IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
US-0820571
(2010-06-22)
|
등록번호 |
US-8494676
(2013-07-23)
|
발명자
/ 주소 |
- Ayers, Joseph
- Witting, Jan
- Ryder, Stephane
- Olcott, Christopher
|
출원인 / 주소 |
|
대리인 / 주소 |
Wilmer Cutler Pickering Hale & Dorr LLP.
|
인용정보 |
피인용 횟수 :
1 인용 특허 :
6 |
초록
▼
A robotic architecture for capturing the autonomous performance advantages the animal models enjoy in the natural environment is disclosed. A biomimesis process is employed to allow selective utilization of basic physical components and adaptation of a common control paradigm for each of different v
A robotic architecture for capturing the autonomous performance advantages the animal models enjoy in the natural environment is disclosed. A biomimesis process is employed to allow selective utilization of basic physical components and adaptation of a common control paradigm for each of different vehicle types. The biomimetic architecture involves five functional elements: a basic biomorphic plant for capturing the biomechanical advantages of the model organism; a neural circuit-based controller consisting of a finite state machine; myomorphic actuators producing linear graded force in response to trains of current pulses for mediating movements; labeled line code output by neuromorphic sensors; and a reactive behavioral sequencer executing command sequences defined within a behavioral library.
대표청구항
▼
1. A method of realizing biomimetic behavior in a robot, comprising: identifying plural physical components of a subject animal, each capable of being physically displaced with respect to each other;identifying a physical maneuver performed by at least one of the physical components of the subject a
1. A method of realizing biomimetic behavior in a robot, comprising: identifying plural physical components of a subject animal, each capable of being physically displaced with respect to each other;identifying a physical maneuver performed by at least one of the physical components of the subject animal;for each of the at least one physical components performing the identified physical maneuver, identifying changes in physical orientation over time;forming a behavioral sequence file comprised of the changes in physical orientation for each of the at least one physical components required to perform the identified physical maneuver over the time required to perform the identified physical maneuver;providing a robot comprising a body with an interior body cavity and an exterior body surface having a substantially duplicate component for each of the identified plural physical components of the subject animal, each of the substantially duplicate components being independently actuatable through selective activation of respective paired, antagonistic actuators;providing a distributed neural circuit-based controller in association with the robot, the controller including a microcontroller and a memory and further comprising a finite state machine for each of the substantially duplicate components, each finite state machine comprising a neural oscillator for dividing each of plural step cycles into elemental phases during which one or more of the respective actuators are activated,a coordinating neuron for checking the step cycle phase of the state machine of one or more adjacent substantially duplicate components to temporally coordinate the activation of each,a pattern generator for coordinating the pattern of movement of the respective substantially duplicate component in response to a command received from the microcontroller, the neural oscillator, and the coordinating neuron, anda recruiter associated with each of the respective actuators for selective activation thereof in response to output from the respective pattern generator;forming a behavioral library in the controller memory, the behavioral library comprising, for a respective physical maneuver, a command sequence, each command of the command sequence for selectively activating one or more of the actuators associated with a respective state machine to achieve a change in physical orientation in a respective substantially duplicate component, the command sequence causing changes in physical orientation for each of the at least one substantially duplicate components required to perform the identified physical maneuver over the time required to perform the identified physical maneuver in accordance the behavioral sequence file; andcommanding the robot to perform the identified physical maneuver through the microcontroller issuing the respective command sequence from the behavioral library to achieve the changes in physical orientation for each of the at least one substantially duplicate components required to, perform the identified physical maneuver over the time required to perform the identified physical maneuver. 2. The method of claim 1, further comprising the steps of: identifying plural physical maneuvers to be performed by the robot;forming a behavioral command sequence data file in the controller memory comprised of a command sequence for each of the identified plural physical maneuvers; andcommanding the robot to perform the identified plural physical maneuvers through the microcontroller issuing the respective command sequences from the behavioral library to achieve the changes in physical orientation for each of the at least one substantially duplicate components required to perform the identified plural physical maneuvers. 3. The method of claim 2, further comprising the steps of: providing at least one sensor in association with the robot, the sensor providing an output in communication with the controller; andselectively altering the step of commanding in response to the sensor output. 4. The method of claim 3, further comprising: identifying step cycles for each of the plural physical maneuvers;identifying a parameter space through which the at least one physical component moves;analyzing the motion of the substantially duplicate component through the parameter space and matching at least one parameter for controlling the substantially duplicate component with mechanical responses of the substantially duplicate component;identifying from the behavioral command sequence data file an expected phase for the at least one substantially duplicate component performing the identified maneuver for each phase of the step cycle;checking the phase of the at least one substantially duplicate component with the oscillator circuit during a step cycle of the identified maneuver;comparing the checked phase of the at least one substantially duplicate component with its expected phase according to the behavioral command sequence data file; andproviding the result of the comparison to the controller. 5. The method of claim 4, further comprising the step of selectively altering the step cycle of at least one of the plural physical-maneuvers by at least one of phase advances and delays in response to the sensor output or the comparison of the checked phase with the expected phase. 6. The method of claim 3, further comprising: wherein the at least one sensor is for sensing at least one of the ambient temperature, collision with an object external to the robot, and velocity relative to one or more geographic coordinates or the external chemical medium surrounding the exterior surface of the robot. 7. The method of claim 3, wherein the at least one sensor is selected from the group consisting of at least one antenna having strain gauges operatively associated therewith and at least one bump sensor. 8. The method of claim 7, further comprising: electrically coupling the strain gauges of the at least one antenna or the at least one bump sensor to the controller; anddetecting a collision between the antenna and an external object using the output of the strain gauges, or between the bump sensor and an external object using the output of the bump sensor, and providing an indication thereof to the controller. 9. The method of claim 3, wherein the step of providing at least one sensor further comprises: providing the robot with at least one optical flow chip and electro-optical image processing capability in the controller to realize autonomous optomotor reflexes. 10. The method of claim 2, further comprising: providing a plurality of sensors;providing sensor fusion signals, wherein combined signals from at least two of the plurality of sensors are communicated to the controller;developing a set of investigatory and reporting behaviors as part of the behavioral command sequence data file, wherein said investigatory and reporting behaviors are selectively modulated in response to the sensor fusion signals whereby objects or conditions external to the robot may be sensed;commanding the robot to investigate and report on objects or conditions external to the robot according to the set of investigatory and reporting behaviors; andidentifying and responding to sensed objects or sensed conditions. 11. The method of claim 2, further comprising: forming conflicting state files in the controller memory, the conflicting state files defining how conflicting controller commands are to be resolved when commanding the robot to perform plural identified physical maneuvers, prior to actuation of the substantially duplicate components required to perform the identified plural physical maneuvers; andcommanding the robot to perform the identified plural physical maneuvers through the microcontroller issuing the respective command sequences from the behavioral library with reference to the conflicting state files implicated by the performance of the respective, identified plural physical maneuvers. 12. The method of claim 1, further comprising: identifying at least one physical maneuver or sequence of maneuvers, or modification to an on-going physical maneuver, that comprises an exteroceptive reflex of the subject animal in response to at least one condition or parameter external to the subject animal;modifying the behavioral sequence file to include a response component associated with the exteroceptive reflex, the response component comprising the changes in physical orientation for each of the at least one physical components performing the identified at least one physical maneuver or sequence of maneuvers, or modification to an on-going physical maneuver, over the time required to perform the exteroceptive reflex;providing at least one sensor in association with the robot for sensing the at least one external condition or parameter and for providing the controller with an indication thereof;modifying the behavioral library, for a respective at least one external condition or parameter as sensed by the at least one sensor, to include a reflex command sequence, each command of the reflex command sequence for actuating a substantially duplicate component to achieve a change in physical orientation therein, the reflex command sequence causing changes in physical orientation for each of the at least one substantially duplicate components required to perform the identified at least one physical maneuver or sequence of maneuvers, or modification to an on-going physical maneuver, over the time required to perform the identified at least one maneuver or sequence of maneuvers, or modification to an on-going physical maneuver, that comprises the exteroceptive reflex in accordance with the behavioral sequence file; andcommanding the robot to perform autonomously in response to the at least one sensor sensing the at least one external condition or parameter through the microcontroller issuing the respective reflex command sequence from the behavioral library to achieve the changes in physical orientation for each of the at least one substantially duplicate components required to perform the identified at least one physical maneuver or sequence of maneuvers, or modification to an on-going physical maneuver, associated with the exteroceptive reflex. 13. The method of claim 12, wherein the exteroceptive reflex is a rheotaxic or orientational response. 14. The method of claim 1, further comprising: providing the robot with a battery compartment and a plurality of batteries therein; anddividing the plurality of batteries into parallel sets of power circuits, each set of power circuits comprising plural ones of the plurality of batteries, the plural ones being connected in series. 15. The method of claim 1, further comprising: providing the controller with an acoustic modem receiver; andsupervising the robot from a remote location via an acoustic modem transmitter. 16. The method of claim 1, further comprising: programming the controller to move the robot with a propensity to investigate a target object, wherein the propensity to investigate is dependent upon the proximity to the target object. 17. The method of claim 1, further comprising: providing the robot with a sonar transceiver in communication with the controller; andmodulating the step of commanding based upon data from the sonar transceiver. 18. The method of claim 1, wherein the step of providing a distributed neural circuit-based controller comprises providing a controller composed of synaptically connected electronic neurons configured as a neural network. 19. The method of claim 1, wherein the step of commanding the robot further comprises: issuing the respective command sequence including the steps of turning actuators associated with the at least one substantially duplicate component on or off according to a respective pulse-width duty cycle. 20. The method of claim 19, wherein the step of commanding the robot further comprises: providing the robot with an environmental sensor in communication with the controller, the environmental sensor for providing an environmental characteristic; andmodulating the respective pulse-width duty cycle according to the sensed environmental characteristic. 21. The method of claim 1, further comprising: providing the controller with a serial bus interface for selectively communicating with a peripheral interface controller. 22. The method of claim 1, further comprising: providing program code to the controller via a removable memory module or a PCMCIA adapter associated therewith. 23. The method of claim 1, further comprising: incorporating into the robot a low-power acoustic modem and imbedded signal-processing platform to allow integrated control and navigation. 24. A method of controlling a biomimetic robotic system for autonomous operation, the robotic system having a biomorphic physical plant that is based on and structured and arranged to substantially duplicate an invertebrate animal model having a neuronal network, the physical plant including a plurality of sensing devices including sensors, actuatable body components, a memory, a plurality of motoring mechanisms each associated with a respective one of the actuatable body components, a state machine for each of the actuatable body components, a behavioral sequencer for ordering a plurality of behavioral responses in the actuatable body components in a desired temporal order, and a neural circuit-based, behavior-based controller that is modeled after the neuronal network of the animal model, for mediating stable locomotion, reactive navigation, and sensing of the robotic system and the environment surrounding it, the method comprising: polling data from the plurality of sensing devices;comparing the polled data with predefined releasers, stored in behavioral libraries stored in the memory, to identify a unique behavioral releaser corresponding to said polled data;identifying a sequence of command state changes for at least one of the state machines to mediate an evoked behavior corresponding to the unique behavioral releaser;placing the sequence of identified command state changes into a temporal location within a command stack of executable sequences, to cause said sequence of command state changes to be executed upon reaching the temporal-based location in said command stack; andexecuting the sequence of command state changes sequentially in accordance with the temporal-based order of the command stack. 25. The method as recited in claim 24, wherein polling data from the plurality of sensing devices includes producing labeled-line code based on sensing device modality. 26. The method as recited in claim 24, wherein preparing to execute the sequence of command state changes includes identifying and suppressing or modifying any conflicting state changes already present in the command stack. 27. The method as recited in claim 24, wherein placing the sequence of command state changes includes varying an intensity of a respective command state change by using different temporal compressions therefor. 28. The method as recited in claim 27, wherein using different temporal compressions includes using pulse width modulation of a respective duty cycle. 29. The method as recited in claim 24, wherein each of the actuatable body components includes at least one elevator, at least one protractor, at least one depressor, and at least one retractor, each being selectively actuatable to enable forward and rearward movement, and at least one extensor and at least one flexor, each being selectively actuatable to enable lateral movement, the method further comprising controlling at least one of the state machines to actuate a corresponding one of said at least one elevator, said at least one protractor, said at least one depressor, said at least one retractor, said at least one extensor or said at least one flexor. 30. The method as recited in claim 24, further comprising: checking a phase of adjacent ones of said actuatable body components prior to execution of a sequence of command state changes on a respective actuatable body component; andresetting the phase of said respective actuatable body components if the phase of any of the adjacent actuatable body components would conflict with the execution of the sequence of command state changes on the respective actuatable body component. 31. The method as recited in claim 24, wherein placing the sequence of command state changes includes performing multi-component sequencing to perform at least one of a rheotaxic behavioral response and an exteroceptive behavioral response. 32. The method as recited in claim 24, wherein the step of comparing further comprises comparing the polled data with predefined releasers, stored in the behavioral libraries stored in the memory, to identify an exteroceptive reflex, and further comprising the step of modulating at least one state machine without altering the command stack of executable sequences based upon identification of an exteroceptive reflex.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.