DETAILED ACTION
The amendments filed 8/3/2026 have been entered. Claims 1-20 are pending.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-5, 7-14, and 16-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Evrard (“Control of humanoid robots to realize haptic tasks in collaboration with a human operator”).
Evrard teaches:
Re claim 1. A system for learning from demonstration (LfD) (section 3.1 “Programming by demonstration of collaborative tasks”, page 57) via physical human-robot-human (pHRH) interactions (section 3.2.1 “Hardware setup and controller”, second paragraph under “Hardware setup” on page 60; and Fig. 3.2 on page 61), comprising:
a memory storing one or more instructions (inherently part of computer from section C.3.6 on page 145); and
a processor executing one or more of the instructions stored on the memory to perform (inherently part of computer from section C.3.6 on page 145):
generating a command corresponding to an operation input received from a first human associated with a first interaction that is a physical human-robot (pHR) interaction with a first robot indicative of a desired command to be performed by the system for LfD via pHRH interactions subject to one or more constraints (velocity command from the teacher in Figure 3.2 on page 61; and second paragraph under “Hardware setup” on page 60: “During the demonstrations, a human operator (the teacher), teleoperates the robot using PHANToM device with 6 degrees-of-freedom force feedback. … The wrist of the robot is constrained to move only along the vertical direction during the whole task, while its orientation is constrained to remain constant.);
implementing the command via an actuator and a robot appendage to create a second interaction that is a pHRH interaction between the system for LfD via pHRH interactions and a second human (robot and operator, Figure 3.2 on page 61; and second paragraph under “Hardware setup” on page 60: “During the demonstrations, a human operator (the teacher), teleoperates the robot using PHANToM device with 6 degrees-of-freedom force feedback. Hence, the teacher had a full feedback of the interaction wrench measured at the gripper of the robot. A second operator (the operator) assists the teleoperated robot to lift a beam, while keeping it horizontal.); and
training a model for the system for LfD via pHRH interactions based on the second interaction that is the pHRH interaction (section 3.2.2 “Experiments” on page 62: “the robot was taught to realize the task with two different dynamics, corresponding to two extreme role distributions among the partners.”; section 3.3 “Learning collaborative tasks” on page 64: “The previous section has introduced a setup and an experimental method to demonstrate a collaborative lifting task to a robotic system. The analysis of the data recorded during the demonstrations shows two distinct patterns in the force-position-velocity space corresponding to the two demonstrated role distributions. In this section, we present the probabilistic framework which we used to encode these patterns, so as to retrieve the skill in autonomous lifting tasks performed with human operators.”; section 3.3.1 “Probabilistic encoding of the task” on page 65: “The data consist in a set of points
D
=
ξ
k
,
k
=
1
.
.
N
d
recorded during the demonstrations of the task (
N
d
is the number of recorded data points). Each point
ξ
is defined as
ξ
=
x
,
f
,
x
˙
T
, where
x
,
x
˙
and
f
are the position and velocity of the robot’s gripper and the vertical force measured at the robot’s wrist at a given time of given demonstration.”; section 3.3.3 “Control scheme” on page 68: “As mentionned earlier in the chapter (see paragraph 3.2.3), the leader and follower data-sets are separated along the force axis. Therefore, if we consider building a controller to reproduce the tasks from these demonstrations, it is natural to consider force as an input to the system. It is also natural to consider the velocity of the gripper as an output of the controller rather than position, so that for a given force at a given position, a reference velocity should be computed.”; section 3.4.1, “reproduction setup” on page 70,: “The setup used for the reproduction of the demonstrated task was the same as for the teaching phase, except that the robot was acting autonomously, instead of being teleoperated.”. In other words, the system of Evrard demonstrates a robot lifting task in which a teacher controls (corresponding to the claimed “first interaction”) a robot cooperating with an operator (corresponding to the claimed “second interaction”). The system records the position, velocity, and force experienced by the robot while interacting with the operator during the demonstration (during the claimed “second interaction”) and uses these data points to create a controller to reproduce the tasks from the demonstrations without the aid of the teacher (corresponding to the claimed “training a model”).).
Re claim 2. Comprising the actuator and the robot appendage (section B.3 “Method” on page 136: “joint actuators”; and Fig. B.1 on page 136: the robot’s arm).
Re claim 3. Comprising a communication interface receiving the operation input associated with the first human indicative of the desired command to be performed by the system for LfD via pHRH interactions (PHANToM device force feedback haptic display, Fig. 3.2, and section 3.2.1 “Hardward setup and controller”, pages 60-61).
Re claim 4. Wherein the first human is located remotely from the system for LfD via pHRH interactions (Fig. 3.2, page 61; and section 4.3, “Taxonomies of coupling” on page 85: “The simplest coupling consists in bilaterally coupling the end-effectors of both robotic systems, so that the interaction on the distant site is reflected on the master site. This way, the operator can feel as if it was directly interacting with the remote environment.”).
Re claim 5. Wherein a remote system for LfD via pHRH interactions generates and implements a remote command corresponding to the command via a remote actuator and a remote robot appendage to implement a movement corresponding to the pHRH interaction between the system for LfD via pHRH interactions and the second human (PHANToM device force feedback haptic display, Fig. 3.2; and section 3.2.1 “Hardward setup and controller”, pages 60-61).
Re claim 7. Wherein one or more of the constraints is a force constraint, an acceleration constraint, a velocity constraint, a proximity constraint, a joint angle constraint associated with the robot appendage, or an interaction constraint (second paragraph under “Hardware setup” on page 60: “The wrist of the robot is constrained to move only along the vertical direction during the whole task, while its orientation is constrained to remain constant.).
Re claim 8. Wherein the desired command associated with the operation input is a wound care command, a physical rehabilitation command, a movement command, or a carrying command (velocity command from the teacher in Figure 3.2 on page 61).
Re claim 9. Comprising a sensor sensing a characteristic associated with the pHRH interaction between the system for LfD via pHRH interactions and the second human (force signal from the robot in Figure 3.2 on page 61).
Re claim 10. Wherein the processor trains the model for the system for LfD via pHRH interactions based on the characteristic (section 3.3 “Learning collaborative tasks” beginning on page 64; first paragraph on page 65: “Probabilistic encoding of the task The data consist in a set of points D = {ξk }, k = 1..Nd recorded during the demonstrations of the task (Nd is the number of recorded data points). Each point ξ is defined as ξ =[x f
x
˙
]T , where x,
x
˙
and f are the position and velocity of the robot’s gripper and the vertical force measured at the robot’s wrist at a given time of given demonstration.”).
Re claim 11. A computer-implemented method for learning from demonstration (LfD) (section 3.1 “Programming by demonstration of collaborative tasks”, page 57) via physical human-robot-human (pHRH) interactions (section 3.2.1 “Hardware setup and controller”, second paragraph under “Hardware setup” on page 60; and Fig. 3.2 on page 61), comprising:
generating a command corresponding to an operation input received from a first human associated with a first interaction that is a physical human-robot (pHR) interaction with a first robot indicative of a desired command to be performed by a robot for LfD via pHRH interactions subject to one or more constraints (velocity command from the teacher in Figure 3.2 on page 61; and second paragraph under “Hardware setup” on page 60: “During the demonstrations, a human operator (the teacher), teleoperates the robot using PHANToM device with 6 degrees-of-freedom force feedback. … The wrist of the robot is constrained to move only along the vertical direction during the whole task, while its orientation is constrained to remain constant.);
implementing the command via an actuator and a robot appendage to create a second interaction that is a pHRH interaction between the robot for LfD via pHRH interactions and a second human (robot and operator, Figure 3.2 on page 61; and second paragraph under “Hardware setup” on page 60: “During the demonstrations, a human operator (the teacher), teleoperates the robot using PHANToM device with 6 degrees-of-freedom force feedback. Hence, the teacher had a full feedback of the interaction wrench measured at the gripper of the robot. A second operator (the operator) assists the teleoperated robot to lift a beam, while keeping it horizontal.); and
training a model for the robot for LfD via pHRH interactions based on the second interaction that is the pHRH interaction (section 3.2.2 “Experiments” on page 62: “the robot was taught to realize the task with two different dynamics, corresponding to two extreme role distributions among the partners.”; section 3.3 “Learning collaborative tasks” on page 64: “The previous section has introduced a setup and an experimental method to demonstrate a collaborative lifting task to a robotic system. The analysis of the data recorded during the demonstrations shows two distinct patterns in the force-position-velocity space corresponding to the two demonstrated role distributions. In this section, we present the probabilistic framework which we used to encode these patterns, so as to retrieve the skill in autonomous lifting tasks performed with human operators.”; section 3.3.1 “Probabilistic encoding of the task” on page 65: “The data consist in a set of points
D
=
ξ
k
,
k
=
1
.
.
N
d
recorded during the demonstrations of the task (
N
d
is the number of recorded data points). Each point
ξ
is defined as
ξ
=
x
,
f
,
x
˙
T
, where
x
,
x
˙
and
f
are the position and velocity of the robot’s gripper and the vertical force measured at the robot’s wrist at a given time of given demonstration.”; section 3.3.3 “Control scheme” on page 68: “As mentionned earlier in the chapter (see paragraph 3.2.3), the leader and follower data-sets are separated along the force axis. Therefore, if we consider building a controller to reproduce the tasks from these demonstrations, it is natural to consider force as an input to the system. It is also natural to consider the velocity of the gripper as an output of the controller rather than position, so that for a given force at a given position, a reference velocity should be computed.”; section 3.4.1, “reproduction setup” on page 70,: “The setup used for the reproduction of the demonstrated task was the same as for the teaching phase, except that the robot was acting autonomously, instead of being teleoperated.”. In other words, the system of Evrard demonstrates a robot lifting task in which a teacher controls (corresponding to the claimed “first interaction”) a robot cooperating with an operator (corresponding to the claimed “second interaction”). The system records the position, velocity, and force experienced by the robot while interacting with the operator during the demonstration (during the claimed “second interaction”) and uses these data points to create a controller to reproduce the tasks from the demonstrations without the aid of the teacher (corresponding to the claimed “training a model”).).
Re claim 12. Comprising receiving the operation input associated with the first human indicative of the desired command to be performed by the robot for LfD via pHRH interactions (velocity command from the teacher in Figure 3.2 on page 61).
Re claim 13. Wherein the first human is located remotely from the robot for LfD via pHRH interactions (Fig. 3.2, page 61; and section 4.3, “Taxonomies of coupling” on page 85: “The simplest coupling consists in bilaterally coupling the end-effectors of both robotic systems, so that the interaction on the distant site is reflected on the master site. This way, the operator can feel as if it was directly interacting with the remote environment.”).
Re claim 14. Wherein a remote system for LfD via pHRH interactions generates and implements a remote command corresponding to the command via a remote actuator and a remote robot appendage to implement a movement corresponding to the pHRH interaction between the robot for LfD via pHRH interactions and the second human (PHANToM device force feedback haptic display, Fig. 3.2; and section 3.2.1 “Hardward setup and controller”, pages 60-61).
Re claim 16. A robot for learning from demonstration (LfD) (section 3.1 “Programming by demonstration of collaborative tasks”, page 57) via physical human-robot- human (pHRH) interactions (section 3.2.1 “Hardware setup and controller”, second paragraph under “Hardware setup” on page 60; and Fig. 3.2 on page 61), comprising:
a memory storing one or more instructions (inherently part of computer from section C.3.6 on page 145); and
a processor executing one or more of the instructions stored on the memory to perform (inherently part of computer from section C.3.6 on page 145):
generating a command corresponding to an operation input received from a first human associated with a first interaction that is a physical human-robot (pHR) interaction with a first robot indicative of a desired command to be performed by the robot for LfD via pHRH interactions subject to one or more constraints (velocity command from the teacher in Figure 3.2 on page 61; and second paragraph under “Hardware setup” on page 60: “During the demonstrations, a human operator (the teacher), teleoperates the robot using PHANToM device with 6 degrees-of-freedom force feedback. … The wrist of the robot is constrained to move only along the vertical direction during the whole task, while its orientation is constrained to remain constant.);
implementing the command via an actuator and a robot appendage to create a second interaction that is a pHRH interaction between the robot for LfD via pHRH interactions and a second human (robot and operator, Figure 3.2 on page 61; and second paragraph under “Hardware setup” on page 60: “During the demonstrations, a human operator (the teacher), teleoperates the robot using PHANToM device with 6 degrees-of-freedom force feedback. Hence, the teacher had a full feedback of the interaction wrench measured at the gripper of the robot. A second operator (the operator) assists the teleoperated robot to lift a beam, while keeping it horizontal.); and
training a model for the robot for LfD via pHRH interactions based on the second interaction that is the pHRH interaction (section 3.2.2 “Experiments” on page 62: “the robot was taught to realize the task with two different dynamics, corresponding to two extreme role distributions among the partners.”; section 3.3 “Learning collaborative tasks” on page 64: “The previous section has introduced a setup and an experimental method to demonstrate a collaborative lifting task to a robotic system. The analysis of the data recorded during the demonstrations shows two distinct patterns in the force-position-velocity space corresponding to the two demonstrated role distributions. In this section, we present the probabilistic framework which we used to encode these patterns, so as to retrieve the skill in autonomous lifting tasks performed with human operators.”; section 3.3.1 “Probabilistic encoding of the task” on page 65: “The data consist in a set of points
D
=
ξ
k
,
k
=
1
.
.
N
d
recorded during the demonstrations of the task (
N
d
is the number of recorded data points). Each point
ξ
is defined as
ξ
=
x
,
f
,
x
˙
T
, where
x
,
x
˙
and
f
are the position and velocity of the robot’s gripper and the vertical force measured at the robot’s wrist at a given time of given demonstration.”; section 3.3.3 “Control scheme” on page 68: “As mentionned earlier in the chapter (see paragraph 3.2.3), the leader and follower data-sets are separated along the force axis. Therefore, if we consider building a controller to reproduce the tasks from these demonstrations, it is natural to consider force as an input to the system. It is also natural to consider the velocity of the gripper as an output of the controller rather than position, so that for a given force at a given position, a reference velocity should be computed.”; section 3.4.1, “reproduction setup” on page 70,: “The setup used for the reproduction of the demonstrated task was the same as for the teaching phase, except that the robot was acting autonomously, instead of being teleoperated.”. In other words, the system of Evrard demonstrates a robot lifting task in which a teacher controls (corresponding to the claimed “first interaction”) a robot cooperating with an operator (corresponding to the claimed “second interaction”). The system records the position, velocity, and force experienced by the robot while interacting with the operator during the demonstration (during the claimed “second interaction”) and uses these data points to create a controller to reproduce the tasks from the demonstrations without the aid of the teacher (corresponding to the claimed “training a model”).).
Re claim 17. Comprising the actuator and the robot appendage (section B.3 “Method” on page 136: “joint actuators”; and Fig. B.1 on page 136: the robot’s arm).
Re claim 18. Comprising a communication interface receiving the operation input associated with the first human indicative of the desired command to be performed by the robot for LfD via pHRH interactions (PHANToM device force feedback haptic display, Fig. 3.2, and section 3.2.1 “Hardward setup and controller”, pages 60-61).
Re claim 19. Wherein the first human is located remotely from the robot for LfD via pHRH interactions (Fig. 3.2, page 61; and section 4.3, “Taxonomies of coupling” on page 85: “The simplest coupling consists in bilaterally coupling the end-effectors of both robotic systems, so that the interaction on the distant site is reflected on the master site. This way, the operator can feel as if it was directly interacting with the remote environment.”).
Re claim 20. Wherein a remote system for LfD via pHRH interactions generates and implements a remote command corresponding to the command via a remote actuator and a remote robot appendage to implement a movement corresponding to the pHRH interaction between the robot for LfD via pHRH interactions and the second human (PHANToM device force feedback haptic display, Fig. 3.2; and section 3.2.1 “Hardward setup and controller”, pages 60-61).
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 6 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Evrard (“Control of humanoid robots to realize haptic tasks in collaboration with a human operator”) as applied to claims 1 and 11 above, and further in view of Liang et al. (“Repairing Deep Neural Networks Based on Behavior Imitation”).
The teachings of Evrard have been discussed above. Evrard fails to specifically teach: (re claim 6) wherein the processor repairs a neural network associated with the model for the system for LfD via pHRH interactions based on the pHRH interaction; and (re claim 15) comprising repairing a neural network associated with the model for the robot for LfD via pHRH interactions based on the pHRH interaction.
Evrard does teach, at page 58, section 3.1.2 “Programming by Demonstration”, “Advances in machine learning allowed to address the generalization problem at two levels: how to extract the important features of a task from a set of demonstrations, and how to generalize to new situations.”
Liang teaches, at the title and abstract, a behavior-imitation based deep neural network repair framework. This can successfully repair buggy deep neural networks with high efficiency.
In view of Liang’s teachings, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to include, with the system as taught by Evrard, (re claim 6) wherein the processor repairs a neural network associated with the model for the system for LfD via pHRH interactions based on the pHRH interaction; and (re claim 15) comprising repairing a neural network associated with the model for the robot for LfD via pHRH interactions based on the pHRH interaction, with a reasonable expectation of success, since Liang teaches a behavior-imitation based deep neural network repair framework. This can successfully repair buggy deep neural networks with high efficiency.
Response to Arguments
Applicant's arguments filed 8/3/2026 have been fully considered but they are not persuasive.
Applicant remarks, on page 8,
Applicant contends that Evrard does not teach or suggest all features of claim 1, as amended. The Office Action cites Section 3.2.2 and 3.3 of Evrard for allegedly disclosing the training a model for the system for LfD via pHRH interactions based on the pHRH interaction of claim 1. Applicant contends that Evrard does not disclose training a model for the system for LfD via pHRH interactions based on the second interaction that is the pHRH interaction, as recited by claim 1. The Office Action alleges that the disclosure of "the robot was taught to realize the task with two different dynamics, corresponding to two extreme role distributions among the partners".
However, it is assumed that "the robot was taught" of Evrard is learning from demonstration (LfD), Evrard fails to disclose the features of amended claim 1 because, at best, Evrard trains a model based on a first interaction, rather than a second interaction created from the first interaction. In Evrard, a human teach performs the demonstration to allegedly train a model. As seen in FIG. 3.2 of Evrard, reproduced below for reference, any alleged training of a model occurs based on the human teacher demonstrating (i.e., the first interaction between the first human and the first robot) how to perform a collaborative task with a human partner. The subject matter of claim 1, on the other hand, requires training a model for a second robot based on a second interaction that is created based on a first interaction. There is no disclosure in Everard that any model is trained based on the interaction between the second human and the second robot (on the left side of FIG. 3.2 of Evrard).
The system of Evrard, at pages 60-70, demonstrates a robot lifting task in which a teacher controls (corresponding to the claimed “first interaction”) a robot cooperating with an operator (corresponding to the claimed “second interaction”). The system records the position, velocity, and force experienced by the robot while interacting with the operator during the demonstration (during the claimed “second interaction”) and uses these data points to create a controller to reproduce the tasks from the demonstrations without the aid of the teacher (corresponding to the claimed “training a model”).
Conclusion
“Phantom Robotic Manipulandum” from Brandeis University is being provided to demonstrate the PHANToM device used to receive the teacher’s input in Evrard is considered a robotic manipulator.
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SPENCER D PATTON whose telephone number is (571)270-5771. The examiner can normally be reached Monday to Friday 9:00-5:00 ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Khoi Tran can be reached at (571)272-6919. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/SPENCER D PATTON/Primary Examiner, Art Unit 3656
Prosecution Insights