Probabilistic state verification for snap assemblies using the relative-change-based hierarchical taxonomy | |
| Kapotoglu et al. | Robots avoid potential failures through experience-based probabilistic planning | |
| CN114594757A (en) | Visual path planning method for cooperative robot | |
| CN116880341B (en) | High-precision motion control system based on industrial Ethernet bus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |