Method and system for autonomous body interaction
A system, method, and apparatus for autonomous body interaction system is provided. A system and method for a robot interacting with a human or other soft or deformable body is provided for the purpose of active application of pressure to specific points or regions of the body. The system and method conducts at least one of: localizing the position of the body, detecting the configuration of the body, identifying the surface region(s) of the body, predicting the underlying anatomy of the body, assessing the state of the body, preparing a plan for manipulation of the body, and executing a predetermined and/or dynamically updated plan on the body.
1. A system, comprising:
a robotic manipulator including an end effector; and
a controller operably coupled to the robotic manipulator and configured to control an interaction between the end effector and a deformable body, the controller configured to:
determine, based on multi-dimensional data detected by one or more sensors, at least one of a position, an orientation, or a shape of the deformable body;
generate a body model for the deformable body based on the at least one of the position, the orientation, or the shape of the deformable body;
generate an interaction plan based on the body model, wherein the interaction plan is generated based at least in part on a plurality of stroke segments, wherein a stroke segment comprises a sequence of interaction goals, and wherein generating the interaction plan further comprises facilitating a continuous trajectory for execution at least in part by inserting one or more additional goals between stroke segments; and
cause the robotic manipulator to execute the interaction plan, including generating commands based on the sequence of interaction goals.
2. The system of claim 1 , wherein the controller is further configured to:
perform tracking of a movement of the deformable body during execution of the interaction plan; and
modify the interaction plan based on the tracking, to at least one of maintain contact with the deformable body, or position the end effector relative to a predefined part of the deformable body.
3. The system of claim 1 , wherein the interaction plan specifies at least one overlay, the at least one overlay including at least one of a variation of a position goal or a variation of a force goal.
4. The system of claim 1 , wherein the controller is further configured to assign labels to the multi-dimensional data using a deep learning model.
5. The system of claim 1 , wherein the interaction plan is configured to cause a contact between the end effector and the deformable body.
6. The system of claim 5 , wherein the deformable body is a human body and the contact between the end effector and the deformable body includes contact sufficient to at least one of displace, compress, or mobilize muscle tissue within the human body.
7. The system of claim 1 , wherein the body model is a current body model, the generating the current body model including modifying a historical body model based on the multi-dimensional data.
8. The system of claim 1 , wherein the interaction plan is a current interaction plan, and the generating the current interaction plan includes adjusting a historical interaction plan based on the multi-dimensional data.
9. The system of claim 1 , wherein the multi-dimensional data is a first multi-dimensional data, and wherein the controller is further configured to:
determine second multi-dimensional data based on signals received from the one or more sensors during the execution of the interaction plan; and
modify the interaction plan during the execution of the interaction plan based on the second multi-dimensional data.
10. The system of claim 9 , wherein modifying the interaction plan includes at least one of modifying a pressure of the interaction plan or modifying a frequency of touch of the interaction plan.
11. A method, comprising:
based on multi-dimensional data detected by one or more sensors, determining, via a controller, at least one of a position, an orientation, or a shape of a deformable body, wherein the controller is operably coupled to a robotic manipulator including an end effector and configured to control an interaction between the end effector and the deformable body;
generating, via the controller, a body model for the deformable body based on the at least one of the position, the orientation, or the shape of the deformable body;
generating, via the controller, an interaction plan based on the body model, wherein the interaction plan is generated based at least in part on a plurality of stroke segments, wherein a stroke segment comprises a sequence of interaction goals, and wherein generating the interaction plan further comprises facilitating a continuous trajectory for execution at least in part by inserting one or more additional goals between stroke segments; and
causing the robotic manipulator to execute the interaction plan, including generating commands based on the sequence of interaction goals.
12. The method of claim 11 , further comprising:
performing, via the controller, tracking of a movement of the deformable body during execution of the interaction plan; and
modifying, via the controller, the interaction plan based on the tracking, to at least one of maintain contact with the deformable body, or position the end effector relative to a predefined part of the deformable body.
13. The method of claim 11 , wherein the interaction plan specifies at least one overlay, the at least one overlay including at least one of a variation of a position goal or a variation of a force goal.
14. The method of claim 11 , further comprising assigning labels to the multi-dimensional data using a deep learning model.
15. The method of claim 11 , wherein the interaction plan is configured to cause a contact between the end effector and the deformable body.
16. The method of claim 15 , wherein the deformable body is a human body and the contact between the end effector and the deformable body includes contact sufficient to at least one of displace, compress, or mobilize muscle tissue within the human body.
17. The method of claim 11 , wherein the body model is a current body model, the generating the current body model including modifying a historical body model based on the multi-dimensional data.
18. The method of claim 11 , wherein the interaction plan is a current interaction plan, and the generating the current interaction plan includes adjusting a historical interaction plan based on the multi-dimensional data.
19. The method of claim 11 , wherein the multi-dimensional data is a first multi-dimensional data, and further comprising:
determining, via the controller, second multi-dimensional data based on signals received from the one or more sensors during the execution of the interaction plan; and
modifying, via the controller, the interaction plan during the execution of the interaction plan based on the second multi-dimensional data.
20. The method of claim 19 , wherein modifying the interaction plan includes at least one of modifying a pressure of the interaction plan or modifying a frequency of touch of the interaction plan.