Method for generating trajectory of exoskeleton and method for setting exoskeleton in motion

Abstract

To provide a method of generating an orbit of an exoskeleton and a method of moving the exoskeleton.SOLUTION: The present invention relates to a method for generating a trajectory of an exoskeleton 1 including two legs each having a foot. This method comprises execution by data-processing means 11a of a server 10a, of steps of: (a) acquiring at least one n-tuple of gait parameters defining a given gait of the exoskeleton 1; (b) generating at least one periodic elementary trajectory of the exoskeleton 1 for the n-tuple of gait parameters, such that the periodic elementary trajectory includes in sequence a first trajectory portion and a second trajectory portion, such that in the first trajectory portion each foot performs a pure rotation, and in the second portion only one foot performs a translation.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to the field of exoskeleton robotics.

[0002] More precisely, it relates to a method for generating a trajectory for an exoskeleton and a method for moving the exoskeleton. [Background technology]

[0003] In recent years, walking-assist devices called exoskeletons have emerged for people with mobility impairments, such as paraplegia. They are external robotic devices "worn" by an operator (human user) via a system of fasteners that connects the exoskeleton's movements to the operator's own. To replicate walking, lower-limb exoskeletons have several joints, typically at least the knees and hips. Actuators allow these joints to move, moving the operator. An interface system allows the operator to give commands to the exoskeleton. A command system translates these commands into commands for the actuators. Further sensors complement the device.

[0004] These exoskeletons are an advancement over wheelchairs because they allow the operator to stand and walk. The exoskeleton is no longer limited by wheels and can theoretically move around in most uneven environments. However, unlike legs, wheels cannot traverse significant obstacles such as steps, stairs, or obstacles of excessive height.

[0005] However, none of these exoskeletons provide anthropomorphic, unassisted, autonomous human walking, i.e., walking that is stable and viable on a variety of terrains.

[0006] In most cases, these limitations are realized by the device's inability to manage balance or the direction of walking itself, two tasks that are usually transferred to the operator, who performs them via crutches, as proposed, for example, in Rewalk patent US7153242 or Ekso-Bionics application US2016038371.

[0007] Rex-Bionics patent EP2231096 describes the only exoskeleton that can be used without external assistance for people who are unable to ensure their own stability. The control principle described in paragraph 0122 clearly explains the need to move the center of pressure (the physical point on the system where the moment of the reaction force applied by the ground is zero) from one part of the support polygon (the convex envelope of the contact points with the ground) to another part of the support polygon.

[0008] This limitation forces extremely slow walking (a few meters per minute; normal walking is 33 meters per minute and over 2 kilometers per hour) with short steps (less than 30 cm, compared to a normal stride length of 50 to 80 cm) in which the feet are in constant flat contact with the ground. The types of environments that can be accessed are therefore limited, as uneven terrain is virtually excluded. Similarly, the slightest obstacle, such as a small object like a stone, can upset the system's balance if the foot is placed on it at any given moment, ultimately creating a risk of the system tipping over.

[0009] In contrast, "natural" human gait is characterized by a series of phases in which the foot is in the process of being flat on the ground, flat in the air, or rolling on the ground, as shown in Figure 1. The ability to roll the foot is essential to walking because it allows the foot to take larger strides and provides stability on a variety of terrain.

[0010] However, the so-called first generation exoskeletons mentioned above do not have actuating feet and do not keep the supporting feet on the ground.

[0011] Executing this roll is in fact complicated for a bipedal humanoid robot or robotic device: even if an interrupted foot structure such as that proposed in application WO2015140353 is provided, once the center of pressure reaches the limit of the support polygon, the system begins to roll around this point and is therefore no longer in static equilibrium.

[0012] In the case of walking, foot roll causes the supporting foot to partially lose contact with the ground, which has several consequences.

[0013] · The support polygon (equilibrium surface) is potentially reduced to a point making it difficult, if not impossible, to maintain a center of pressure within the support polygon; The system is underactuated, i.e. it can no longer act over all of its degrees of freedom. All movement is no longer possible.

[0014] In such situations, the traditional form of shuffling gait (flat foot walking) as described in the following documents no longer works effectively: Kajita S., KF (2003). Biped Walking pattern generation by using preview control of Zero-Moment Point. ICRA (pp. 1620-1626), or Principle described in Rex-Bionics EP2231096.

[0015] A natural way to think about this is to move the swinging leg forward and bring the second foot on the ground back into balance on the support polygon. During this time, the system is free to rotate around the support foot and is in the process of "tipping" somewhat. This is called dynamic walking because the body goes through a series of unstable postures, but only temporarily (if it "stopped", the person would fall mid-walk).

[0016] In this dynamic walking approach, it is complicated to move the swinging legs quickly and assume a posture that re-establishes balance, at least in a short time. Indeed, if the legs are designed to follow a set trajectory within a pre-calculated time, there is a risk that the legs will strike the ground too early or too late due to the uncontrollable behavior of a malfunctioning system in the event of a slight disturbance (it is impossible to correct a slight deviation from the planned trajectory). This can cause discomfort to the operator, loss of balance, and even a fall, even on simple terrain.

[0017] For this reason, all first generation exoskeletons (and many humanoid robots) try to avoid this type of situation by keeping the supporting foot flat, resulting in the aforementioned limitations on walking speed, step length, types of terrain tolerated, and general stability of the gait.

[0018] As a result, a new walking paradigm for exoskeletons is proposed in application WO2018130784, combining the principles of "virtual constraints" and "hybrid zero dynamics" (HZD) to enable fast and natural walking, without the risk of falls or imbalance, even on difficult and unpredictable terrain.

[0019] Traditionally, the trajectory, i.e., the change in each degree of freedom, is expressed as a function of time. The "dynamics" of a system is a function of:

number

number

number

[0020] Conversely, the HZD is the dynamics of the non-actuated degrees of freedom. This dynamics is called "zero" because the command corresponds to the inability / unwillingness to act, i.e., the command is 0. It is also called "hybrid" because the impact of the foot on the ground produces a discrete instantaneous phase that intersects with the continuous phase.

[0021] In the so-called "virtual constraint" method, the principle is to define a trajectory that is not time-dependent but is modified by a parameter that directly follows the configuration, in order to select the degrees of freedom that are modified by the parameter. This parameter is called a phase variable. An example of such a phase variable is the angle between the heel / hip axis and the vertical direction, which constitutes the non-actuated degrees of freedom mentioned above.

[0022] The phase variable allows to define the "progression" of the steps. More precisely, at each step, the phase variable switches successively from an initial value to a final value before it is assigned to the initial value again, which is the starting point of the next step. To simplify matters, the values ​​of the phase parameter may be normalized between 0 and 1.

[0023] Each value of the change parameter corresponds to a value of an actuation degree of freedom that the system must be forced to obey; it is these relationships (one for each actuation degree of freedom that is controlled in this way) that are called virtual constraints.

[0024] If the system follows this trajectory exactly in the degrees of freedom upon which it is possible or desired to act, in other words, if virtual constraints are obeyed in this degree of freedom, then the changes in the system are entirely determined by the changes in the non-actuated degrees of freedom, which follow their own dynamics, the HZD.

[0025] Therefore, a good choice of virtual constraints can make this dynamics contain attractive periodic "orbits", i.e. stable orbits to which the system is naturally attracted.

[0026] Although this method HZD gives great satisfaction, the difficulty lies in generating the trajectory (this also applies to the "sliding" method). In fact, it is observed that the trajectories obtained have "almost no roll" of the foot, i.e. the foot remains substantially horizontal (heel and toes hardly lift off the ground), contrary to the natural human gait shown in Figure 1 above, where a roll is shown.

[0027] However, although in theory algorithms for generating trajectories fully allow for roll phases as pronounced as in natural human gait, they generally are based on methods for optimizing non-convex nonlinear problems under constraints, and therefore favor "optimal" trajectories that are considered more stable to the adverse effects of more "anthropomorphic" trajectories, which are however primarily preferred by the human operator of the exoskeleton. Summary of the Invention

[0028] It is therefore desirable to have a new solution for generating orbits that increase the natural side of the orbit without compromising the stability of the orbit.

[0029] Thus, according to a first aspect, the invention relates to a method for generating a trajectory for an exoskeleton comprising two legs, each having a foot, comprising the following steps, performed by the data processing means of a server: (a) obtaining at least one n-tuple of gait parameters defining a given gait of the exoskeleton; (b) generating at least one periodic elementary trajectory of the exoskeleton for each set of n gait parameters, such that the periodic elementary trajectory includes a sequence of first and second trajectory segments, wherein each foot performs a pure rotation during the first trajectory segment and only one foot performs a translation during the second trajectory segment; The present invention is provided to perform the following.

[0030] According to advantageous and non-limiting characteristics: The periodic base orbit cyclically repeats a sequence of a first orbital portion and then a second orbital portion.

[0031] During the second trajectory segment, the foot that performs the translation is the trailing foot first, and the leading foot first performs the pure rotation.

[0032] During the second track segment, the initially leading foot remains stationary.

[0033] At the end of the first track segment, the leading foot is flat on the ground.

[0034] Step (b) is performed using at least one neural network.

[0035] The exoskeleton accepts a human operator, and step (a) involves determining an n-tuple sequence of gait parameters for the exoskeleton desired by the operator.

[0036] The generated trajectory of the exoskeleton includes a new periodic elementary trajectory and a transition to a new periodic elementary trajectory for each n-tuple of the sequence.

[0037] According to a second aspect, the present invention relates to a method for moving an exoskeleton having multiple degrees of freedom, at least one of which is actuated by an actuator controlled by data processing means; the method comprising step (c) of executing, by the data processing means of the exoskeleton, a trajectory of the exoskeleton generated by the method according to the first aspect, thereby causing the exoskeleton to walk.

[0038] According to a third aspect, the present invention relates to a system comprising a first server and an exoskeleton, each of the first server and the exoskeleton including data processing means; the data processing means being configured to execute the method for generating a trajectory for an exoskeleton according to the first aspect and / or the method for moving an exoskeleton according to the second aspect.

[0039] According to fourth and fifth aspects, the present invention relates to a computer program product comprising code instructions for performing the method for generating a trajectory for an exoskeleton according to the first aspect and / or the method for moving an exoskeleton according to the second aspect when the program is executed on a computer, and also to a storage means readable by a computing device having recorded thereon a computer program product, the computer program product comprising code instructions for performing the method for generating a trajectory for an exoskeleton according to the first aspect and / or the method for moving an exoskeleton according to the second aspect.

[0040] Other features and advantages of the invention will appear on reading the following description of preferred embodiments, which description is given with reference to the accompanying drawings, in which: [Brief explanation of the drawings]

[0041] [Figure 1] Shows a person walking. [Figure 2] FIG. 1 is a diagram of an exoskeleton used in accordance with the method of the present invention. [Figure 3] 1 is a diagram of an architecture for implementing the method of the present invention; [Figure 4] FIG. 10 illustrates the structure of an optimization problem describing all phases of the walking trajectory with foot roll of the exoskeleton in an exemplary embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0042] Architecture According to two complementary aspects of the present invention, it is proposed that: How to generate a trajectory for an exoskeleton. A method for making the exoskeleton 1 move (applying the trajectory generated by the method of the first aspect).

[0043] Referring to Figure 2, the exoskeleton 1 described above is an articulated mechanical system of the bipedal robotic type, actuated and controlled, equipped with two legs, and more precisely accommodating a human operator, whose lower limbs are firmly attached (particularly by means of straps) to the legs of the exoskeleton 1, which may roughly be a humanoid robot. Here, the term "walking" means moving the robotic device 1, which actually creates alternative support for the legs in a standing position in a way that creates displacement.

[0044] The exoskeleton 1 has multiple degrees of freedom, i.e., deformable joints (generally via rotation), i.e., joints that can move relative to each other. Each of the multiple degrees of freedom can be either "actuated" or "non-actuated".

[0045] An actuated degree of freedom denotes a joint with an actuator controlled by the data processing means 11c, i.e., this degree of freedom is controlled and can move based on it. Conversely, a non-actuated degree of freedom denotes a joint without an actuator, i.e., this degree of freedom follows its own dynamics and the data processing means 11 does not exercise direct control over it (but a priori exercises indirect control via other actuated degrees of freedom). In the example of Figure 1, the contact of the heel with the ground is precise in time, and therefore the exoskeleton 1 is free to rotate relative to this contact point. The angle between the heel / hip axis and the vertical then constitutes a non-actuated degree of freedom.

[0046] The exoskeleton will of course have at least one actuated degree of freedom, preferably multiple actuated degrees of freedom, and at least one non-actuated degree of freedom, i.e., a degree of "mis-actuation" as previously mentioned. The number of non-actuated degrees of freedom is called the mis-actuation degree.

[0047] The data processing means 11c designates a computing device (typically a processor, or preferably an external processor if the exoskeleton 1 is "remotely controlled", but preferably embedded in the exoskeleton 1, see below) configured to process instructions and generate intended commands for the various actuators, the latter of which may be electrical, hydraulic etc.

[0048] The present application is not limited to any architecture of the exoskeleton 1. Examples will be taken from those described in applications WO2015140352 and WO2015140353.

[0049] Therefore, preferably, in accordance with the description of these applications, the exoskeleton 1 comprises a foot structure for each leg, which includes a support surface against which the foot of the leg of the person wearing the exoskeleton rests.

[0050] The support plane comprises a front platform and a rear platform, with pivot connections in the feet connecting the front platform to the rear platform and providing a non-actuated degree of freedom.

[0051] However, one skilled in the art will know how to adapt the method to any other mechanical architecture.

[0052] According to a preferred embodiment, the inventive method for generating trajectories and gaits may include a first and a further second server 10a, 10b in a structure as shown in FIG.

[0053] The first server 10a is a trajectory generating server and the second server 10b is a possible learning server.

[0054] In fact, the generation of the trajectory of the exoskeleton 1 may use a neural network, in particular of the "feedforward" type (FNN "feedforward neural network") as proposed in application FR1910649. The second server 10b is then a server for implementing the method for learning the parameters of said neural network. It should be noted that this method is not limited to the use of a neural network, but may also use any known technique for generating trajectories in its entirety, and moreover.

[0055] In any case, it is quite possible that these two servers are integrated (CONFONDUS), but in practice the second server 10b is most often a remote server, while the first server 10a is able to operate embedded in the exoskeleton 1 in real time, as shown in Figure 2. According to a preferred embodiment, the first server 10a implements a method for generating a trajectory for the exoskeleton 1 by means of a neural network that uses parameters retrieved from the second server 10b, and the exoskeleton 1 then directly applies said trajectory generated in situ to make itself move.

[0056] Each of these servers 10a, 10b is typically a single computer connected to an extended network 20, such as the Internet, for data exchange, although in practice communication can be at least temporarily interrupted once the neural network has been trained and embedded in the second server 10b. Each server includes processor-type data processing means 11a, 11b (particularly the data processing means 11b of the second server has strong computing power, since training is long and complex relative to the simple use of the trained neural network) and, if applicable, data storage means 12a, 12b, such as computer memory, e.g., a hard drive. When generating trajectories using the neural network, the memory 12b of the second server 10b may store the training database.

[0057] It is understood that there may be multiple exoskeletons 1 each embedded with a first server 10a (which may be of limited power and size and generate trajectories only for the dedicated exoskeleton 1), or there may be multiple exoskeletons 1 each connected to a more powerful first server 10a and capable of integrating with a second server 10b (and having the ability to generate trajectories in flight (VOLEE, air passage) for all exoskeletons 1).

[0058] Principles of the Invention As explained herein, the term "trajectory" of an exoskeleton conventionally refers to the change in each degree of freedom (especially the actuation degree of freedom) expressed as a function of variable time or phase.

[0059] Furthermore, it is known to define a "complex" trajectory as a sequence of periodic trajectories, called "elementary," intersecting with transitions. The term "periodic trajectory" refers to any trajectory that applies (can be repeatedly applied) over the duration of a step (PAS, walking step, step) starting from the initial state of the exoskeleton 1 (the moment of foot contact) at the beginning of the step and returning to the same state at the beginning of the next step (this encompasses any walking on a flat surface, but also on ramps, stairs, etc.). Periodic trajectories are also said to form "limit cycles." Therefore, the aforementioned periodic trajectory can be applied stably over any number of steps.

[0060] In other words, each basic trajectory is associated with a given gait of the exoskeleton 1 (defined by an n-tuple of gait parameters) and allows maintaining this gait in a stable and feasible way (i.e., minimizing the cost function as much as possible, as will be understood, obeying all the constraints of the optimization problem). As explained, the gait parameters correspond to gait "features", such as step length (stride), step frequency, upper body (BUSTE, chest-up) inclination, as well as step height in the case of staircase intersections, instantaneous rotation angles for curved movements, and even morphological characteristics of the operator (a subgroup of gait parameters called patient parameters), such as hip, weight, femur or tibia length, center of gravity position (value of forward offset), lateral clearance of the upper body in the framework of rehabilitation activities, etc.

[0061] The gait "constraints" described above can vary and may depend on the type of gait desired, e.g., a "pied-plat" gait or a "HZD." The method is not limited to any type of desired gait.

[0062] A transition corresponds to a change in gait, i.e., a change in the value of a gait parameter (e.g., an increase in step length). That is, it is about knowing the initial set of gait parameters and the final set of gait parameters, and therefore the initial periodic trajectory (associated with the initial set of gait parameters) and the final periodic trajectory (associated with the final set of gait parameters). The aforementioned transitions are trajectory fragments, allowing to switch from the initial periodic trajectory to the final trajectory. It should be noted that there is always an "initial" or a "final" transition, as shown below.

[0063] The underlying idea of ​​this method is to note that it is possible to generate two sub-sections of a periodic basic trajectory in a more clearly anthropomorphic way. In the first section of the trajectory, called the "roll," both feet are in contact with the ground, each performing a pure rotational movement. That is, the rear foot (at least the "tip," typically in contact with the front platform of the foot structure, advantageously initially in full contact, i.e., flat) gradually lifts its heel. The front foot (contacting only the heel) gradually assumes a position at the tip of the trajectory (the first two drawings in FIG. 1). Then, in the second section, called the "swing" or simply the "step," the front foot remains in contact with the ground, while the other foot takes off (DECOLLECTION) (this is called "foot clearance") and performs a translational movement (which may, of course, include various rotations). In other words, the rear foot first breaks contact with the ground, moves forward, and resumes contact with the ground with its heel (the last three drawings in FIG. 1). At this stage, the two feet are in contact with the ground opposite to what they were at the beginning of the first part (the rear foot becomes the front foot), but with the same shape (the rear foot is in contact at the tip, the rear foot is more flat, the front foot is in contact only at the heel), i.e., it is possible to take a step and repeat the first and second parts, etc., in a symmetrical manner (giving the "periodic elementary" nature of the trajectory).

[0064] It should be noted that the concepts of "front" and "back" are defined with respect to the direction of walking, i.e. there is always a front foot and a back foot, and these feet reverse position with each step.

[0065] Each periodic basic orbit therefore consists of a first part and a second part, and the sequence of the first and second parts forms a cycle that can be repeated (by alternating between left and right feet in each cycle).

[0066] The existence of these first and second parts can be caused by adding conditions during the generation of the orbit and by applying a continuity of the orbital parameters from one part to the other. These conditions are advantageously:

[0067] · For the first part, there is a pure rotation of each foot (uniform rotation around a single axis of rotation), preferably with the leading foot flat (sliding) at the end of the first part. The second part is a clearance and translation of the (first) trailing foot only, until it touches the ground again (becoming the new leading foot, hence the word "first"). During this time, there is a pure rotation of the (first) leading foot. Furthermore, the leading foot does not move.

[0068] It should be noted that this division into two parts is not what humans actually do, and that in fact two parts are indeed present in human gait, but are less clear (part of the forelimb rotation is accompanied by translation). However, this "exaggeration" actually makes it possible to make the exoskeleton 1 walk in a much more natural way. Furthermore, although these conditions may seem very restrictive, it is observed that in real conditions the stability does not change at all and the generation of trajectories does not become any more complicated.

[0069] As can be seen, it is also possible to define these first and second portions within the transitions, in particular within the first and last transitions.

[0070] First and second orbital segments Figure 4 shows a preferred example of an initial transition, then a step, and then a final transition. Starting with the first part (roll phase), the step, i.e., the periodic basic trajectory, shall be described first. In the example of Figure 4, the left foot (dotted line, right foot solid line) is the front foot. Those skilled in the art may naturally transpose this example to be the case in the figure.

[0071] This is therefore a case of a "cyclic roll" which defines the first part of a periodic "cyclic" trajectory. The movement of each foot is defined in three views, with three speeds by three axes and three rotations by three axes. It can be seen that the following is produced:

[0072] For the right leg (hind leg), continuity and v x =v y =v z =Ωr=Ωy=0, and simply Ωx≠0 to allow pure rotation. For the left foot (front foot), continuity and x =v y =v z = Ωr = Ωy = 0, and simply Ωx ≠ 0 to allow pure rotation, with the output position p = 0, i.e., the sliding ("zero" position). Each foot experiences a non-zero normal force during locomotion that represents the reaction of the ground on the foot.

[0073] Once the front foot reaches the output position, control passes to the "cyclic step" case of the second part of the periodic trajectory. It can be seen that the following is caused to occur:

[0074] For the left (front) leg, continuity, v=0, i.e. pure rotation. - Right foot (hind foot): continuity and foot clearance. The right foot is in the air with no force being applied to it.

[0075] This ends with a "CYCLICITE": the complete configuration of the exoskeleton 1 at the end of this second part must be symmetrically equivalent to the configuration at the start of the cyclic roll (first part). The cyclic roll is started again by reversing the left foot (LF) and right foot (RF).

[0076] In the case of Figure 4, let us assume that we start with an initial transition from a given starting position (immobile, standing, etc.). As explained here, we can again have a first and second portion of this transition by analogy with the first and second portions of the periodic basic trajectory. Here, we assume that we start with a "left foot behind" starting roll, but this is an arbitrary choice and the opposite is also possible. In reality, the two feet are approximately side-by-side. As mentioned above, a pure rotation of the feet is generated, but in reality, it is even possible to keep the feet stationary when flat, with only the pelvis moving. The idea is to generate the dynamics of the movement.

[0077] Then, linked to the second part of this initial transition (the initiation step), the right foot remains stationary (or at least performs only a pure rotation) and the left foot remains in a take-off position until it reaches the ground again, and the cyclical orbit, or more precisely the first part (the cyclical roll), can begin.

[0078] Similarly, Figure 4 shows the case of the final transition with a periodic first orbital section. By symmetry with the initial transition, it starts with the second section (stop step) and ends with the first section (stop roll, i.e., end roll). It should be noted that the stop step section is identical to the second section of the periodic orbit with a fixed left foot and a lifting right foot (although this is also optional, the complete reverse may be performed). The only difference is that in this type of transition, the calculated movement generally shows that the rear foot only moves to the level of the front foot, rather than ahead of it in walking. In the stop roll section, the foot can again rotate, and the exoskeleton 1 remains fixed until it reaches a final posture (e.g., a fixed posture, standing, etc.).

[0079] method According to a first aspect, a method for generating a trajectory of an exoskeleton 1 is proposed, implemented by the data processing means 11a of the server 10a. The said method for generating a trajectory of the exoskeleton 1 starts with step (a) of obtaining at least one n-tuple of gait parameters defining a given gait of the exoskeleton 1. It also starts with step (a) of obtaining further n-tuple sequences of gait parameters progressively (e.g. for new commands from the operator of the exoskeleton).

[0080] In main step (method step) (b), the method comprises generating at least one periodic elementary trajectory of the exoskeleton 1 for an n-tuple of gait parameters, where the periodic elementary trajectory comprises a sequence of first and second trajectory segments, where in the first trajectory segment each foot performs a pure rotation and in the second trajectory segment only one foot performs a translation (foot clearance, the other is stationary but performs a pure rotation), as described herein.

[0081] For each new n-tuple of parameters in the sequence of n-tuples, a new periodic orbit is determined and a transition to this new periodic orbit is determined.

[0082] In this case, the method of generating the trajectory advantageously comprises the determination of n sets of gait parameters of the exoskeleton 1, i.e. the repetition of step (a) (repeated regularly, if applicable).

[0083] In fact, the exoskeleton 1 is an exoskeleton that accepts a human operator, but it is the human operator's posture (and possibly button presses) that determine the aforementioned parameters (unlike in the case of a normal robot, which can directly accept a start request, including setpoints for walking speed and / or direction).

[0084] For this purpose, the operator may be provided with a sensor jacket 15 that allows them to detect their upper body configuration (upper body orientation), as described herein. The direction in which the operator turns their upper body is the direction in which they want to walk, and the speed is given by the intensity with which the operator positions their upper body forward (how much they lean). The start request may correspond to the operator pressing a button (or a specific posture) signaling their intention to start walking, and thus to instructing the data processing means to determine the aforementioned parameters. Certain parameters, such as the instantaneous turning angle or step height in the case of a staircase meeting, may be predetermined or acquired by other sensors 13, 14.

[0085] The generation of the trajectory itself is not limited to any known technique, the purpose of the present invention is only to apply the above mentioned constraints during generation in a way that obtains the first and second parts.

[0086] The optimization tools described here are particularly known and are in particular capable of generating a given trajectory according to selected walking constraints and parameters. For example, in the case of HZD trajectories, the trajectory generation problem is preferably formulated in the form of an optimal control problem that can be solved by the so-called direct collocation algorithm. Reference is made to the following documents: Omar Harib et al., Feedback Control of an Exoskeleton for Paraplegics Toward Robustly Stable Hands-free Dynamic Walking.

[0087] As also explained, it is also possible to use a neural network trained on a training trajectory database.

[0088] It should be noted that it is possible to consider using a first neural network to generate a first portion of the trajectory and a second neural network to generate a second portion of the trajectory, and thus it is sufficient to train the first network on a training database of the first portion of the trajectory and the second network on a training database of the second portion of the trajectory.

[0089] According to a second aspect, a method for causing an exoskeleton 1 to move (exercise) is proposed, which comprises implementing the aforementioned method according to the first aspect (steps (a), (b)) to generate a trajectory for the exoskeleton, and then (noted as step (c)) executing the aforementioned trajectory so that the exoskeleton 1 walks.

[0090] Steps (b) and (c) may be repeated in a manner that constantly corrects the trajectory of the exoskeleton 1 in real time.

[0091] Equipment and Systems According to a third aspect, the invention relates to a system for implementing the method according to the first aspect and / or the second aspect.

[0092] As will be described, the system comprises a potentially integrated first server 10a, a possible second server 10b, and an exoskeleton 1.

[0093] The first server 10a includes data processing means 11a for implementing the method of the first aspect.

[0094] The exoskeleton 1 comprises data processing means 11c configured to implement the method according to the second aspect and, if necessary, data storage means 12 (in particular of the first server 10a), inertial measurement means 14 (inertial measurement unit), means for detecting foot impacts on the ground 13 (contact sensors or possibly pressure sensors), and / or a sensor jacket 15.

[0095] It has a number of degrees of freedom, at least one of which is actuated by an actuator controlled by the data processing means 11c in the framework of implementing the method according to the third aspect.

[0096] computer program products According to a fourth and fifth aspect, the present invention relates to a computer program product comprising code instructions for executing (by processing means 11a, 11c) the method for generating a trajectory for an exoskeleton 1 according to the first aspect and / or the method for moving an exoskeleton 1 according to the second aspect. The present invention also relates to storage means readable by a computing device in which the computer program product can be found.

Claims

1. A method for generating a generated trajectory of an exoskeleton comprising two legs, each having a foot, comprising the following steps performed by a data processing means of a server: (a) obtaining at least one n-tuple of gait parameters defining a given gait of the exoskeleton; (b) generating at least one periodic elementary trajectory for the exoskeleton while applying a condition to the n-tuple of gait parameters; and The condition is: the periodic basic orbit includes a sequence of first and second orbital segments, during which each foot performs a pure rotation and during which only one foot performs a translation, the translation beginning with the translating foot contacting the ground only at its toe and ending with the translating foot contacting the ground only at its heel; at the end of the first trajectory segment, a leading foot of the exoskeleton that is located forward of a trailing foot of the exoskeleton at the start of at least one of the periodic basic trajectories is flat on the ground; and during the second trajectory portion, the foot that performs the translation is the trailing-first foot and the leading-first foot remains stationary.

2. The method of claim 1 , wherein the condition is that the periodic base orbit cyclically repeats the sequence of the first orbital portion and then the second orbital portion.

3. 3. The method of claim 1, wherein step (b) is performed using at least one neural network.

4. 3. The method according to claim 1, wherein step (b) is performed using an optimization tool, the optimization tool being capable of generating at least one of the periodic basic trajectories according to the conditions and the n-tuple of gait parameters.

5. the exoskeleton accepts a human operator; 5. The method of claim 1, wherein step (a) comprises determining an n-tuple sequence of gait parameters for the exoskeleton desired by the operator.

6. The method of claim 5 , wherein the generated trajectory of the exoskeleton includes, for each n-tuple of the sequence, a new periodic elementary orbit and a transition to the new periodic elementary orbit.

7. The condition is: limiting the movement of the foot to pure rotation during the first track portion; 7. The method of claim 1, further comprising: during the second trajectory portion, limiting the movement of one foot to pure rotation or allowing the translation of one foot while forcing the other foot to remain immobile.

8. 8. The method of claim 1, wherein the condition includes forcing the foot, which is performing the translational movement in the second trajectory portion, to contact the ground only with the heel at the end of the second trajectory portion.

9. 9. The method of claim 1, wherein the condition includes forcing the foot that performs the translational movement in the second trajectory portion to contact the ground only at the tip end at the start of the second trajectory portion.

10. 10. The method of claim 1, wherein the condition includes forcing another foot to be flat on the ground at the start of the second trajectory portion.

11. 11. The method of claim 1, wherein the condition includes forcing one other foot to remain immobile during the second track portion.

12. 1. A method of moving an exoskeleton having multiple degrees of freedom, wherein at least one of said multiple degrees of freedom is actuated by an actuator controlled by data processing means, comprising:

12. A method comprising the step (c) of executing the trajectory of the exoskeleton generated by the method of any one of claims 1 to 11 by a data processing means of the exoskeleton, causing the exoskeleton to walk.

13. A system comprising a first server and an exoskeleton, each of the first server and the exoskeleton including a data processing means, 13. A system, wherein the data processing means is configured to execute a method for generating a trajectory for an exoskeleton according to any one of claims 1 to 11 and / or a method for moving an exoskeleton according to claim 12.

14. 13. A computer program comprising code instructions for carrying out the method for generating a trajectory for an exoskeleton according to any one of claims 1 to 11 and / or the method for moving an exoskeleton according to claim 12, when the program is run on a computer.

15. 12. A storage means readable by a computer device having recorded thereon a computer program product comprising code instructions for executing a method for generating a trajectory for an exoskeleton according to any one of claims 1 to 11 and / or a method for moving an exoskeleton according to claim 12.

Images