Transcript Current State of Powered Gait Devices

Definitions
Powered Locomotion Devices
1. Orthosis:
–
a rigid, non-moving brace for weak or ineffective joints
or muscles
2. Prosthesis:
–
an artificial device to replace a missing part of the body
3. Functional Electrical Stimulation (FES):
–
Surface or surgically implanted current electrodes that
active groups of muscles
4. Gait:
–
Sequence of foot movements, walking
Orthosis
Powered Locomotion Devices
Comparison of hybrid walking systems for paraplegics: analysis of study methodology (Ijzerman
1999)
Moorong Medial Linkage Orthosis (Moorong MLO) (Middleton 1998)
•
Arcuate sliding link centered on the hip joints with roller bearings
Walkabout Orthosis (Middleton 1997)
•
Medially-mounted hinge joint linking two KAFO
Self-Fitting Modular Orthosis (SFMO) (Popovic 1993)
•
Three modules to support patient, FES-aided gait
•
Project has been abandoned
Weight Bearing control (WBC) orthosis (Yano 1997)
•
Consists of: rigid frame for support, reciprocating hip joint; gas powered foot device that
varies sole thickness and push button sequential control system
FES and Hybrid Walking Systems
Powered Locomotion Devices
Advantages/Disadvantages
•
•
•
Effects of SCI, muscles used, training requirements, cost, spasticity reduction, reliability (Solomonow 1992)
Builds muscle mass and stroke volume (Merati et al 2000)
High energy expenditure
–
•
Oxygen demand is above 50 % of the VO2 peak (Merati et al 2000)
Slow ambulation
–
•
~tenfold less than wheelchair (Merati et al 2000)
Cosmesis and difficult to don/doff
–
14 subjects, 3 using RGO, 4 using only FNS (Merati et al 2000)
Parastep System, Sigmedics Corp. (Frank Zeiss)
•
Only commercially available FES system
Cleveland FES Center and Case Western Reserve U (CWRU) (feswww.fes.cwru.edu)
•
16-channel FES with implanted electrodes and a walker. Surface are impractical for everyday use (Kobetic 1999)
–
•
Using a switch initiated gait, paraplegic could stand for 8 minutes and walk for 20 meters
Isocentric reciprocal gait orthosis (ISO-RGO) vs. FNS or orthosis only (Marsolais 2000)
–
–
Slower walking (.2 m/s) and increased energy cost (.5 Kcal/m)
Better stability and walking distance
Self-Fitting Modular Orthosis (SFMO) (Popovic 1993)
•
•
•
•
Separate knee, hip and ankle modules placed on a pair of jeans
Maximal average speed = 0.4 m/s where upper body support was 60 % of total
Approximately same oxygen uptake as RGO+FES
Has abandoned project
Controlled Brake Orthosis (Goldfarb and Durfee 1994)
•
•
Uses magnetic particle brakes to dissipate excessive torque
Dr. Durfee (U. of Minnesota) is looking for second round of funding
Advanced Gait Orthosis
Powered Locomotion Devices
Hydraulic system (Seireg et al 1981)
•
•
•
Five degrees of freedom (2 at hip, 1 at knee, 2 at ankle)
Good and simple design and analysis
Bulky and unusable because of the current state of technology
Powered Gait Orthosis (PGO) – 4 bar linkage and CAM system (Ruthenberg et al 1997)
•
•
•
•
•
One degree of freedom run by a linear DC motor
More of a research tool than a a practical means for paraplegic gait
Battery pack and control system can be fastened to the back of the corset
Mechanized hip and knee with cam-modulated linkage for knee function
Peak power usage is the same as human walking
History of active exoskeletons (Vukobratovic 1990)
•
•
•
•
Hydraulic powered exoskeleton (1968)
Two pneumatically driven exoskeletons (1970-1973)
Electrical exoskeleton using servoelectric D.C. drives (1974)
Compact, computer controlled active suit for “dysthrophics” (1980)
Dynamic Knee Brace System (DKBS) (Irby 1999)
•
•
Allows flexion during swing, restricts it during stance
Footswitches are inputs, finite state controlled linear solenoid to control knee
Intelligent Orthosis (IO) (Suga et al 1998)
•
•
A KAFO that controls the amount of resistance to joint movement using rotary encoder and heel switch
All subjects had a weakened quadriceps
Current Useful Technology
Powered Locomotion Devices
1. Biped robots
•
MIT leg lab “spring flamingo” (Pratt 1999)
–
–
•
Have a 6 degree of freedom biped robot walking on unknown sloped terrain
Applied a neural network mechanism for a stable adaptive control
Anthropomorphic biped robot BIP2000 (Espiau 2000)
–
–
Bipedal robot with 15 active joints include hip
Can walk and turn in unknown sloped terrain
2. Feedback controllers
–
–
Sensory nerve signal to predict EMG signal for FES (Strange 1999)
EMG control of actuators (Fukuda 1998)
3. Shape memory alloy (SMA) actuated arm and hand prosthesis (Mavroidis 1999)
–
–
–
Early stages of development, hasn’t been applied to locomotion
Advantages: small size, high force to weight ratio, low cost
Disadvantages: low strain, limited life cycle, non-linear effects, low bandwidth and efficiency
4. DARPA exoskeleton project (http://www.darpa.mil/dso/thrust/md/exoskeletons/program.html)
–
This program will be used to develop technologies, such as actively controlled exoskeletons, to
enable a soldier to handle more fire-power, wear more ballistic protection, and carry more
ammunition and supplies
Powered Actuators
Powered Locomotion Devices
1. Shape Memory Alloys (SMA) Actuators
–
–
–
Frequency of actuation: 5 Hz w/o cooling; 15-20 Hz w/ liquid coolant
Low efficiency and cyclic abilities due to heat transfer
Not a feasible option for actuation
2. Pneumatic Muscle Actuators (PMA) (Caldwell 1998) & McKibben Artificial Muscle (Tondu
2000)
–
–
–
–
–
Highly flexible, soft actuators
Strains of 30 %
Max. bandwidth for antagonistic pairs is 5 Hz
Active stress is 3 MPa
Has built orthotic devices out of and controllers for MIMO
3. Series Elastic Actuators (Pratt 1995)
–
–
Advantages: greater shock tolerance, lower reflected inertia, more accurate and stable force
control, less inadvertent damage to environment and capacity for energy storage
Disadvantages: low zero motion force bandwidth
4. Electroactive Polymers (Dr. John Madden, MIT Newman Lab)
–
–
–
–
Safe stress is 3 MPa, Specific power of 39 W/kg
Can survive 100,000 cycles at 2 % (work being done on fatigue characteristics)
3 % Efficiency (could be as high as 20% if can recover stored electrical energy)
JPL's NDEAA Technologies Group (ndeaa.jpl.nasa.gov/nasa-nde/lommas/eap/EAP-web.htm)
Ankle
Powered Locomotion Devices
The Odstock Dropped Foot Stimulator (http://www.mpbe-sdh.demon.co.uk/fes.htm)
•
•
•
single channel FES that corrects dropped foot by stimulating the common peroneal nerve using
self adhesive skin surface electrodes placed on the side of the leg
Clinical study finished in 1995, 178 patients have been treated 25 have stopped
Cost is for 1 year is £830
Peroneal Nerve Stimulator (PNS) (Voigt 2000)
•
•
Creates an exaggerated dorsiflexion with excessive subtalar eversion
Could not find any excessive and potential harmful mechanical loads
Spring-Type AFO (Brunner 1998)
•
•
Stiff AFO with a 10-15o cut at the ankle
Showed that allowing the ankle joint to move facilitates normal gait
Dorsiflexion Assist Controlled by Spring AFO (DACS-AFO) (Hachisuka 1998)
•
•
•
Generates a dorsiflexion assist moment during plantar flexion and no moment during
dorsiflexion using a spring located at the calf
The initial dorsiflexion angle of the ankle joint is adjustable and three springs with
different moments are available.
None of the five subjects that they tested said that they preferred the DACS-AFO
Spring-assisted dorsiflexion AFO (Lehmann 1986)
•
Effective in ground clearance, but not stiff enough for lengthening contraction after heel-strike
Rehabilitative Devices
Powered Locomotion Devices
Walking Assistance and Rehabilitation Device (WARD) (Gazzani 1999)
•
Treadmill with Body Weight Unloading apparatus
•
Six of seven patients improved score on ambulation scale
Modified crank and rocker mechanized gait trainer (Hesse 2000)
•
Simulates gait, supports subjects and controls their center of mass
•
Two subjects improved dramatically
Lokomat robotic gait Orthosis (http://www.aut.ee.ethz.ch/~jezernik/research.msql)
•
•
Supported by a harness, robotic orthosis driven by DC electric motors moves the patient’s legs
Currently performing gait-pattern adaptation experiments
Control Systems
Powered Locomotion Devices
Bipedal walking robot using Cerebellar Model Articulation Controller (CMAC) (Hu 1999)
•
Adaptive CMAC neural network control is stable and accounts for disturbances
Modified crank and rocker mechanized gait trainer (Hesse 2000)
•
Simulates gait, supports subjects and controls their center of mass
•
Two subjects improved dramatically
Comparison of machine learning (ML) techniques (Jonic 1999)
•
Adaptive network based fuzzy inference system (ANFIS)
•
Minimal number of and most comprehensible rules
•
Entropy minimizing inductive learning (IL) and radial basis function (RBF) neural network
•
Best generalization
Finite state control of FES (Sweeney 2000)
•
Systems receive feedback from sensors on body or from the body’s own natural sensors
Neural network controlled FES maintains high accuracy with two force sensors under foot (Tong 1999)
Fuzzy Walking Pattern (FWP) controller for SMA biped robot (Tu 1998)
Physiology
Powered Locomotion Devices
Paraplegic Standing (Matjacic 1998)
•
Paraplegic standing with ankle stiffness of 8 Nm/o
•
Models the torque around the ankle for a two-link inverted pendulum
•
Did not look at lateral stability
Feedback control of unsupported paraplegic standing (Hunt 1999)
•
inputs are ankle torques and body inclination and outputs a FES signal to the plantar flexors
Passive stiffness increases in paretic patients (Lamontagne 2000)
•
Use stiffness as an energy storage for toe-off
•
Proves that orthotic should help
•
Ankle moment is linear with angle
Biomechanics of the Foot (Mann 1997)
Lower Limb Orthoses (Michael 1997)
Normal and Pathological Gait (Perry 1997)
Improved Muscle-reflex actuator for large-scale neuromuscular models (Winters 1995)
Model of the intrinsic and reflex contributions to ankle stiffness dynamics (Kearney)
Lower joint powers during stair climbing at different slopes (Riener et al 1999)
Power Sources
Powered Locomotion Devices
Fuel Cells
•
Methanol-powered alkaline fuel cell used to power piezoceramic actuator (Leo 1999)
Sensors
Powered Locomotion Devices
Overview (Veltink 1999)
•
Describes body-mounted sensors for muscle activation, force and movement
Using the natural sensors of subject as feedback signals to control FES (Haugland 1999)
•
Replaced heel switch with implanted electrode for a peroneal stimulator
•
Provided a hand grasp FES system with sensory feedback from fingertips
•