GB2278041A - Exoskeletal system - Google Patents

Description

2278041 EXOSKELETAL SYSTEM The present invention relates to systems for

enhancing the ability of a human to walk, jump, exert loads, carry loads, and the like, and, more particularly, to exoskeletal systems which are connected to a human operator at a limited number of points, but including to the operator's feet, and which serves to enhance the operator's ability to transport himself and any loads, including the apparatus itself, which the operator may wish to carry.

Various technologies have been developed over the years to rapidly and efficiently transport a human from one place to another. Similarly, considerable technology exists for carrying out various otherwise difficult tasks, such as lifting and transporting heavy loads.

However, in certain applications, existing technologies are inadequate in that circumstances serve to prevent access to a machine which would otherwise be able to transport the human and/or carry out the desired task. To cite just one example, it is generally agreed that in various contexts, the properly equipped rescue worker, sportsman, mountain climber, and the like, can still serve an indispensable role through his ability to traverse terrain which is otherwise impassable and to accomplish tasks which are beyond the ability of machines.

It has, thus, long been desired to develop technology which will enhance the ability of a human to walk and carry out 1 various tasks, such as lifting and carrying various loads, while on his feet.

Several attempts have been made in the past few decades. One such system, designated Hardiman, was developed several decades ago by General Electric and involves a robot-like suit which the operator wears over his eniire body and which he controls through a series of actuators and servo controllers. The suit is cumbersome, technically complex and inflexible. Only the arms of the suit have bilateral abilities with the machine legs being used as simple support structures based on a master-slave system.

More recently, a system designated Pitman has been proposed by the U.S. Advanced Weapons Technology Group at Los Alamos National Laboratory. The Pitman system, like its Hardiman predecessor, involves a suit which is worn by the user. Unlike the Hardiman suit, the Pitman is more flexible and features various subsystems offering protection against armorpiercing rounds, chemical and biological weapons, and laser and nuclear effects. The Pitman proposal does not explain how to control motion of the suit.

Further developments have been made in creating an anthropomorphic structure which will enhance the capabilities of a human but which will not involve cumbersome and inflexible suits, but rather will include members which are somewhat distanced from the operator, allowing the operator to comfortably bend, run and carry out various other bodily movement relatively unconstrained by the system.

2 Efforts to develop such exoskeletal systems have been largely confined to the development of mechanized arms. Illustrative of these efforts is the work of Stephen Jacobsen at the University of Utah, which includes his 10 degrees of freedom hydraulically powered anthropomorphic arm. The arm is a teleoperated slave and kinematically corresponds one-to-one to its master. The load-reflective master is equivalent to the operator's arm. When the slave grips an object, the exoskeletal master applies loads to the operator's arms and fingers which are directly proportional to the loads applied to the object. A disadvantage of this system, which is shared by all master/slave systems, is its great complexity, requiring as it does, two sets of actuators, sensors and electronics.

More recently, H. Kazerooni, initially at the Massachusetts Institute of Technology and more recently at the University of California, Berkeley, has developed mechanized arms using systems based on the direct actuation of the slave. The powered "extenders" augment the user's natural arm strength. Physical contact at the interface between the extender and the operator allows for the transfer of mechanical power and information signals. Thus, control of the extender is accomplished without any type of joystick, keyboard, or master/slave system.

In Kazerooni's system the user inserts his arm into a rubber cylinder located inside the extender. Piezoelectric load cells, placed between the cylinder and extender's outer shell, sense the interactive loads between the user's arm and the machine. A second set of load sensors determine the weight and acceleration of the load being maneuvered. Parallel processors 3 define the dynamics of operator motion, matching machine impedance to that of the operator. Because load reflection occurs naturally, the human arm feels a scaled-down version of the actual loads applied to the extender.

While the Kazerooni load control system is considerably less complex than previous systemst such a system cannot be applied to enhance the capabilities o f the user's legs, which requires a more sophisticated system to sense and properly account for the inherent stability problems encountered in such a system.

In many applications, it is the enhancement of the functioning of the user's legs, such as the ability to walk quickly and for an extended time period, which is of utmost importance. Such systems allow the user to sense the center of gravity of the man-machine ensemble and dynamically couple the user with the machine.

At least one system has been developed to enhance the capabilities of the user's legs. The system, known as the Spring Walker, is disclosed in U.S. Patent No. 5,016,869. The Spring Walker features leg extensions connected to a set of springs and pulleys which alternately store and release energy to provide more efficient locomotion. The Springwalker, which features a pair of backward-facing knee joints, makes possible trampoline- like jumps and large steps, through use of energy storing springs, with the user being supported by the machine. Control of the device is by direct mechanical connection between the user and the machine. While the device is useful in certain 4 limited contexts, the Springwalker suffers from a stability problem when standing. The device is inherently unstable when not in motion. In addition, the coupling of the user and machine is such that the freedom of movement of the user is severely limited.

There is thus a widely recognized need for, and it would be highly advantageous to have, an impedance controlled exoskeletal system which will enhance the ability of the user to walk, run, jump, and the like, which will provide mechanical load amplification, including, for example, the ability to carry payloads, including the weight of the system itself and at least part of the weight of the userf which will exhibit balance and stability, and which the user will be able to don and remove quickly and easily.

SUMMARY OF THE INVENTION

According to the present invention there is provided an exoskeletal apparatus to help a user carry a payload while moving across the ground, comprising: (a) a right leg member attached to the right foot of the user so as to transmit loads in the right leg member directly to the ground; (b) a left leg member attached to the left foot of the user so as to transmit loads in the left leg member directly to the ground; and (c) a payload/user link connected to the payload and further connected to the right leg member and to the left leg member, the payload/user link located so that loads exerted by the user on the payload/user link act to guide the movement of the apparatus.

In preferred embodiments according to the present invention, the leg members are connected to foot members which are attached to the feet of the user.

Further according to the present invention, the exoskeletal apparatus to helri a user carry a payload while moving across the ground further includes (1) a pair of sensors for sensing loads between each of the foot members and the respective foot of the user; (2) a pair of sensors for sensing loads between each of the foot members and the ground; (3) a sensor for sensing loads between the user and the payload/user link; (4) an actuator for the right leg member; (5) an actuator for the left leg member; (6) a control system connected to the sensors and to the actuators for controlling the movement of the apparatus.

According to yet further features in the described preferred embodiments the right leg member and the left member 6 each includes a foot rest located substantially between the foot of the user and the ground and at least two segments, the segments being connected to each other through a joint, and the payload bearing member includes apparatus for at least partially supporting the weight of the user.

I The present invention successfully addresses the shortcomings of the presently known configurations by providing a relatively flexible and easy to deploy impedance controlled exoskeletal system which is capable of enhancing the user's walking, running, jumping, payload bearing and load exertion abilities, and which has suitable balance and stability.

7 BRIEF DESCRIPTION OF THE DRAWINGS AND APPENDICES

The invention is herein described, by way of example only, with reference to the accompanying drawings and appendices, wherein:

FIG. 1 is a side view of an apparatus according to the present invention as it might appear when user is in the standing position; FIG. 2 is a view of the apparatus of Figure 1 when the user is walking; FIG. 3 is a view of the apparatus of Figure 1 when the user is walking or running; FIG. 4 is as shown in Figure 3 but with the addition of apparatus to support part of the weight of the user; FIG. 5 is as shown in Figure 3 but with the addition of secondary machine pads located below the feet of the user; FIG. 6 is a schematic side view of an apparatus according to the present invention showing the interfaces between the user and the apparatus and between the apparatus and the ground or payload; FIG. 7 is a schematic depiction of the elements making up an apparatus according to the present invention; FIG. 8 is a schematic depiction of some of the key components of an apparatus according to the present invention; FIG. 9 illustrates the overall concept of master/slave system; FIG. 10 illustrates the overall concept of a system according to the present invention wherein a human directly controls a slave; 8 FIG. 11 is a schematic depiction of tracking with impedance control; FIG. 12 shows the system of Figure 11 in an idealized condition; FIG. 13 is a front cross-sectional view of a passive payload bearing apparatus according to the present invention; FIG. 14 is a side view of the apparatus of Fig. 16; FIG. 15 is a front view of an apparatus as in Figures 13 and 14, showing more construction details; FIG. 16 is a side view of the apparatus of.Figure 15; FIG. 16A is a view of the hip assembly of the apparatus of Figures 15 and 16; FIG. 16B is another view of the hip assembly of Fig. 16A; FIG. 16C is a view of the shoe portion of the apparatus of Figures 15 and 16; FIG. 16D is another view of the shoe portion of Fig. 16C; FIG. 17 is a side view of a shoe featuring an energy storage device attached to the sole; FIG. 18 is a bottom view of the shoe of Figure 17; FIG. 19 is a side view pictorial illustration of resilient element 74 of Fig. 6 which is sandwiched between interface 48 of Fig. 6 and the ground; FIG. 20 is a pictorial illustration of payload 40, payload stabilizer 68, resilient element 90 and frame 14 of Fig. 6; 9 0 FIG. 21 is a detailed side view illustration of the association between payload frame 14 and the belt 56 of Fig. 6; FIG. 22 is a pictorial illustration of an exoskeleton apparatus constructed and operative in accordance with an alternative embodiment of the present invention which is suitable for facilitating pedaling action by a human user; FIG. 23 is a pictorial illustration of connector 25 of Fig. 6; FIG. 24 is a schematic vectorian illustration of the basic structure of a control system used in the operation of the present invention; FIG. 25 is an ideal schematic illustration of the system illustrated in Fig. 24; FIG. 26 is a schematic illustration of the operation of the apparatus of the present invention, while attached to a fitness bicycle; FIG. 27 is a schematic illustration of a control system useful in the operation of the apparatus while attached to the fitness bicyle; and FIG. 28 is a schematic illustration of the servo velocity loop based on the CTT method.

Appendix A is a software simulation of an exoskeleton which is useful for facilitating stationary pedaling action. Appendix A has been filed with this application and Is contained in the Patent Office file so that it may be read In conjunction with the embodiments of the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention is of an exoskeletal impedance controlled robotic apparatus which can be used to enhance the ability of a user 'to walk and carry payloads.

The principles and operation of an impedance controlled enhanced walking exoskeletal apparatus according to the present invention may be better understood with reference to the drawings and the accompanying description.

Referring now to the drawings, Figure 1 illustrates a possible embodiment of an apparatus according to the present invention with the user in a standing position. The apparatus includes a right leg member 10 which is preferably attached at its lower end in some suitable fashion to a right foot member 18 which is, in turn, attached in some suitable fashion to the right foot of the user. Placement of right leg member 10 and right foot member 18 is such that loads in right leg member 10 can be transmitted directly to the ground whenever right foot member 18 is in contact with the ground, without having to go through any part of the user's body.

Throughout the specification and claims the term 'ground' is used to mean that portion of the user's environment which interacts with the foot members of the apparatus. In most cases this will be the ground surface. However, in certain other cases, the 'ground' may be, for example, the pedal of a bicycle, the rung of a ladder,.and the like. Use of the term 'ground' is made for convenience only and is not intended in any way to limit the scope of the present invention.

11 The apparatus also includes a left leg member 12, which may be attached to a left foot member 20, which are similar to right leg member 10 and right foot member 18, and which are preferably identical to them in every respect.

As can best be seen in Figure 6, each of leg members 10 and 12 is directly or indirectly connected near their upper ends through a suitable connector 25, which may be active or passive and which may include further components, to a payload frame 14 which is, in turn connected to the user through a frame/user interface 42 and also to a payload 40 through a frame/payload interface 16. The connections are such that at least part of the load exerted by the payload is transmitted to right leg member 10 and/or to left leg member 12 when direct or indirect contact is made between one or both of the leg members and the ground. In this way, at least a part of the total load is transmitted directly to the ground without requiring any portion of the user's body to carry the load.

Frame/user interface 42 is preferably connected to the user, for example, by attachment to the user's back, such that loads, which includes forces and moments, exerted by the user on frame/user interface 42 act to guide the movement of the apparatus. In Figures 1-6 frame/user interface 42 is depicted as being attached to the back of the user. Other locations are also possible, as is described below.

Each of leg members 10 and 12 preferably includes at least two segments 22. Each pair of adjoining segments 22 of a single leg member, 10 or 12, are connected to each other through a joint 24, which may be passive or active, rigid or flexible, 12 and which may include one or more additional components as described in more detail below. Segments 22 may be of the telescoping variety with one segment sliding within, or relative to, another. In such a case, joint 24 indicates the point of interaction of the telescoping segments 22. The configuration described allows the user to readily kneel, resting one or both knees on the ground.

Each of leg members 10 and 12 is preferably attached near its lower end to foot members, 18 and 20, preferably located substantially between the foot of the user and the ground, or at least sharing with the feet of the user an area of contact with the ground to allow for the transmittal of loads to the ground.

Figures 1-4 depict an apparatus according to the present invention as it might appear when used by a user in various positions. The user depicted in Figure I is in the standing position. Here it is desirable that the apparatus ensure the stability of the user and of the apparatus and communicate to the use information on the location of the center of gravity of the man/machine ensemble to enable the user to act to stabilize the ensemble. Stabilization may be accomplished through any suitable inertial stabilization system, including, but not limited to, systems including gyroscopes.

Depicted in Figure 2 is a typical walking position where, in addition to stabilization the ensemble, the apparatus must also facilitate the interaction between the user and the ground.

Shown in Figures 3, 4 and 5 is a typical walking or running position. Here the apparatus must ensure the proper acceptance of the load produced by the weight of the user, the apparatus and the payload. In addition, the apparatus works to stabilize the payload with respect to the user.

In Figure 4 the apparatus features support apparatus 30, typically in the form of a saddle, for supporting at least some of the weight of the user which could further enhance the ability of the user to travel at increased speeds and for longer times.

In Figure 5 the apparatus of Figure 3 is shown as having additional joints which tend to further separate the feet of the user, which rest on a top portion 32 of the footrest, from the ground, which comes in contact with the bottom portion 34 of the footrest, which is displaced from top portion 32. Such a configuration effectively lengthens the legs of the user and makes it possible to travel at greater speeds.

In a preferred embodiment of an apparatus according to the present invention, depicted in Figure 6 and described partly above, frame/payload interface 16 is located in the vicinity of the back of the user and supports a payload 40. Frame/payload interface 16 is, in turn, supported by payload frame 14 which is attached, through a frame/user interface 42 to the back of the user.

Each foot member, 18 and 20, preferably includes a structural member 44 which is connected to the corresponding leg member, 10 or 12, and a pair of interfaces -- a user/structural member interface 46 and a structural member/ground interface 48. Each of these interfaces may include various passive and/or 14 active components and preferably includes one or more load cells to measure one or more forces and/or one or more moments. Preferably, load cells measure three forces and three moments.

The configuration shown in Figure 6 makes it possible for the system to sense the load exerted by payload 40 and by the apparatus and user and enables the appropriate corrections to be made so that the loads generated by payload 40 do not adversely affect the user.

Preferably, the corrections are effected through a series of actuators (not shown) which are located in any suitable locations, including, but not limited to, joints 24 between adjoining segments 22. The signals from the various load sensors are received and analyzed by a control system, described in more detail below, which generates commands to the various actuators to enable the apparatus to function properly.

The overall system is depicted schematically in Figure 7. The entire system includes the user, the machine, the payload and a series of interfaces between the machine and the user, between the machine and the payload, and between the machine and the ground. Loads are communicated in both direction at each interface. Preferably the system also includes a power supply to augment the muscle power of the user.

Depicted in Figure 8 are some of the components of the machine. Sensor data is fed to a CPU unit which computes the desired trajectories of the various links, which links may be rigid or flexible. The information is used to drive a servo control unit which sends appropriate commands to the various actuators, which may comprise various hydraulic and/or pneumatic pistons. Feedback sensors close the feedback loop, monitoring the activity of the actuators and sending the information to the servo control unit. The actuators generate forces which drive various mechanisms which, in turn, cause the various links to move in the desired fashion.

The system of the present invention, unlike a master/slave system, calls for movements of the user to activate the machine directly. A typical master/slave system is shown schematically in Figure 9. In such a system commands from the user control the master. The master, in turn, commands the slave which interacts with the environment.

By contrast, in a system according to the present invention, shown schematically in Figure 10, there is direct dynamic interaction between the user and the slave. The system is controlled using impedance control, either passive or powered. The advantage of such a system is its simplicity brought about by the absence of the master which normally greatly complicates the system by requiring a second set of sensors, actuators and related equipment.

One possible impedance control scheme is depicted schematically in Figure 11. The user generates certain forces which are picked up by load sensors. Signals from the load sensors are then analyzed and the impedance control subsystem produces velocity commands to the servo loop, which interacts with the environment. The velocities produced by servo loop are fed back to the user to give the user a real time sense of the status of the man/machine ensemble. An ideal system, having 16 ideal sensors, instantaneously acting servo loop, and a perfectly accurate impedance model, would appear as shown in Figure 12.

In one set of embodiments of systems according to the present invention, the device is passive in that it includes no actuators or power supply,. One example of such embodiments is shown in Figures 13-16. Figures 13 and 14 are simplified front and side views, respectively, of a device which is useful in carrying payloads. The device includes a pair of leg members, 110 and 112, each of which is connected at its lower end to a foot member, 118 and 120, which are, in turn, attached in some suitable fashion to the feet of the user in such a way that loads in a leg member, 110 or 112, or both, are transmitted directly to the ground whenever the user's foot which is attached to that leg member contacts the ground.

The upper ends of leg members, 110 and 112, are connected in some suitable fashion to a harness 150 or frame which, in a preferred embodiment, surrounds the user around his midsection and is preferably connected in some suitable fashion to the user. Harness 150 can accommodate a payload (not shown), for example, by strapping the payload to, or by placing the payload on, harness 150.

The operating principle of such a payload bearing device is that at all times while the user is walking, at least one of his feet is in contact with the ground. Since the foot members are suitably attached to the user's feet, whenever one of the user's feet contacts the ground the corresponding foot member also contacts the ground and is thus able to transmit any loads 17 which are being supported by the corresponding leg member. The leg members are joined together through a common harness which, in turn, supports the payload to be transported. Proper design of the system, as described in more detail below, effects the shifting of the weight of the system, including payload, as the user's feet alternately contact the ground and lift off from the ground. The load shifting is accomplished in ways which are analogous to the functioning of the human hip joint, as is described in more detail below.

A more detailed version of a passive payload bearing device described generally above is shown in front view and in side view in Figures 15 and 16, respectively, and in the close-up views of portions of the apparatus shown in Figures 16A-16D.

The apparatus includes a load carrying frame 400 onto which various loads (not shown) can be attached or placed in some suitable fashion. Rigidly attached to load carrying frame 400 is a hip stabilizing assembly 402 which is described in more detail below with reference to Figures 16A and 16B. Hip stabilizing assembly 402 is connected, through a hip joint 404, to a thigh rod 406, which is, in turn, connected to a knee joint 408. Knee joint 408 connects thigh rod 406 to a shank rod 410 which is, in turn, connected to an ankle joint 412. Ankle joint 412 is rigidly connected to the user's shoe 444, preferably to the heel of shoe 412.

Load carrying frame 400 (which is not shown in Figure 16 for clarity of presentation) is preferably rigid. The interface between the user and load carrying frame 400 is preferably located in the mid-section of the user, including the back area.

18 The interface is preferably effected through a series of properly located trunk contact plates 414. Contact plates 414 are located so as to allow the user to transmit to load carrying frame 400 horizontal forces which are required to guide the apparatus.

Contact plates 414 are so placed so to allow for the relatively free relative vertical motion between the user and contact plates 414 such that virtually none of the vertical loads associated with the payload being carried are transmitted to the user.

As can be appreciated, whenever the user lifts one of his legs off the ground there is a strong tendency for that portion of the apparatus on the side of the leg which is being lifted to collapse downward under the combined weigh of the load and the weigh of the apparatus itself. To counter this tendency, use is made of hip stabilizing assembly 402 which is shown in more-detail in Figures 16A and 16B. Hip stabilizing assembly 402 includes an upper member 420, one portion of which is connected rigidly to load carrying frame 400, as through a pair of bolts 422 which grasp a portion of load carrying frame 400 in a slot 424 formed in a portion of upper member 420.

The lower portion of upper member 420 is pivotally connected about transverse pivot 426 to an upper arm 428 extending from hip joint 404. It is to be noted that the axis of pivot of transverse pivot 426 is perpendicular to that of hip joint 404. The combination of the two pivots enables a circular motion with two degrees of freedom between load carrying frame 400 and thigh rod 406.

19 To counteract the tendency of the apparatus to slump to one side when the user lifts his leg, the following arrangement is provided. Bolted to upper arm 428 is an insert 430 which extends laterally so that it is substantially parallel to, and below, an extension of upper member 420. The extension of upper member 420 is connected to a disc 434---viaa connector 432, which may be an adjustable screw, and which passes through insert 430. Located between disc 434 and insert 410 and between insert 430 and the extension of upper member 420 are suitable load damping materials, such as, for example, various rubbers, springs, and the like.

Hip stabilizing assembly 402 operates as follows. When the user lifts one leg off the ground, that portion of the apparatus on the side of leg being lifted tends to fall, or pivot about the other foot toward the user's body. The downward movement causes a pivoting motion about pivot point 426. However, the pivoting motion is damped and largely eliminated by theresistance of the damping materials which comes into play as insert 430 is made to approach the extension of upper member 420. In this way the hip is stabilized and prevented from dropping significantly while the leg is not in contact with the ground.

Also as the user lifts a leg and bends his knee to move forward, knee joint 408 is made to bend. It is important that knee joint 408 straighten prior to the placement of the foot back on the ground. To accomplish this, the apparatus is provided (see Figures 15 and 16) with a pair of thigh guides 436 which are connected to thigh rods 406 in such a way that forward motion of the user's thigh Dushes a thigh guide 436, and therefore also thigh rod 406, forward.

The bending of knee joint 408 brings about the stretching of an elastic member 438, which may be a elastic band, such as a bungee, and which is tied to thigh rod 406 and to shank rod 410. Elastic member 438 is preferably guided by a guide wheel 440 which is mounted on, for example, thigh rod 406. As elastic member 438 stretches it stores energy which is then used to bring thigh rod 406 and to shank rod 410 into alignment in the fully extended configuration prior to placing the foot on the ground.

Ankle joint 412 is shown in more detail in Figures 16C and 16D and includes a coiled spring 442 which connects between shank rod 410 and a heel connector 444 which is attached to the shoe, preferably the heel, of the user. Ankle joint 412 is designed to dampen the axial shocks and to provide a certain flexing without inhibiting the freedom of movement of the ankle of the user.

The efficiency of systems according to the present invention can be further enhanced by providing any of a number of suitable energy storage mechanisms, such as a spring, elastic member, or any suitable hydraulic or pneumatic system, and the like, which are capable of storing at least part of the energy available when the user's foot is lowered to the ground and which can then reinject at least a portion of the stored energy into the system so as to reduce the energy requirement of the user and/or system. Two examples of such devices, which can be used in conjunction with the shoes of the user, are shown in Figures 17-18 and in Figure 19.

21 Shown in Figures 17 and 18 is a spring arrangement which is connected to the sole portion of the user's shoe. Three spring assemblies 200 are shown in Figure 18. For clarity, only one of these is shown in the side view (Figure 17). As can best be seen in Figure 18, spring assembly 200 preferably includes a first spring 202 and a second spring..204 of different diameter properly mounted on a suitably shaped spool 206 which is connected to the sole of the shoe. The elasti:cities of the springs and their lengths are such that when the shoe first approaches the ground, only one of the springs is compressed. Further approach of the shoe to the ground results in the compression of both springs, with a higher overall elasticity. In this way the shock-absorbing properties of the assembly are increased while the assembly serves also to store energy which is released once the shoe begins to lift from the ground.

Another energy storage sysem for use on shoes is shown in Figure 19. The heel of the shoe includes a suitable spring 300, only the top and bottom coils of which are shown in crosssection, which loops are connected, respectively, to the shoe and to a displaceable heel 302. Disp2Laceable heel 302 is also connected to a pivot rod 304 which is pivotable about a pivot point 306 rigidly connected to the shoe near its anterior end.

During normal walking the user first places his heel on the ground and then rolls his foot forward as the user's weight shifts forward so as to place the foot flat on the ground and then so as to contact the ground only with the anterior tip of the foot just prior to disengagement of the foot from the ground.

The mechanism of Figure 19 takes advantage of this 22 walking sequence to facilitate walking. Specifically, spring 300 normally urges displaceable heel 302 downwards. When heel 302 first contacts the ground the force between the shoe and the ground compresses the spring, displacing heel 302 toward the shoe (shown in broken lines). As the shoe rolls forward, spring 300 displaces heel 302 downward which pivots pivot rod 304, parts of which are still in contact with the ground, so as to cause a certain lifting of the shoe from the ground, thereby returning some of the stored energy to the user and facilitating his walking.

Reference is now made again to Fig. 8 which is a block diagram of a preferred mode of control for the apparatus of Fig. 6. The apparatus of Fig. 8 includes a CPU 50 such as a Intel 80486, associated with a plurality of force and torque sensors, also termed herein "load sensors",. . and a plurality of servo systems. In the illustrated embodiment, the plurality of load sensors includes the following:

a. A load sensor 52 associated with the interface 16 between payload 40 and frame 14 which is operative to monitor interaction forces and torques between the frame 14 and the payload 40; b. A load sensor 54 associated with the interface 42 between frame 14 and a frame supporting member associatedwith the body such as a belt 56 (Fig. 6). Load sensor 54 is operative to sense interaction forces and torques between the human torso and the payload frame 14.

C. A load sensor 58 associated with the interface 46 23 between the structural member 44 and the user which is operative to measure interaction forces and torques between the human's foot and structural member 44; d. A load sensor 60 associated with the interface 48 between the structural member 44 and the ground. Load sensor 60 is operative to measure interaction forces and torques between the ground and structural member 44.

e. An inertial sensing system 62 associated with payload 40.

Load sensors 52, 54, 58 and 60 may each comprise a load cell such as an JR3 multi-axis force-torque sensor, commercially available from JR3, Inc., 22 Harter Ave., Woodland, CA 95695, USA.

Inertial sensing system 62 may comprise three accelerometers arranged to measure accelerations along each of the X, Y and Z axes. Each accelerometer may, for example, comprise a Model 3110 accelerometer, commercially available from EuroSensor, ICSensors, 20 - 24 Kirby St., London EC1N 8TS, UK.

The plurality of servo systems comprises the following:

a. A plurality of servo systems 64 including a servo system 64 for each of joints 24, 27 and 29; and b. a servo system 66 which is associated with a payload stabilizing element 68, as illustrated in detail in Fig. 23.

Each servo system 64 may comprise a servo actuator 80 (Fig. 6) controlled by a servo amplifier which receives feedback from the servo actuator via position and velocity sensors 82 (Fig. 6). The servo actuator may, for example, comprise a model 24 BMB-045 servo motor, commercially available from Bental Motors, Bental Development Ltd., Kibbutz Merom Golan, 12905, Israel. The servo amplifier may comprise a Mosfet PWM servo amplifier, commercially available from Elmo Motion Control Ltd., POB 463, Petah Tikva 49103, Israel. A suitable position sensor is an H27-HR-27 (Serie H -Taille 11) potentiometer, commercially available from MCB, 48 Avenue Kleber, Courlis 1, 97200, Colmbes, France. A suitable velocity sensor is a tachometer, commercially available from Inland Motor Division, Kollmorgent Corporation, Radford, VA, USA.

Two methods are now described which may be employed by CPU 50 in order to control the servo actuators, using input data provided by the sensors.

METHOD 1 may comprise the following steps:

a. Receive inputs from the sensors measuring the following inputs: 1. 60). 2.

3.

FGS = ground-structural element load vector (sensor FUL = frame-belt load vector (sensor 54) FUS- user-structural element load vector (sensor 58) Each load vector includes three linear force values and three torque values or moments.

b. Use conventional impedance control methods to compute, for each foot, a vector V as a function of the above input force and torque values, where:

V = desired velocity vector of the structural element 44 of an individual foot, relative to the frame 14. The V vector includes three linear velocity values and three angular velocity values, corresponding to the X, Y and Z axes, respectively.

For example, to control a swinging leg during walking or running, a value is assigned to V which brings FUS as close as possible to zero. To control a supporting leg during relatively slow motion, a value is assigned to V which brings FUL as close as possible to zero.

Conventional impedance control methods are used in telerobotics and are described in the following references, inter alia:

1. Sheridan, T. B., Telerobotics. automation, and human supervisory control, MIT Press, Cambridge, Massachusetts, USA, 1992.

2. Hogan, N., "Impedance control: an approach to manipula- tion: part I theory", Journal of Dynamic Systems, Measurement and Control, 107, pp. 17, March 195; 3. Hogan, N., "Impedance control: an approach to manipulation: part II - implementation", Journal of Dynamic Systems, Measurement and Control, 107, pp. 8- 16, March 1985; 4. Hogan, N., "Impedance control: an approach to manipula tion: part III - applications", Journal of Dynamic Systems, Measurement and Control, 107, pp. 17 - 24, March 1985.

5. Kazerooni, H. and Mahoney, S. L. "Force augmentation in human-robot interaction", Proceedings of the 1990 American Con trol Conference, Vol. 3, pp. 2821 - 2826, San-Diego, CA, USA, 1990.

6. Strassberg, Y. "A control method for bilateral teleop- erating systems", Ph.D. dissertation, Dept. of Mechanical Engi- 26 neering, University of Toronto, Canada, 1992.

C. Receive input regarding the positions and velocities of each joint 24. Use conventional inverse kinematics and inverse dynamics techniques to compute, as a function of the above de sired velocity vector V, final outputs which enable the desired velocity vector V to be achieved. These final outputs include the desired angular velocity of each segment or link 22 relative to each segment 22 adjacent thereto, which is operationalized by the velocity of the servo actuator 80 associated with each joint 24. Conventional inverse kinematics and inverse dynamics techniques are used in a variety of robotics applications and are described in the following references, inter alia:

1. Chae, H. A. et al, "Model based control of a robot manipulator", MIT Press, Cambridge, MA, USA, 1988; 2. Desoer, C. A. and Vidyasagar, M. "Feedback systems:

input-output properties", 3. Leigh, J. R.

theory", Academic Press, 4. Spong, M. W., control", John Wiley and Sons, 1989; 5. Spong, M. W. and Vidyasagar, M. "Robust linearcompen- sator design for nonlinear robotic control", IEEE Journal of robotics and automation, Vol. RA-3, No. 4, pp. 345 - 351, 1987; and 6. Vidyasagarr, M., "Nonlinear systems analysis", Presn- tice-Hall, 1978.

METHOD 2 may comprise the following steps:

Academic Press, 1975; "Functional analysis and 1980; and Vidyasagar, M., "Robot dynamics and linear control 27 a.

inputs:

I.

torque; and 2. Sensor 62 -- inertial accelerations and velocities I b. Using the output of step a,-compute the relative velocity of the payload 40 relative to frame 14 such that:

1. in the direction of motion of the human user, the center of mass of the payload 40 will remain, as far as possible, stationary relative to the human user.

2. in the horizontal direction perpendicular to the direction of motion of the human user, the center of mass of the payload 40 will remain, as far as possible, stationary.

C. Using the output of step b, generate commands for servo system 66 which controls payload stabilizer 68.

Reference is now made back to Fig. 6. A particular feature of foot-ground interface member 48 of Fig. 6 is that it includes a resilient element defining an area of contact with the ground along the longitudinal axis defined by the foot such that a zero-moment point between the resilient element and the ground shifts in accordance with motion of the human user. Specifically, interface 48 is typically operative to store energy when the human user's heel contacts the ground and to release the stored energy when the human user's toes push off the ground.

The resilient element is preferably configured to have varying stiffness such that the stiffness is relatively small during heel contact and relatively large during push off.

The foot-ground interface member may, for example, Receive inputs from the sensors measuring the following Sensor 52 -- frame/payload interaction force and 28 include a foot-contacting member, a ground-contacting member in hinged relationship with the foot-contacting member and a resilient member, such as a spring, which is attached at one end to the foot-contacting member and at another end to the groundcontacting member.

Preferably, the apparatus of Fig. 6 includes a plurality of resilient elements, which may include the following resilient elements:

a. A resilient element 70 (Fig. 21) associated with inter face 42, which is operative to absorb forces operating on the human's torso. Resilient element 70 is preferably configured to correspond to a relatively small portion of the human back such that a relatively large portion of the human back, and particu larly the shoulder portion thereof, is unimpeded and remains free to move. The resilience may be provided by springs and/or by a resilient material such as a resilient foam.

Resilient element 70 is preferably fixedly associated with belt 56. Belt 56 and resilient element 70 are preferably removably mountable on the human body.

b. A resilient element 72 associated with interface 46.

Resilient element 72 may generally resemble resilient element 70.

The resilience may be provided by springs and/or by a resilient material such as a resilient foam.

C. A resilient element 74 associated with interface 48. Resilient element 74 is illustrated in Fig. 19.

d. A resilient element 90 associated with interface 16 between the frame 14 and the payload 40 which is operative to 29 absorb interaction forces between the frame and the payload. The resilient element 90 is illustrated in Fig. 20.

In the illustrated embodiment, resilient element 74 is passive, however, it is appreciated that resilient element 74 may, alternatively, be active, i.e. it may be ultimately controlled by the CPU 50 of Fig. 8.

Segments 22 preferably are configured to be both light and structurally strong and may, for example, comprise hollow metal rods.

Payload stabilizing element 68, which is illustrated in detail in Fig. 20, is operative to stabilize the payload 40 relative to the human user so as to prevent the payload 40 and its contents from applying detrimental loading to the human user, even when the user carries out sideways movements or bending and straightening movements.

Typically, an electronics box 84 is provided in association with the payload 40 which includes electronic components of the apparatus such as CPU 50, inertial sensing system 62, the servo amplifiers of servo systems 64 and 66, and a power source 86 (Fig. 8).

Fig. 20 is a pictorial illustration payload stabilizer of payload 40, 68, resilient element 90 and frame 14. AS shown, the payload 40 comprises a container into which may be placed articles or loads which the human user wishes to carry from place to place. The payload 40 is mounted on payload stabilizer 68 with two degrees of freedom, as indicated by arrows 92 and 94.

Payload stabilizer 68 is fixedly mounted on the hori- zontal portion of frame 14. Stabilizer 68 includes two upstanding portions 95. Pivotably associated therewith, preferably driven by a motor 96, is a frame 97. Payload 40 is pivotably associated with frame 97 via two pivoting joints 98, wherein the pivoting motion is preferably driven by a motor 99. Motors 96 and 99 may each comprise, for example, a direct drive DC torque motor, commercially available from Inland Motor, Kollmorgent Corporation, Radford, VA, 24141.

Sensor 52 is mounted on frame 14. Intermediate sensor 52 and payload 40 is resilient element 90 which may comprise a spring.

Fig. 21 is a detailed side view illustration of the association between payload frame 14 and the belt 56 of Fig. 6.

Fig. 22 is a pictorial illustration of an exoskeleton apparatus constructed and operative in accordance with an alternative embodiment of the present invention. The apparatus of Fig. 22 is suitable for facilitating pedaling action by a human user. The apparatus of Fig. 22 may be generally similar to the appara- tus of Fig. 6. For example, elements 510, 522, 524 and 525of Fig. 22 may respectively be similar to elements 10, 22, 24 and 25 of Fig. 6. However, unlike connector 25 of Fig. 6, connector 525 of Fig. 22 may have only one degree of freedom, namely in the plane of rotation of the pedal 530.

Although payload 40 and the elements associated therewith are omitted in Fig. 22, it is appreciated that, alternatively, payload 40 and the elements associated therewith may be provided in conjunction with the pedaling facilitating apparatus 31 of Fig. 22.

Fig. 23 is a Pictorial illustration of connector 25. As shown, connector 25 includes a first hip joint 27 and a second hip joint 29. In the illustrated embodiment, each hip joint has one passive pivotal degree of freedom, as indicated by arrow 31, and two active pivotal degrees of freedom, as indicated by arrows 33 and 35. Motion in the direction indicated by arrow 33 is actuated by motor 37. Motion in the direction indicated by arrow 35 is actuated by motor 39. Position and velocity information in all three of the directions indicated by arrows 31, 33 and 35 is preferably utilized in step c of METHOD I above.

Control apparatus for an exoskeleton device suitable for facilitating pedaling action is now described with reference to Figs. 24-28.

The exoskeletal apparatus of the present invention is dominated by the movements of the human operator and is operative to bear the external loads carried by the operator. Alternatively, the apparatus may partially bear the weight of the operator himself.

Operatively, the human operator is responsible for the stability of the apparatus and initiates the desired directions of movement. The apparatus allows the operator any degree of spatial freedom required for management control of the apparatus, and intervenes only when the system approaches critical conditions, which may pose a threat to the stability of the system.

The conditions under which the apparatus operates may be divided into three categories, the first two defining the term 32 "critical": 1) a region of global instability, 2) a region wherein the apparatus is responsible for maintaining its global stability, 3) a region wherein the operator is responsible for maintaining global stability of the system.

The control systim of the present invention preferably performs the following tasks:

a) controlling the interaction between the operator and the apparatus and between the apparatus and the surroundings; b) allowing the operator to drive, control and stabilize the apparatus; C) ensuring global stability, including essential status variables which the operator has no control over; d) checking performance standards in conditions of uncertainty.

Referring now to the vectorial schematic illustration of Fig. 24, it should be noted that the control system of the present invention employs principles of impedance control. These principles were first published by Hogan N. in "Impedance Control: An Approach to Manipulation: Part III Applications", Journal of Dynamic Systems, Measurement and Control, vol. 107, March 1985, pp. 17-24.

The operator applies a force Fh to the apparatus. This force Fh is measured by a force sensor 500 which outputs a force signal Fh', which is then converted by the desired impedance model 502 into a desired velocity signal Vc, Vc is the velocity at the point where force was applied. The Jacobian determinant (J) of the system, which is the ideal approximated algorithm, converts the desired velocity signal to a generalized velocity 33 signal qc, which represents the velocity in a joint coordinate system. The generalized velocity signal qc is transmitted to the servo velocity loop 504, which outputs the real joint velocities q. By multiplying q by the inverse Jacobian, the real velocity at the point of operatorapparatus interaction is computed.

In order to simplify the operation of the system, let all elements of the schematic illustration adopt ideal values, where, Fh = Fh' q qc and J J,.

Fig. 25 illustrates this ideal situation. Theoretically, the system of Fig. 25 is capable of controlling the impedance, or interaction, between operator and apparatus, and to reach desired performance values. The higher the desired impedance value is, the higher the velocity values reached by the apparatus, using smaller forces, Ph. Thus, the control system allows the operator to have control over the velocity V of the apparatus, by applying force Ph, using the desired impedance model.

In an ideal situation, the higher the desired impedance value, the less operator-apparatus interaction exists, and as forces or moments develop between operator and apparatus, velocity V will increase and strive to minimize the interactive forces Ph. Thus an apparatus "dominated" by its operator is created. In the practical sense, it is not possible to achieve an ideal system, and interactive forces will develop, mostly due to the dynamic interaction between the apparatus and the 34 terrain. The control system is directed toward reducing these forces to a minimum.

Referring now to Fig. 26 the operation of the control system and its components will be examined. Fig. 26 illustrates the operation of the apparatus while attached to a fitness bicycle as opposed to a person. The global stability effect is thereby eliminated because the apparatus is supported by a stationary structure, and thus examination of the components is facilitated. The human operator remains in command and the loading of the apparatus takes place using the loading mechanism of the bicycle. The operation of the apparatus while attached to the fitness bicycle will now be described in detail.

The mechanical system preferably has th4ee degrees of freedom, including two active degrees, powered by controlled closed-loop servo activators, and one passive degree of freedom. The configuration of these degrees as illustrated by Fig. 26 reflects the following considerations:

1. Kinematic similarity to the bio-mechanical structure of the human leg. Considering kinematics alone, one degree of freedom would suffice for propelling the bicycle pedals. However, in order to increase the similarity of the bicycle structure to the structure of the human leg, which may be viewed as having three major degrees of freedom, i.e. thigh, knee and ankle, a configuration including three degrees of freedom was adopted for the experiment.

2. Suitability to walking experiments: A three degree configuration, kinematically adapted to the human leg, is also suitable for walking experiments where a fourth degree of freedom, representing transverse stability for walking, is not necessary.

The mechanical structure of the bicycle is indicated by reference numeral 598. Two force gages 600 and 602 are placed on either side of an arm 604, which represents the passive degree of freedom. Gage 600 measures the force generated by operatorapparatus interaction, and gage 602 measures the force generated by the apparatus and the bicycle pedal 606.

The domination of the apparatus by the operator is possible as a result of the output of these two gages, the signals of which represent the external forces at work on the apparatus. The bicycle pedal 606 is attached by means of a mechanical arm 608 to the pedal axis 610. It is at this point that the load moment, the direction of which is always opposite the direction of the rotation vector (indicated by arrow 612), is measured.

The control system includes two main loops. An internal velocity loop based on the Computed Torque Technique, and an external force loop based on the Impedance Control Method. This configuration enables domination of the apparatus even in the face of critical conditions.

The elements of the bicycle experiment are schematically illustrated in Fig. 27. The human operator represented by block 700 applies force, while riding the bicycle, to the foot supports of the apparatus. FhR is a vector of the force applied by the operator to the right foot support 702, and FhL is a vector of the force applied by the operator to 36 the left foot support 704. FhR and FhL are measured by means of two force gages whose outputs are FhR' and Fhl,', respectively. The forces applied by the human operator are applied to the servo activators of the apparatus as well. As such, the forces FhR and are converted by the Jacobians jTR and jTL to load moments the annaratus.

FhL ThR and Thl operating in the joints of Simultaneously, the interactive forces FPR and FPL between the pedal and foot support, of the right and left side respectively, are measured. The output signals of gages 706 and 708, respectively, are multiplied by force increase coefficients 710 and 712, respectively, and the results of the multiplication are used to compute the total forces EFR and EFL applied to the right and left foot supports, respectively.

EFR and EFL are converted, using the desired impedance model 714, to desired foot support velocities VRC and VLC. This conversion operation takes into account status measurements of the bicycle pedals as well, so as not to include ineffective velocity components such as components in a direction that does not contribute to the actual velocity.

The desired velocities VRC and VLC are converted by the Jacobian models J'R and J'L to desired velocity values, qRC and qLC, in the joint axes. These values are output to the servo loops 716 and 718, respectively, which produce the real joint velocities qR and qL, respectively. The outermost loop of the control system, illustrated in Fig. 27, is ultimately closed by the human operator.

The velocity loops will now be described with reference 37 to Fig. 28. The velocity loops have three major tasks:

1. velocity control (checking the desired velocities output by the impedance controller); 2. ensuring ruggedness parameters of the system; 3. decoupling control.

The control of the velocity loops is based on the Model Based Control algorithms, the most prominent of which are described by Chae et al in "Model Based Control of a Robot Manipulator", MIT Press, 1988. In the case of the bicycle experiment, a CTT (Computed Torque Technique) control method is most suitable. Based on knowledge of the real model of the controlled system, all linear and non-linear coupling elements of the model are eliminated, as a first step. Then the velocity loop is closed, based on the assumption that no coupling elements exist in the system. In effect, it is not possible to eliminate all coupling elements from the model, because the real model is unknown.

With reference to the schematic illustration of Fig. 28 the following parameters are defined:

n = number of degrees of freedom Fex = external force q e Rn = generalized coordinates M(q) e Rnxn = inertia matrix h e Rn coriolis, centripetal and gravity forces U e Rn control, with dimensions of generalized 1 in the face of uncertainties in the forces Fex E Rn = external forces vector 38 where:

j e R= the Jacobian matrix of the apparatus Ud e Rn interferences and N < represents the cartesian spatial dimension N < 6 M(q)q11 + h(q,q') = U + jTFex + Ud is an equation which serves as a model for the apparatus of the present invention, representing the motion of a body, having n degrees of freedom and interacting with an external force Fex, Appendix A is a Fortran listing of a software implementation of a portion of the apparatusof Fig. 22. Appendix A is a software simulation of an exoskeleton which is useful for facilitating stationary pedaling action.

Biomechanical considerations useful in designing an exoskeleton device for facilitating the actions of walking, running and carrying a payload are now described.

Generally, gaits of the man-machine system may be divided into 'the three following phases: double support, when both legs are in contact with the ground; single support, when one leg is supporting and the other does not; and flight, when there is no contact with the floor.

Phase detection may be based on the outputs of the force sensors installed in the machine-ground interfaces (22).

During double-support, the control is responsible for decreasing the error between the reference (desired) and measured 39 dynamic forces at the interfaces of the machine with the man, ground and payload. Dynamic servitude of the machine to the human operator may be maintained, therefore, by dictating a velocity command to the actuators of the machine in a manner which decreases the error between the reference and measured dynamic forces at the machine's interfaces.

During the Single-Support and Flight, the control operates in a similar way to the operation during Double-Support; except is in the setting of reference levels for desired interface dynamic-force values: these phases (namely-Single Support and Flight) differ from the Double support by having one or two legs which are not in contact with the ground. The force applied, and thus measured, at interface (22) of a swinging leg (swinging = not in contact with ground) is zero or very close to zero. In order to maintain the desired correspondence between machine's and operator's swinging leg kinematics, the reference (desired) force value at interface (182) of a swinging leg is set to zero.

In the case of Single-Support, therefore, the setting of reference levels at interfaces (22) and (182) of the swinging leg are set to zero and those of reference levels at interfaces (22), (182) of the supporting leg, as well as references at interfaces (181) and (20), are set to the values of desired dynamic forces which are calculated in the manner described before.

During the Flight phase of gait, all reference levels are set to zero, since during Flight each one of the compounds of the system is moving in space as a free body. When the human operator moves his body segments with respect to one another, the machine follows his motion by actuating the actuators so as to bring all interface forces to their zero reference values.

The control logic described above typically does not relay on the sequencing of gait phases, because it depends only on phase detection, i. e. - on detecting whether each leg is in contact with the ground, or not and not on sequencing, i.e. - the order of the appearance of phases. This is regarded as an advantage because it enables the controller to successfully control the system in situations which may not be classified as "normal gait". For example, the same control scheme deals successfully with the task of transition from standing to kneeling, and vise versa from kneeling to standing. When the operator starts to flex his knees and ankles so as to vertically lower his trunk, the forces at the man-machine interface, in the trunk-pelvis region (interface 181) increase above their reference value, and the difference between the forces measured at interface 22 and interface 182 also increases above its reference level; the control logic immediately dictates velocity commands to the machine's actuators so as to decrease these errors, resulting in kneeling action of the machine which fully traces the kneeling action of the human operator.

It is appreciated that features shown and described herein in the context of separate embodiments may be provided in any suitable combination and that, conversely, each embodiment shown and described herein which includes multiple features may alternatively be provided with only one of or only some of the 41 multiple features.

It should be appreciated by persons skilled in the art that the present invention is not limited to the preferred embodiments.

Claims (25)

1. A foot-ground interface member comprising: a resilient element defining an area of contact with the ground along the longitudinal axis defined by the foot such that a zero-moment pc;int between the resilient element and the ground shifts in accordance with motion of the human user.
2. 2. A foot-ground interface membe according to claim 1 which is operative to store energy when the human user's heel contacts the ground and to release the stored energy when the human user's toes push off the ground.
3. An exoskeleton control method comprising the steps of: receiving at least one load input; employing an imped@Lnce control technique to compute a desired velocity for at leastl a portion of the exoskeleton; and computing, for each of a plurality of segments in the exoskeleton, an angular velocity for each segment relative to another segment adjacent thereto.
4. 4. A method according to claim 3 wherein said step of computing comprises the step of employing an inversekinematics -- the plurality of angular velociti technique to comput- Les.
5. 5. A method according to claim 4 wherein sa4,J step of computing comprises the step of employing an inverse dynamics 43 technique to compute the plurality of angular velocities.
6. 6. A foot-grIund interface member comprising: a resilient element which is configured to have varying stiffness such that the stifffness is relatively small during heel contact I and relatively large during push off.-
7. 7.

   A foot-g=und interface membep comprising Ecot-contacting member; groundcontacting member in hinged relationship with the foot-contacting member; and a resilient member attached at one end to the foot-contacting member and at another end to the groundcontacting member.
8. 8. Apparatus according to cla?im 7 wherein the resilient member comprises a spring.
9. 9. steps of:

   input; computing a relative velocity for the payload which maintains the center of mass of the payload generally stationary relative to a bearer of the payload; and generating a payload stabilizing command according to the result of th-, _: computing step.

   A payload stabilizing control method comprising the receiving at least a load input and an inertial 44
10. 10. An exoske-etal apparatus to help a user carry a payload while moving across the ground, comprising:

    a right leg member attached to the right foot of the user so as to transmit loads in said right leg member direct ly to the ground; a left leg member attached to the leJ--t foot of the user so as to transmit loads in said left leg member directly to the ground; a right foot member connected to said right leg member; a left foot member connected to said left leg member; apparatus for storing and connected to said right foot member; apparatus for storing and connected to said left foot member; and a payload/user link connected to the payload and further connected to said right leg member and to said left leg member, said payload/user link located so that loads exerted by the user on paid payload/user link act to guide the movement of the apparatus.

    recovering energy recovering energy
11. 11. An exoskeletal apparatus to help a user carry a payload while moving across the ground, comprising:

    -o the right f oot of a r-4ght leg member attached t --ansmit loads in said right leg member direct- the user so as to tly to the ground, said right leg member including a him assembly; a left leg member attached to the left foot of the user so as to t=ansmit loads in said left leg member directly to the ground, said right leg member including a hip assembly; a rright foot member connected to said right leg member; a left foot member connected to said left leg member; and a payload/user link connected to the payload and further connected to said right leg member and to said left leg member, said payload/user link located so that loads exerted by the user on said payload/user link act to guide the movement of the apparatus.
12. 12. An apparatus as in claims 10 or 11 wherein said payload/user link is located in the vicinity of the back of the user.
13. 13. An apparatus as in any of claims 10-12 wherein said payload/user link is substantially ring shaped and surrounds the user substantially at the midsection area.
14. 14. An apparatus as in any of claims 10-13 wherein said right leg member and said left leg member each includes at least two segments, said segments being connected to each other through a joint.
15. 15. An apparatus as in any of claims 10-14 wherein said payload/user link includes apparatus for at least partially 46 supporting the weight of the user.
16. 16. An apparattus as in any of claims 10-15 further compris ing:

    a sensor for sensing loads between each of said foot members and the respective foot of the user; a sensor for sensing loads between each of said foot members and the ground; a sensor for sensing loads between the user and said payload/user link; said actuators an actuator for said right leg member; an actuator f or said lef t leg member; a control system connected to said sensors and to for controlling the movement of the apparatus.
17. 17. An apparatus as in claim 16 wherein said control system is an impedance control system.
18. 18. An apparatus as in any of claims 10-17 further compris ing energy storage apparatus for alternately storing energy from the user and/or the apparatus and/or the payload and returning at least a portion of said energy to the user and/or the apparatus and/or the payload.
19. 19. An appara'-as as in claim 18 wherein said energy storage apparatus are connected to said foot members.

    47
20. 20. An appara-tus as in claim 18 wherein said energy storage apparatus includes a spring mechanism.
21. 21. An apparatus as in claim 18 wherein said energy storage apparatus includes a hydraulic mechanism.
22. 22. An apparatus as in claim 18 wherein said energy storage apparatus includes a pneumatic mechanism.
23. 23. An apparatus as in any of claims 11-22 wherein said hip assembly includes apparatus for stabilizing the apparatus when only one of the user's feet is in contact with the ground.
24. 24. An apparatus as in any of claim 23 wherein said apparatus f or stabilizing the apparatus includes an inertial stabilization system.
25. 25. An apparatus as in any of claims 10-24 further comprising an ankle joint.

    48