Fuzzy logic based control

Details Fuzzy logic based control Fuzzy logic based control

2000-12-01

2003-10-14

Patel, Ramesh (Department: 2121)

Data processing: generic control systems or specific application

Generic control system, apparatus or process

Optimization or adaptive control

C700S048000, C700S028000, C700S051000, C700S246000, C700S247000, C700S250000, C706S001000, C706S009000, C706S052000, C706S900000, C701S027000, C701S040000, C701S044000, C701S057000

Reexamination Certificate

active

06633783

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to a controller for a robot or a human assist device, more specifically, a fuzzy logic controller for controlling the motion of a robot or a human assist device in accordance with human postural dynamics.
2. Description of the Prior Art
The mechanism behind human postural balance is a challenging problem to engineers, biologists, physiologists, medical specialists, and mathematicians, which is discussed in K. Barin, “Evaluation of a generalized model of human postural dynamics and control in the sagittal plane” Biological Cybernetics, 61:37-50, 1989; H. Hemami, F. C. Weimer, C. Robinson, C. Stockwell, and V. S. Cvetkovic, “Biped stability considerations with vestibular models” IEEE Transactions on Automatic Control, 23:1074-1079, 1978; and L. M. Nanshner and G. McCollum “The organization of human postural movements: A formal basis and experimental synthesis” Behav. Brain Sci., 8:135-172, 1985.
An important practical reason for studying postural control is the application of the results in restoring the locomotion function of disabled persons and for developing external assistive devices to enable people to perform tasks with less effort, as discussed by B. Dariush, M. Parnianpour, and H. Hemami “Stability and a control strategy of a multi-link musculoskeletal model with applications in FES” IEEE Transactions on Biomedical Engineering, 45:3-14, 1998.
The most obvious task performed by the human postural control system is the maintenance of the upright bipedal stance. This task is essentially one of generating a series of muscular contractions that produce moments of force about the joints of the musculoskeletal system in order to counteract the effects of gravity. These muscular contractions operate virtually continuously, and it is now well known that in the course of evolution these ‘anti-gravity’ or ‘postural’ muscles underwent physiological adaptations to enable them to perform efficiently.
U.S. Pat. No. 5,432,417 entitled “Locomotion Control System for Legged Mobile Robot”, assigned to the same assignee of this invention relates to a locomotion control system for a biped walking robot. A walking pattern is pre-established in advance in terms of ZMP (Zero Moment Point) at which point a horizontal moment acting on the robot that is generated by the ground reaction force is zero, a trajectory of the body's attitude. When walking, an actual ground reaction force is detected to determine an actual ZMP position, which is compared with a target ZMP position determined by the ZMP trajectory. The system controls movement of the legs, thighs, and trunk of the robot such that actual ZMP reaches the target ZMP.
Heretofore, human postural dynamics was too complex to be performed by a robot or an artificial device such as human assist systems including artificial legs. There is a need for a system that enables a robot or an artificial human assist device to act in accordance with human postural dynamics.
SUMMARY OF THE INVENTION
In accordance with the present invention, a fuzzy logic controller controls the motion of a robot or a human assist device in accordance with human postural dynamics. The fuzzy logic controller is advantageous when the muscular reactions are required, which are too complex for conventional quantitative techniques. The fuzzy controller parallels the human ability of the use of reason to control processes with imprecise, incomplete, or unreliable information.
In accordance with one aspect of the invention, a fuzzy control system is provided for controlling a posture of an object including a robot and a human assist device that has at least a part that is moved by a pair of actuators provided at one end thereof. The system comprises a system dynamics unit for detecting regulation error of said part, the regulation error being an error of the position of said part relative to a desired position; and a fuzzy controller, responsive to the regulation error and regulation error rate, for performing fuzzy inference process to provide outputs for driving said pair of actuators, whereby the pair of actuators provide antagonist/agonist forces to said part for bringing said object into a regulated posture,
In accordance with a specific aspect of the invention, the system dynamics unit includes a controller for controlling the posture of the object, said part being an artificial leg and said pair of actuators acting as muscles for moving the leg. In one embodiment, the object is a biped walking robot.
In accordance with another aspect of the invention, the fuzzy controller comprises fuzzification interface; a knowledge base including a set of control rules on how to control the system; an inference mechanism for evaluating which control rules out of said set of control rules are relevant at the current time and providing outputs; and defuzzification interface for converting the outputs of the inference mechanism into the input to the system dynamics.
In accordance with further aspect of the invention, the fuzzy controller has two inputs corresponding to scaled regulation error and regulation error rate respectively, and has outputs that are to be scaled to produce neural firing rate inputs that generate the muscle forces for said pair of actuators. The knowledge base includes a rule table of conditional statements, the scaled regulation error and the regulation error rate being if-parts of the conditional statements and the outputs being then-parts of the conditional statements. The defuzzification interface, responsive to the outputs from the inference mechanism, produces a single crisp number using centroid defuzzification method.
In accordance with another aspect of the invention, a fuzzy control system for controlling a posture of a biped walking robot having a leg, a thigh and a trunk is provided. The leg, thigh and trunk respectively is moved by at least a pair of actuators provided at one end. The system comprises a musculoskeletal system unit for detecting regulation error of the leg, the regulation error being an error of the position of the leg relative to a desired position; and a leg segment fuzzy controller, responsive to the regulation error and regulation error rate, for performing fuzzy inference process to provide outputs for driving said pair of actuators, whereby the pair of actuators provide antagonist/agonist forces to the leg for bringing the robot into a regulated posture.
In one embodiment, the musculeskeletal system unit detects regulation error of the thigh, the regulation error being an error of the position of the thigh relative to a desired position, the system further comprising a thigh segment fuzzy controller, responsive to the regulation error and the regulation error rate of the thigh, for performing fuzzy inference process to provide outputs for driving the thigh.
In another embodiment, the musculeskeletal system unit detects regulation error of the trunk, the regulation error being an error of the position of the trunk relative to a desired position, the system further comprising a trunk segment fuzzy controller, responsive to the regulation error and the regulation error rate of the thigh, for performing fuzzy inference process to provide outputs for driving the trunk.


REFERENCES:
patent: 4990838 (1991-02-01), Kawato et al.
patent: 5231335 (1993-07-01), Mega et al.
patent: 5241651 (1993-08-01), Ueda
patent: 5247432 (1993-09-01), Ueda
patent: 5432417 (1995-07-01), Takenaka et al.
patent: 5459659 (1995-10-01), Takenaka
patent: 5701965 (1997-12-01), Kamen et al.
patent: 6445983 (2002-09-01), Dickson et al.
patent: A-4-256130 (1992-09-01), None
patent: A-5-27981 (1993-02-01), None
patent: A-5-150991 (1993-06-01), None
patent: A-5-298422 (1993-11-01), None
patent: A-5-337849 (1993-12-01), None
patent: A-6-44077 (1994-02-01), None
patent: A-6-242956 (1994-09-01), None
patent: A-6-251122 (1994-09-01), None

Affiliated with

Also associated with

Rating

Fuzzy logic based control does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Fuzzy logic based control, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Fuzzy logic based control will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3118171