Posts Tagged ‘John M. Holland’

1983 – “Kludge” Omnidirectional Mobile Robot – John M. Holland (American)

kludge-rajan85-1 001-x640 (24)

1983 – "Kludge" Omnidirectional Mobile Robot by John M. Holland.

kludge-rajan85-1 001-x640 (22)

Kludge with legs contracted.


Kludge at a 1984 exhibition.


kludge-rajan85-1 001-x640 (23)

John M. Holland.

kludge-rajan85-1 001-x640 (4)

kludge-rajan85-1 001-x640 (5)

kludge-rajan85-1 001-x640 (6)

kludge-rajan85-1 001-x640 (3)

kludge-rajan85-1 001-x640 (7)

kludge-rajan85-1 001-x640 (8)

The focus in this post is on the unique mobility base, and not on its navigation and sensor qualities.

Patent Information:

kludge-rajan85-1 001-x640 (9)

Publication number    US4573548 A

Publication date    4 Mar 1986
Filing date    23 Jul 1983
Priority date    23 Jul 1983
Inventors    John M. Holland
Original Assignee    Cybermation, Inc.
Mobile base for robots and the like
US 4573548 A
A mobile base for robots or other devices requiring a transport mechanism is disclosed incorporating a plurality of wheels which are simultaneously driven and steered by separate drive sources so as to allow the mobile base to change direction without rotation of the mobile base. In an additional embodiment, each wheel is located on an extensible leg assembly which can be rotated to project outwardly from the mobile base and thereby provide additional stability to the base. This adaptive, retractable leg synchro-drive mobile base uses a third drive source to perform the extension and retraction of the leg assemblies and provides that the wheels maintain their orientation while extension or retraction occurs while the mobile base is in translation and that the wheel orientation returns to its previous state if retraction or extension occurs while the base is not in translation.

kludge-rajan85-1 001-x640 (10)

kludge-rajan85-1 001-x640 (11)

kludge-rajan85-1 001-x640 (12)

kludge-rajan85-1 001-x640 (13)

kludge-rajan85-1 001-x640 (14)

kludge-rajan85-1 001-x640 (15)

kludge-rajan85-1 001-x640 (16)

kludge-rajan85-1 001-x640 (17)

Source: Basic Robotic Concepts, John M. Holland, 1983.

Synchro Drive (Author)
The synchro drive system shown in Fig. 3-18 features three or more wheels (in this case four) that are mechanically synchronized to each other for both steering and power. Synchronization can be accomplished by the use of chains (as shown), or by gears. Each wheeled "foot" assembly contains a 90° miter gear arrangement as shown in Fig. 3-19. The housing of the foot is driven by the steering chain, while the inner shaft is connected to the drive chain. The

Fig. 3-18. Synchro-drive using chain coupling.

Fig. 3-19. Steering action of the synchro-drive foot assembly.
system offers some interesting characteristics when it is driven. Since the wheels steer together, the base does not change its rotational orientation when the robot executes a turn. For this reason, the upper torso of the robot (which contains the vision and ranging systems) is pivoted and mechanically linked to the steering chain. By driving the steering chain with a stepper motor (and gear reducer) the robot can execute very precisely controlled turns.
Notice that the miter gear (Fig. 3-19) must be on the opposite side of the power shaft from the wheel. This is because of the interplay between steering and wheel drive. If the power chain is stationary (the robot is not moving), and the steering chain causes the foot to execute one complete revolution, the wheel power shaft will experience the equivalent effect of one revolution in the opposite direction. With a miter gear having a ratio of 1:1 and located as shown, the wheel axle will revolve once in such a way that the wheel rolls around an arc as shown in the figure. This action is much easier on treated
surfaces than having the wheel pivot about its center line. With the 1:1 ratio, the robot will not wobble as the turn is executed, if the wheel radius (r) is equal to pivot radius (r'). Unfortunately, this may mean that in the inboard rotational position (the right-most wheel in Fig. 3-18A) is displaced sufficiently under the robot to cause a deterioration of the zone of stability. If the pivot radius (r') is shortened, the robot will appear to "belly dance" as the steering is operated. If this is objectionable, the miter gear can be selected to have a ratio equal to the ratio of the circumferences of the two circles associated with r and r'. This of course means that it will be the ratio of the two radiuses.
The system has some attractive qualities, but it is not overly stable because it cannot adapt to steep terrain. For robots operating on flat surfaces, it is a good alternative.
•    Efficiency: Fair to good
•    Simplicity of Control: Excellent
•    Traction: Good
•    Maneuverability: Excellent
•    Navigation: Excellent
•    Stability: Fair to poor depending on the number of wheels
•    Adaptability: None
•    Destructiveness: Excellent
•    Climbing: Poor
•    Maintenance: Fair to good
•    Cost: Low to moderate
Adaptive Synchro Drive with Retractable Leg Assemblies
The adaptive synchro drive represents a compromise between the complexity of a walking robot, and the simplicity (and poor stability) of the synchro drive. The requirement that led to this design was to provide a robot that could negotiate relatively steer ramps, and that could climb over mild curbs without falling over. The solution was to add a degree of adaptability to the synchro drive. This was to be accomplished in the simplest possible manner, and the result is shown in Figs. 3-20, 3-21, and 3-22.
The adaptive synchro drive has the same basic power and steering chain arrangements in the base as did the original system, except that at the positions where the wheels were, there are now pivoted

kludge-rajan85-1 001-x640 (18)
Fig. 3-20. Synchro-drive leg assembly.
leg assemblies. A set of two chains in each leg transmits the drive and steering control to the same type "foot" assembly used in the synchro drive. Fixed to each leg at the pivot point is a chain sprocket connected to a third stepping motor and gear box arrangement. The power and steering shafts run concentrically through this sprocket into the leg. This arrangement allows the legs to rotate about their pivots (Fig. 3-23), thus changing the effective area of the base. As the legs rotate, the linkages are such that the wheels continue steering in the original direction. Since there is no capability of steering the wheels individually, they cannot be caused to "toe in" during collapsing. For this reason, a certain amount of rolling motion is required to allow this action to take place without damaging the surface on which the robot is running. To accomplish this, the action of the retraction (collapse) motor can be locked to the main drive motor. If the legs must be retracted in a short distance, the robot may have to roll forward and backward once or twice. Alternatively, it can execute a continuous turn with the drive motor turned off. Another disadvantage to this system is that the base orientation cannot be controlled. Thus if the robot approaches a load with the legs extended, it must make do with whatever the orientation of the legs might be. A modification to the steering chain may be added to overcome this problem.

kludge-rajan85-1 001-x640 (20)
Fig. 3-21. Diagram of a synchro-drive with retractable leg assemblies.
This modification would consist of three movable idlers taking up chain slack between the wheels. In the normal position of these idlers, the wheels would all steer in the same direction. In the other idler position the wheels would all steer tangentially and the robot base could pivot on its own axis. Unfortunately, this reduces the simplicity and economy of the design.
Like the synchro drive, this carriage can steer through 360 degrees without moving. This eliminates the need for backing up. There are several advantages to avoiding this maneuver, including the elimination of rear facing obstacle detection systems and the elimination of polarity reversing circuitry on the main power motor. This second factor improves efficiency since solid-state polarity reversing circuitry always induces some power loss.
This whole arrangement is of course an elaborate compromise, but it was one that satisfied our needs. The lack of independent wheel steering control was traded off for the simplicity of control and economy of having only one drive and one steering motor. The two main sacrifices that had to be made were in the area of efficiency and maintenance.

kludge-rajan85-1 001-x640 (19)
Fig. 3-22. Photograph of a synchro-drive with retractable leg assemblies (battery removed from foreground).
My experimental version of this system (named "Kludge") has been very successful, and the efficiency is better than expected. With 8-inch diameter tires and its legs extended, Kludge can climb over 4 X 4-inch timbers (actually 3.5 X 3.5 inches). My totally unbiased assessment of this approach is:
•    Efficiency: Fair to good
•    Simplicity of Control: Excellent
•    Traction: Good
•    Maneuverability: Excellent
•    Navigation: Excellent
•    Stability: Fair with legs retracted, excellent with them extended
•    Adaptability: Adaptable to ramps and small single steps
•    Destructiveness: Excellent with precautions mentioned above
•    Climbing: Poor
•    Maintenance: Fair (I hope)
•    Cost: Moderate

kludge-rajan85-1 001-x640 (21)
Fig. 3-23. Rotational collapsing and the change of the zone of stability.

….I would also like to thank Mr. Robert Pharr of Roanoke, Virginia, whose mechanical expertise brought the mobile robot "Kludge" from a concept to a reality. Thanks must also go to my technician, Mr. William Grady Spiegel, for his support in breadboarding many of the circuits associated with Kludge and other systems discussed in this book…

kludge-rajan85-1 001-x640 (1)

The Old Robots conveniently has a pdf of this article.

kludge-rajan85-1 001-x640 (2)

Source: ROBOTICS AGE January 1985 [edited]
John M. Holland
Cybermation, Inc.
When human beings attempt to solve a problem, they tend to match successful past solutions to the new situation. While this problem solving technique is extremely helpful in day-to-day situations, it can be misleading when we attempt to solve unique new problems. The trouble is that this whole conceptual process is so unconscious that we are unaware of the assumptions we make along the way. The problem of robot mobility is an excellent example.
When we start thinking about robot mobility systems, we immediately catalog the solutions to mobility problems in other fields. Most present-day mobile robots use a version of the mobility system originally designed for either a tricycle, a wheelchair, an automobile, a tank, an all-terrain vehicle, a wagon, or a combination of these. In some cases, the designer has turned to nature for inspiration and the result may resemble a spider or even an elephant.
Many mobile robots are well adapted to the problems they are designed to solve. For example, robots like the Ohio State University (OSU) Hexapod or the Odetics Odex I walker are required for certain rough-terrain applications. In fact, an article by R. B. McGhee, et. al. (see references) shows that walking robots may actually be more efficient than wheels or treads on soft surfaces. Still, it is very important to realize the original problem for which the technique was developed.
Attempting to apply existing vehicle designs to robots quickly points out the difference between the intelligence and sensory capabilities of a robot "driver" and a human operator. The robot driver will be a relatively stupid, nearly blind computer. Expecting a robot driver to perform the classical parallel parking maneuver for an automobile is optimistic in the extreme.Solutions based on animal models have an additional problem since animals are constructed from different materials than those available for the robot. For example, muscle tissue provides both an excellent lightweight servomechanism, and compliant springiness that can be used to store and recover kinetic energy.
Before designing a robot mobility system, we must determine the robot's intended capabilities and make several trade-offs for cost vs. performance.
The first major trade-off is between walking and rolling. While a walking robot can go almost anywhere, it will tend to be very complex mechanically, difficult to control, expensive to build, slow, and (on finished surfaces) relatively inefficient. Some of these difficulties can be eliminated, but only at the expense of making others worse. Depending on the applications, the ability to climb stairs, rubble, and undefined obstacles, may outweigh all other considerations. On the other hand, it may be more economical to replace stairs in the robot's environment with ramps, thus making a rolling robot acceptable.
There are always restrictions on the robot as well. These restrictions include the width and height clearance available, the maximum weight, the damage that it can be permitted to inflict on its running surface, etc.

As the robot design progresses, it is sometimes necessary to back up and modify or subdivide the original performance envelope. For example, two models of the robot may become necessary to fulfill all requirements, or perhaps an ability may be dropped rather than modify the robot's operating environment.
The example used in Table 1 evaluates applications we had in mind for our Cybermation robots and is an approximate evaluation of our first prototype. These robots would be expected to perform teleoperated and autonomous functions in demanding industrial applications, including explosives factories, clean-rooms, and nuclear reactor buildings. Thus, as reflected in the table, initial cost is not as important as the maintenance costs and reliability. The ability to climb steps was dropped in favor of requiring ramps and lifts.

The base consists of three wheel assemblies located on retractable legs. We call this the Synchro-drive since a set of chains is used to synchronously steer and drive all three wheels. The robot has three sets of motors, gear boxes, and chains; one for driving the wheels, one for retracting and extending the legs, and one for steering the wheels. Additionally, the steering chain is connected to a spine shaft running up through the center of the base. The robot's turret is mounted on a flange attached to this shaft, and rotates with the shaft in such a way that the turret always points in the same direction as the wheels.
This configuration gives the Synchro-drive some interesting capabilities. For example, the base does not rotate as the robot executes a turn. Not only does this save energy (by not requiring rotational acceleration and deceleration of the base), but it also allows the robot to maintain a sense of direction, by measuring the angle of the turret and base. One of the greatest advantages of the Synchro-drive is that steering and drive commands represent a pure polar coordinate reference system. This greatly simplifies navigation.
Furthermore, since the Synchro-drive has a true zero turning radius, it does not need reverse. This means that rear-facing sensors, and two (expensive) quadrants of the drive motor control can be eliminated.

Finally, the wheel assembly is designed to allow turns without translation. This was accomplished by off-setting the wheel from the center of the steering axis and placing its driving gear in such a way as to impart a rolling action during steering. The robot can thus turn in place, without damaging carpets, tile, or wood floors.


Table 2. Evaluation table for the Synchro-drive robot with concentric shafts and fixed legs.

The Synchro-drive approach evolved after many (informal) cycles through the evaluation process just described, and yet the approach still had short-comings at the point shown in Table 1.
The prototype (nicknamed Kludge) showed that the basic mode of movement was largely superior to other modes being considered, but the chains were a real problem. First, chains don't like to operate in a horizontal plane, and at least 180 degrees of engagement or purchase is required on each sprocket. This meant that many idlers had to be installed, which lowered efficiency and increased costs. Secondly, the chains stretch over time and must be adjusted. Additionally, the chains took up a lot of room and forced the robot's center of gravity to be higher than necessary. As a general rule, chains are noisy and dirty by nature. Finally, the wheels had to be realigned each time the chains were adjusted or tightened.
Each of these problems could be lessened by one measure or another, but the approach kept coming up short of our goals.
The problem then became how to build a robot that had all the good qualities of Kludge but was clean, reliable, easily assembled and repaired, had a lower center of gravity, and required no realignment. It wouldn't hurt if the new approach was (in the jargon of the patent office) clearly novel as well. This would allow us to obtain patent protection for the engineering investment.
The solution was to use a unique combination of concentric shafts and bevel gears. With this configuration, the moving parts could be enclosed inside hollow tubes comprising the robot frame members. This eliminated pollution, and greatly reduced maintenance. Accurately keyed gears kept the steering in alignment at all times.
The second production prototype (K2A) contains only a handful of different types of bearings and gears that are used repeatedly throughout the design. Furthermore, the new approach allows the batteries, drive motor and gear box to be slung between the leg members, lowering the center of gravity. By doing this, and by extending the fixed legs slightly beyond the edge of the base, the robot is about 80 percent as stable as the first Kludge prototype with its legs fully extended, and about 160 percent as stable as the first prototype with its legs retracted. Although an extensible-leg version using concentric shafts is planned, the cost savings on the current model outweigh the loss of high-end stability, at least for most current applications.
The result of these improvements is shown in Table 2. Notice that for the relatively small loss of stability, the savings in other areas are substantial. As an additional advantage, the maneuver required for extending and retracting the legs was eliminated.
The emergence of mobile robots as an important economic reality will require the rethinking of the basic precepts of mobility. These new mechanical "beasties" will encompass an enormous variety of forms, each governed by the niche it is intended to fill. Exactly as in nature, those robots that best fill the requirements of their niche will flourish and evolve, and those that are hastily or ill-conceived will become extinct.
We have used the Cybermation Synchro-drive as an example, but the basic process of fitting a solution to the problem can be used in the development of any robotic system.
McGhee, R.B., KW. Olson, and R. L. Briggs "Electronic Coordination of Joint Motions for a Terrain Adaptive Robot." Society of Automotive Engineers, Inc. Warrendale, PA.
Raibert, M. H., et al. "Dynamically Stable Legged Locomotion." The Robot Institute, Carnegie-Mellon University, Pittsburgh, PA. September 1981.
Holland, J. M., Basic Robotic Concepts. Howard W. Sams Publishing Co, 1983.

"Kludge" was a prototype. It evolved into the successful K2A then the K3A, subject to another post.

The company Cybermation became Cybermotion, and now KineLogic.

See other early Humanoid Robots here.
See other early Mobile Robots here.

1984 – FETAL I Omnidirectional Robot – William H. T. La (Vietnamese/American)



FETAL I had its major public appearance at the International Personal Robots Congress (IPRC) held in Albuquerque, New Mexico in 1984.

iprc3-ra-aug84 001-x640

Fetal I, constructed by Bill La, is a three-wheeled vehicle capable of moving in any direction, A later prototype, Fetal II [no picture available], was presented a Golden Droid award at the 1984 IPRC.



FETAL I at the IPRC 1984. Images by Richard Moyle via David Buckley's Historic Robots.


IPRC-Robotics Age Aug 1984.
Sunday was the day everyone had been waiting for, the awards brunch, Nelson Winkless, the official historian for IPRC, acted as master of ceremonies. After his opening remarks and thanks to the many behind-the-scenes personnel, Nels got down to the most important part of the program, the presentation of the Golden Droid awards.
The Golden Droid awards were presented in three categories: the most entertaining going to Bruce Taylor for his robot Henry; the most useful being presented to Reza Falamak and his EZ Mower Robot; and the open award going to Bill H. T. La for his Fetal II [Ed. not shown.]. After the picture-taking and congratulations were over, it was back to the exhibition hall for the final afternoon of the show, Although Sunday's attendance was somewhat less than the two previous days, the enthusiasm was still evident.


Dr. Bill La with his wife.

Venture – Volume 6, Part 2 – Page 88

William H. T. La, 33, was a Vietnamese exchange student and maker of toys when he invented the Alexis while playing with an erector set. He found that by placing castings on three wheels, the wheels could move in any direction.


Patent information:

omni-1 001-x640

omni-2 001-x640

Source: Robotics Age, Feb 1984.





Publication number    US4237990 A
Application number    US 06/000,570
Publication date    9 Dec 1980
Filing date    2 Jan 1979
Inventors    Hau T. La
Original Assignee    Hau T

Omnidirectional vehicle
US 4237990 A
A wheeled vehicle provided with three individually driven wheels rotatable on horizontal axes. The wheels are disposed at the corners of a triangle. The periphery of each wheel is defined by a plurality of rollers rotatable on respective axes which are each at an angle to the axis of the respective wheel. The axes of three rollers, one for each wheel, when each such roller is in its lowermost position, form a triangle. Each roller axis may be at right angles or perpendicular to, or at 45 degrees to, or at some other acute angle to the respective wheel axis, and the triangle may be an equilateral triangle in a typical embodiment of the invention.
According to a preferred embodiment of the invention, no two of the wheel axes are aligned or parallel to each other. In a typical construction, the wheels are at the corners of an equilateral triangle and the wheel axes intersect at the center of this triangle. The vehicle may be driven over a surface, or it may be inverted and an object with a surface engaged on the wheel rollers may be moved with respect to the stationary "vehicle". Controls for motors driving the wheels of the vehicle are provided to produce rectilinear movement, rotational movement or curvilinear movement of the vehicle over the surface.

Alexis, the Omnidirectional Wheelchair.

omni-c2 001-x640

An earlier prototype of the omnidirectional wheelchair.

Source: Basic Robotic Concepts, John M. Holland, 1983
A version of this drive was developed by the Veterans Administration as a transport system for paraplegic persons (Fig. 3-28). This system powers only the axial motion of each wheel, allowing the smaller outer wheels to roll freely. The effect is that these rollers act as force translators. This effect can be seen for the case of forward drive shown in Fig. 3-29B. Notice that only the two rear wheels are powered for this maneuver and that their outward force vectors cancel. This scheme greatly reduces the complexity of the carriage, but some loss of traction will occur. The small diameter of the rollers will also cause difficulty on surfaces that are not perfectly smooth.
•    Efficiency: Good
•    Simplicity of Control: Excellent
•    Traction: Good to fair (for single-axis drive)
•    Maneuverability: Excellent
•    Navigation: Excellent to fair (for single-axis drive)
•    Stability: Fair to good depending on mounting
•    Adaptability: None
•    Destructiveness: Excellent
•    Climbing: Poor
•    Maintenance: Poor to good (for single-axis drive)
•    Cost: Moderate (in production) to good (for single-axis drive)

omni-c3 001-x640


Dr. William H. T. La is the inventor of the Alexis, at the V.A. Institute in Palo Alto. The Alexis is a nonconventional "smart" wheelchair that uses a system known as "metamotion", employing three independently motor-driven nonparallel wheels linked by an on-board computer. This feature, patented by Dr. La in 1980, allows the Alexis to move directly to any point on the horizontal plane by the rider's manipulation of a joy stick that sends an electronic signal to the computer that controls the directions of the three wheels. The Alexis also has a control feature that enables a rider unable to manipulate the joy stick to operate the Alexis by head motion. As a result, any rider of the Alexis, regardless of disability level, could make it "turn on a dime" and maneuver it in cramped spaces.

The world's most futuristic wheelchair was designed and patented by Stanford University in 1982. It's omni-directional wheels made it truly revolutionary for its time. It was named in honor of Kim Alexis.

The Evening Independent – Sep 24. 1984  
'Alexis' the wheelchair called a significant step
As technology advances, entrepreneurs put it to use—quite often, to the advantage of victims of disease and infirmity, On this page, the focus is on a new computer assisted "sports car" wheelchair.
Knight-Ridder Newspapers
SAN JOSE, Calif, — Robert Smith slides into the $1,300 bucket seat of his sleek, computer-assisted wheelchair, fingers the control panel at his left hand and the joystick in his right, then zips off for a quick spin around a local shopping center,
At a top speed of 12 mph, the machine isn't exactly primed for the Daytona 500, But "Alexis," as Smith has dubbed this roadster-like tricycle for the disabled, which he helped design, enjoys advantages unknown to stock car racers.
At the shopping center, those advantages quickly proved themselves, Smith didn't swerve when shoppers stepped in his path because Alexis, unlike conventional wheelchairs, can move sideways as well as forward and backward.
The vehicle is part ballerina, too — it can pirouette within its own footprint, whereas ordinary wheelchairs have a turning radius of about one yard. This gives Alexis the kind of maneuverability that is critical in tight spaces, such as between racks of clothing in apparel stores or between a desk and wall at the office.
And belying the notion that wheelchairs will always be drab, pitiful contraptions, Alexis got some admiring glances as it sidestepped and twirled for the curious shoppers, A list of the world's 10 sexiest machines, Smith knows, wouldn't include present-day wheelchairs — an image he hopes Alexis will shatter so that, eventually, paraplegics and others can pride themselves on what may be their only means of powered transportation.
Indeed, Smith was thinking of Kim Alexis when he christened the wheelchair. She's the stunning blonde who last year modeled a red bathing suit in Sports Illustrated's swimsuit edition.
Smith, 24, who graduated from Stanford University in 1982 with a master's degree in mechanical engineering, and four others designed and built the futuristic wheelchair over a six-month period at the Veterans Administration Rehabilitation Research and Development Center in nearby Palo Alto, Tim Prentice, a high school student at the time, provided rough sketches of what was to become Alexis.
Smith's goal was to innovate, to build a machine that would include a Zilog Z-80A microcomputer to adjust the speed of the three independently powered wheels so the vehicle would move in precise directions at precise speeds. The microcomputer performs this command-and-control loop 20 times a second.
"It's almost as if you were going to build a sports car," Smith said of his design approach. "You can either soup up a Chevy Nova or start with a clean sheet of paper and design a Corvette."
Each wheel consists of eight natural-rubber rollers instead of treads, allowing the wheelchair to travel in all directions without any drag. The Boeing Co, owns the patent on this design for its shop carts.
The wheels — two in front and one in back —and motors are concealed by rounded fiberglass covers adorned with a red racing stripe. Adding to this airstreamed, classy look is the removable bucket seat manufactured by Recaro, the type commonly found in sports and racing cars.
"Actually, I'm kind of tired of that designs" Smith confided, "It looks like a vacuum cleaner,"
Perhaps, but Alexis doesn't sound like one. Its advanced electronics create very little noise. Moreover, the microcomputer won't let Alexis travel so fast that it tips over, and because all three motors are identical, unlike conventional models with left and right motors, parts are interchangeable.
That's important because wheelchair manufacturers don't stock many spare parts.
International Texas Industries of San Antonio has purchased the patent on Alexis and will manufacture it in Wichita, Kan. The first delivery is due Dec, 31. Initially, 500 of the wheelchairs will be test marketed only in the San Francisco area, so engineers here will be able to easily spot and correct any glitches.
The standard model will sell for $4,000 to $5,000, Smith estimates. Wheelchairs currently on the market, which is dominated by a company called Everest & Jennings, vary from $4,000 to $15,000, depending on the number and type of accessories the buyer needs.
Experts in and out of the VA agree that Alexis marks a significant step in wheelchair technology.
"This is the single most important contribution to mobility of the disabled since electric-powered wheelchairs were introduced," said Larry Leifer, an associate professor of mechanical engineering at Stanford University and director of the VA's research and development center.
"Most offices are inaccessible to wheelchairs. We expect Alexis to give (disabled) people a wider choice of places to live — without modifying that house, which is very expensive — and we expect it to give them broader employment opportunities."
David McGowan, executive director of the Adult Independence Development Center in Santo Clara, agreed.
"I'd say it's a rapid evolution," McGowan said. "It would be a significant change.
"Although federal law mandates that new projects be accessible to the disabled, we live in a reality where most buildings are not accessible. Doorways and hallways are not wide enough, for example. That would make (wheelchair) maneuverability critical."
The VA, according to McGowan, is at the forefront of such innovations because of the financial resources available to it.
As with any new mechanism, development of the Alexis prototype has fostered its share of headaches. Today, the wheelchair — itself disabled — sits in Smith's cramped laboratory near the VA Medical Center waiting for a new joystick. The original throttle was much too stiff for the kind of fingertip control for which Smith is striving.
In addition, the two fiberglass and nylon chassis plates began showing immediate signs of wear, forcing a switch to pure fiberglass to ensure durability while still offering a smooth ride. Another problem was getting parts: Smith often found that, because of the hospital environment, drugs and toilet paper seemed higher on the VA's list of items to be ordered.
The test pilot who puts Alexis through its' paces is Peter Axelson, 27, another mechanical engineer at the research and development center. Axelson lost the use of his legs eight years ago when he fell 180 feet while learning to rock climb as an Air Force cadet.
In subsequent years, he founded Beneficial Designs of Santa Cruz where he designed the sit ski, similar to sled.
"The initial sense of being able to move in any direction in Alexis is incredible," he said. "That's the most profound feeling, I believe that most people would get into Alexis and not try to move in any direction but backward and forward because moving sideways is so unusual."
The wheelchair negotiates tight indoor spaces better than it does curbs, hills and other outdoor obstacles. Yet, all devices have their limits, and Alexis is no exception, Axelson said.
"It's like the difference between a long distance runner and a dancer, Alexis is agile."

Patent information:




Publication number    US4715460 A
Publication type    Grant
Application number    US 06/673,965
Publication date    29 Dec 1987
Filing date    20 Nov 1984
Also published as    EP0201592A1, WO1986003132A1
Inventors    Robert E. Smith
Original Assignee    International Texas Industries, Inc.
Omnidirectional vehicle base
US 4715460 A
An omnidirectional wheelchair base 7 includes upper 10 and lower 20 flexible base plates held in spaced-apart alignment by a pair of front supports 14 and 16 and a rear support 18. A pair of front wheels 22 and 26 are provided, each mounted on the front supports 14 and 16, respectively, and each having an axis of rotation wherein the angle between the axes of rotations of each of the front wheels is less than 180°. A rear wheel is mounted on the rear support and has an axis of rotation less than 180° from each of the front wheels.

Publication number    WO1986003132 A1
Publication type    Application
Application number    PCT/US1985/002281
Publication date    5 Jun 1986
Filing date    19 Nov 1985
Also published as    EP0201592A1, US4715460
Inventors    Robert E Smith
Applicant    Int Texas Ind Inc

Alexis Wheelchair – Last word:
Alexis is an innovative electric wheelchair using a "wheels within wheels" design. It is unique in that it can turn in its own footprint and move sideways. The Rehab R&D Center licensed Intex Industries to make Alexis commercially available in 1987, and Intex made 40 pre-production units for field trials in the San Antonio area. During subsequent redesign efforts, the company filed for bankruptcy, preventing further commercialization at the time.

From the video blurb: Unfortunately, Alexis never made it to market because Jon King, Intex CEO, embezzled millions of investment capital. He was later convicted and spent 10 years in federal prison for his crime.

Mobile Vocational Assistant Robot (MoVAR): 1983-1988.
Overview of project

The MoVAR project used a unique, patented, 3-wheeled omni-directional base with a PUMA-250 arm mounted on it. It was desk-high and designed to go through interior doorways (see Figure 2). All electronics and power components for the motors and sensors were mounted in the mobile base. A telemetry link to a console received commands and sent position and status information. The mobile base had a bumper-mounted touch sensor system to provide autonomy in the face of obstructions, a wrist-mounted force sensor and gripper-mounted proximity sensors to assist in manipulation, and a camera system to display the robot's activities and surroundings to the user at the console. The robot console had three monitors: graphic robot motion planning, robot status, and camera view. It had keyboard, voice, and head-motion inputs for command and cursor control, and voice output.

Omni-directional mobile robot called MoVAR.

Figure 2: MoVAR robot base with instrumented bumpers and joystick; the PUMA-250 carries a camera for remote viewing, a six-axis force sensor and gripper with finger pad-mounted proximity sensors. A wireless digital link allows the mobile base computer to communicate with the user console. A later phase of this project added instructable natural language input capability, coupled to an internal world model, so that typed-in commands such as, "move to a position in front of the desk, and move west when the bumpers are hit" could be executed. Path planning was not a targeted research area for this project due to the many other research groups active in this domain.

This project was stopped in 1988 when VA funding was terminated. The hardware and software were subsequently transferred to the Intelligent Mechanisms Group at the NASA Ames Research Center (Mountain View, CA) for use in the development of real-time controllers and stereo-based user interfaces for semi-autonomous planetary rovers.

See other early Humanoid Robots here.
See other early Mobile Robots here.