DE102020121561B3 - Control method for a bilateral teleoperation system, controller, teleoperation system and use of the controller and / or the teleoperation system - Google Patents

Abstract

Control method for a time-delayed bilateral teleoperation system, the method comprising the steps: - definition of a virtual time-delayed slave reference position on the master side, which is connected via a master-side virtual spring device with a virtual spring stiffness to the master position for force and position feedback, - definition of a virtual time-delayed master reference position on the slave side, which is connected to the slave position via a virtual spring device on the slave side with a virtual spring stiffness, - specification of the position to be assumed via the input device on the master side, transmission of the position to be assumed over the forward communication channel to the slave side and specification of the transmitted position Position as master reference position on the slave side, - Transfer of the slave position reached with the robot device to the master side via the return communication channel and selective Specifying the slave position via a controller as the position of the slave reference on the master side, - whereby the controller selectively only passes on the slave position reached as the position for the slave reference on the master side as long as it can be ensured that the bilateral teleoperation system is asymptotically stable is.

Description

The procedure of teleoperation is a crucial tool for performing robotic tasks in remote scenarios. In teleoperation, a spatial position to be assumed for a remote robot device located on a secondary or slave side is specified on a primary side or also a master side. For example, the robot device can be a handling device in the form of an electrically and / or pneumatically driven gripper arm, which is intended to assume a spatial position or a positional position. The position to be taken is specified via the primary page or the master page. The master and slave sides are locally spaced from one another and are only connected to one another via an outward and return communication channel for data communication. The master and slave side form a control loop to the extent that the position to be assumed is transmitted from the master side to the slave side via the forward communication channel and the transferred master position is specified as the position to be assumed by appropriate motor controls for the robot device. The position reached by the robot device is determined or scanned by means of appropriate sensor devices and reported back to the master side via the feedback channel and taken into account as a value for the further control of the robot device.

By adding a human operator to the previously described control loop instead of a fully autonomous teleoperation system, many of the task requirements, particularly in the areas of perception and cognition, can be bypassed. The human operator can, for example, use a corresponding input device on the primary or master side to select and specify a position to be assumed by the robot device on the secondary or slave side. The stability of the teleoperation system used and in particular of the control method used are very relevant here, since status and control variables must be exchanged between remote locations and locations (distance between the master and slave side). Time delays and packet losses in the control loop that occur due to the selected communication channels must be taken into account in the control system. To do this, additional techniques must be used to prevent the time delays or packet losses from destabilizing the system.

In recent decades, numerous papers have proposed new approaches to solving the above-mentioned problem. In addition, procedures based on passivity such as wave variables [ 1 ], [ 2 ], Energy storage [ 3 ] and the "Time Domain Passivity Approach" (TDPA, German passivity approach in the time domain) [ 4th ] are characterized by their robustness and their valuable properties. For example, the approaches described above are applicable to linear and non-linear systems, they are a sufficient condition for stability, and they only use input / output information regardless of system parameters. In addition to teleoperation, these approaches have been successfully applied to a number of other applications, including the control of a number of elastic actuators [ 5 ], [ 6th ], the multi-rate control [ 7th ], [ 8th ], the hierarchical whole-body control [ 9 ], [ 10 ] and the explicit force control [ 11 ], [ 12th ], to name just a few.

The most important prerequisite for bilateral teleoperation is the stability of the system or the control method, which is achieved in most control approaches at the expense of reduced performance, e.g. B. by a remaining and desired position distance between the master and slave robots. A number of approaches have been proposed with regard to time-delayed position synchronization. In [ 13th ] proposed the "Passive-Set-Position Modulation" (PSPM, dt. Passive-sentence position modulation) to store the energy derived from a virtual damper in a virtual energy storage device and to bring it into the system later without violating the continuous passivity condition. However, the effects of the samples are partially ignored and can cause instability of the system if the passivity condition of Colgate [ 14th ] is not fulfilled. A solution to this problem was found in [ 3 ] suggested. Nevertheless, there is still the need to carry out forward movements in the systems known from the prior art in order to “charge” the energy store several times, which can increase the physical and psychological stress of the task to be carried out via the teleoperation system. In [ 15th ] a wave-variable approach was introduced to enable position synchronization of two agents. Nevertheless, problems arise with regard to the transparency of the task [ 16 ], which in addition to attenuating the energy using a worst-case scenario, also suffers from wave reflection problems.

In addition, in [ 17th ] proposed an approach based on the TDPA to passivate a position-position teleoperation architecture. Although passivity is ensured in an adaptive manner, which is an advantage over wave-variable methods [ 18th ], this approach suffers from a positional drift between the master and slave sides. To address this problem, two types of solutions have been proposed. Add-on drift compensators for TDPA have been used in [ 19th ], [ 20th ], [ 21 ] suggested and in [ 22nd ] extended to multi-degree of freedom telemanipulation. Regardless of this, the drift compensation depends on the existence of so-called passivity gaps, which are rare with high delays. On the other hand, in [ 23 ] proposed a modification of the TDPA based on r-passivity that does not suffer from positional drift. However, since the approach is based on the power-based TDPA, its conservatism limits the permissible force feedback and thus the transmitted impedance between the master and slave side. In this case, the position synchronization is at the expense of a further increasing discrepancy between the impedance of the local (master side) and the remote task (slave side).

the DE 10 2020 113 409 A1 describes a method for controlling a slave system by means of a master system with haptic feedback in the master system, which originates from the slave system, wherein both systems communicate bidirectionally via a telecommunication channel, the time-varying communication delays in at least one direction and in the Usually has in both directions. From the DE 10 2014 004 115 B3 Another method for controlling a master-slave system has become known.

From the DE 20 2019 001 448 U1 an actuator system, in particular for teleactuation, with a first actuator, in particular for operation by a user, a second actuator, in particular for executing a movement of the user, a transmission channel between the first actuator and the second actuator and a controller is known, the Controller is designed so that the controller can measure the energy of the first actuator as target energy, wherein the transmission channel is designed to transmit the target energy to the second actuator and the controller is designed as a function of the transmitted target energy, the damping of the second actuator.

the DE 10 2016 105 682 A1 describes a regulated haptic system as well as a method for operating a regulated haptic system comprise the measurement of a contact force with the aid of a first force sensor and the subsequent generation of a first force signal, which can optionally be amplified and which is sent to a control unit, which then controls an actuator, to generate a scaled force on a haptic control unit that counteracts the measured contact force.

the DE 10 2015 100 694 A1 discloses a teleoperation system comprising: a slave having a drive unit that drives a gripping end effector, a kinematic coordinate of the end effector and a gripping force Feffektor being determinable with a camera which is preferably integrated in the slave and which is aligned with the end effector, a master, which is remotely connected to the slave, with at least one operating unit on which a user can exert a gripping force FG, the gripping force being transmitted to the slave, and with a visual user interface that displays the image of the camera, where the following applies that FG is linearly dependent on the kinematic coordinate and Feffector.

From the EP 3 538 328 B1 a system and method for instructing a robot is known. The system includes an immersive haptic interface so that operator interaction with a master robotic arm is reflected from a slave robotic arm that is arranged to interact with a workpiece. The interaction of the slave robot arm is reflected back to the master robot arm as haptic feedback to the operator. The dynamic system is continuously simulated forwards and new commands are calculated for the master robot arm and the slave robot arm.

1. (i) Die Ansätze und Verfahren können lediglich Systeme mit geringer Zeitverzögerung und/oder bekannter konstanter Zeitverzögerung stabilisieren.
2. (ii) Die ausreichenden Stabilitätsbedingungen der Teleoperationssysteme müssen im Frequenzbereich dargestellt werden, was Systemparameter erfordert, die in vielen praktischen Implementierungen möglicherweise nicht verfügbar sind.
3. (iii) Die Verfahren führen zur einer Positionsdrift zwischen der vom Master befohlenen und der durch die Robotereinrichtung eingenommenen Slaveposition.
4. (iv) Es kommt zu einer Diskrepanz zwischen der Impedanz der lokalen (Masterseite) und der entfernten Umgebung (Slaveseite).

The prior art approaches and methods described above have one or more of the following disadvantages:

1. (i) The approaches and methods can only stabilize systems with a small time delay and / or known constant time delay.
2. (ii) The sufficient stability conditions of the teleoperation systems must be represented in the frequency domain, which requires system parameters that may not be available in many practical implementations.
3. (iii) The methods lead to a position drift between the slave position commanded by the master and the slave position assumed by the robot device.
4. (iv) There is a discrepancy between the impedance of the local (master side) and the remote environment (slave side).

The present invention is based on the object of improving the systems and methods known from the prior art to the effect of controlling the time-delayed To improve teleoperation systems in such a way that there are no energy losses or inputs due to the time delay in order to specify a stable system or method.

According to the invention, the object is achieved by the features of the method according to claim 1, the controller for a bilateral teleoperation system according to claim 5, the bilateral teleoperation system according to claim 6 and the use according to claim 7.

The time-delayed bilateral teleoperation system comprises a master side with a haptic input device for specifying a position to be taken x _m and a slave side with a robot device for taking the position x _m specified on the master side, the master side with the slave side for data exchange via a time-delayed back and forth Return communication channel are connected.

• - Definition einer virtuellen zeitverzögerten Slave-Referenzposition x̃_s auf der Masterseite, welche über eine masterseitige virtuelle Federeinrichtung mit einer virtuellen Federsteifigkeit k_m mit der Masterposition x_m zur Kraft- und Positionsrückkopplung verbunden ist,
• - Definition einer virtuellen zeitverzögerten Master-Referenzposition x_md auf der Slaveseite, welche über eine slaveseitige virtuelle Federeinrichtung mit einer virtuellen Federsteifigkeit k_s mit der Slaveposition x_s verbunden ist,
• - Vorgeben der einzunehmenden Position x_m über die Eingabeeinrichtung auf der Masterseite, Übertragen der einzunehmenden Position x_m über den Vorwärtskommunikationskanal auf die Slaveseite und Vorgeben der übertragenen Position x_m als Master-Referenzposition x_md auf der Slaveseite,
• - Übertragen der mit der Robotereinrichtung erreichten Slaveposition x_s an die Masterseite über den Rückkommunikationskanal und selektives Vorgeben der Slaveposition x_s über eine Steuerung als Slave Referenzposition x̃_s auf der Masterseite,
• - wobei die Steuerung die erreichte Slaveposition x_s selektiv nur so lange als Slave-Referenzposition x̃_s auf der Masterseite weitergibt, wie sichergestellt werden kann, dass das bilaterale Teleoperationssystem asymptotisch stabil ist.

The control method according to the invention for the time-delayed bilateral teleoperation system comprises the steps:

• - Definition of a virtual time-delayed slave reference position x̃ _s on the master side, which is connected via a master-side virtual spring device with a virtual spring stiffness k _m to the master position x _m for force and position feedback,
• - Definition of a virtual time-delayed master reference position x _md on the slave side, which is connected to the slave position x _s via a slave-side virtual spring device with a virtual spring stiffness k _s ,
• - Specifying the position x _m to be assumed via the input device on the master side, transferring the position x _m to be assumed via the forward communication channel to the slave side and specifying the transferred position x _m as the master reference position x _md on the slave side,
• - Transfer of the slave position x _s reached with the robot device to the master side via the return communication channel and selective specification of the slave position x _s via a controller as slave reference position x̃ _s on the master side,
• - whereby the controller selectively passes on the reached slave position x _s as slave reference position x̃ _s on the master side only as long as it can be ensured that the bilateral teleoperation system is asymptotically stable.

In the method according to the invention described above, the parameters or variables for, for example, the positions (x) and for the spring stiffnesses (k) can be one-dimensional, but also multidimensional. The dimensionality of the master position, for example, can preferably be determined by the joint degrees of freedom of the input device and that of the slave position by the degrees of freedom of the robot device on the slave side. According to the invention, it can also be provided that the dimensions of the degrees of freedom and thus those of the variables of the method according to the invention differ from one another on the master and slave side, whereby preferably only selected variables of the variables available on the master side are mapped in the method according to the invention connected to selected variables on the slave side. In particular, provision can also be made to average several values on the master and / or slave side in order to reduce the degrees of freedom and to couple the average value with variables on the opposite control side.

Alternatively, provision can also be made to switch the assignment of the variables. For example, on the master side, an input device with one degree of freedom can be connected to a robot device with several degrees of freedom via the control method according to the invention, whereby the human operator can determine for which of the several degrees of freedom of the robot device the position to be assumed is to be specified via the input device. Thus, by switching over an input device with one degree of freedom, the position to be assumed can be specified one after the other for all degrees of freedom of the robot device with several degrees of freedom.

The robot device according to the invention, such as a handling device, is constructed from a series or any combination of joints and segments, the resulting body extends from a first fastening end to an opposite end effector. The robot device can be connected or fastened to a stationary or alternatively to a mobile structure, in particular to a mobile platform, via the first fastening end. A so-called end effector, such as a gripping device, can be arranged at the end effector.

The robotic device can comprise both actively driven and non-driven passive joints.

The term input device is understood to mean a device which is in turn formed from segments and joints, such as in particular a haptic input device such as a joystick, whereby a human operator preferably adjusts the alignment or configuration of the input device by applying forces by changing the joint positions can affect. The changes in the joint positions can be measured, in particular, by means of sensor devices and passed on to corresponding evaluation _{devices for determining the predetermined position x m} to be assumed. A force feedback to the human operator through the virtual time-delayed slave reference position x̃ _s on the master side can in turn be output via corresponding actuating devices in the joints of the input device. The virtual time-delayed slave reference position x̃ _s on the master side is coupled to the master position x _m for force feedback via a master-side virtual spring device with a virtual spring stiffness k _m. The strength of the force feedback can be adjusted via the spring stiffness k _m. The spring stiffness should be selected to be at least large enough to overcome the stiffness of the input device and at least the frictional forces that occur.

In order not to feed in any energy due to the time delay between the master and slave side, the invention provides for the master position and the delayed slave reference position on the master side to be decoupled, in the sense that the master position does not directly follow the delayed slave reference position, the slave position however, it follows the delayed reference position of the master. the 1 shows a mechanical model to illustrate the control method provided according to the invention, which reproduces the principles of modeling as mechanical elements. In particular, the control provided according to the invention, which only forwards the slave position as a reference slave position as long as it is, is shown as a transmission with a clutch.

According to the invention, a new local slave reference position is introduced on the master side, which forms the new reference of the slave position for the master. The position of the newly introduced slave reference is influenced by the delayed slave position. As a result, the slave position has an indirect influence on the master position via the newly introduced virtual slave reference position, while the master position has a direct influence on the slave position. A control according to the invention that is implemented on the master side is designed so that the delayed slave reference on the master side only follows the slave position as long as it is ensured that the system is asymptotically stable and therefore converges over time. Assuming that the slave has a stable trajectory after the control, the slave will also converge with respect to the reference position of the master due to the stability of the open control loop.

wobei x_m (t) die bevorzugt durch den menschlichen Bediener an der Eingabeeinrichtung vorgegebene Verschiebung ist, wobei der Vektor der über die Gelenke der Eingabevorrichtung vorgegebenen Verschiebungen x_m (t) die durch den Bediener vorgegebene Position x_m darstellt. Mit m_m als Masse der Eingabeeinrichtung, wobei b_m der Viskositätskoeffizient des Masters ist, α das lokale adaptive Dämpfungselement auf der Master-Seite ist, mit k_m als diskrete virtuelle Federkopplung zwischen Master und der Slave-Referenz auf der Masterseite, und ẽ_m(t) = x_m(t) - x̃_s(t) der Positionsfehler zwischen der Master- und Slave-Referenzposition ist, wobei sich die vorherige Beziehung bei Verwendung eines diskreten Zeitindexes schreiben lässt, als: ẽ_m(k) = x_m(k) - x̃_s(k), mit t = k * T_S mit T_S als Abtastrate der Positionsdetektion, mit k = 1,2,3,...,n ∈ N*. Unter der Annahme, dass die lokale Servorate des Masters mit der Aktualisierungsrate von x̃_s(k) schnell genug ist, kann die Mastersteuerung (k_m(x_m(k) - x̃_s(k)) als kontinuierlich angesehen werden [13]. Die virtuelle Federsteifigkeit k_m und/oder k_s kann anhand von [14] passiv parametrisiert werden, im Hinblick auf die Parameter der Steuerung, die Abtastfrequenz und die inhärente physikalische Dämpfung der Vorrichtung, wodurch der Effekt der Diskretisierung eliminiert und die Stabilität gewährleistet wird. Dies weist den Vorteil auf, dass das Gesetz der Schaltsteuerung in diskreter Zeit definierbar ist.In the bilateral teleoperation system according to the invention according to FIG 2 the master and local slave reference dynamics are given by: $$m_m{\ddot{x}}_m\left(t\right)+b_m{\dot{x}}_m\left(t\right)+\alpha {\dot{x}}_m\left(t\right)+k_m{\tilde{e}}_m\left(t\right)=0$$

where x _m (t) is the displacement preferably specified by the human operator at the input device, the vector of the displacements x _m (t) specified via the joints of the input device representing the _{position x m} specified by the operator. With m _m as the mass of the input device, where b _{m is} the viscosity coefficient of the master, α is the local adaptive damping _{element on the master side, with k m} as a discrete virtual spring coupling between the master and the slave reference on the master side, and ẽ _m (t) = x _m (t) - x̃ _s (t) is the position error between the master and slave reference position, whereby the previous relationship using a discrete time index can be written as: ẽ _m (k) = x _m (k) - x̃ _s (k), with t = k * T _S with T _S as the sampling rate of the position detection, with k = 1,2,3, ..., n ∈ N *. Assuming that the local servo rate of the master is _{fast enough with the update rate of x̃ s} (k), the master control (k _m (x _m (k) - x̃ _s (k)) can be viewed as continuous [13]. The virtual spring stiffness k _m and / or k _s can be passively parameterized on the basis of [14] with regard to the parameters of the control, the sampling frequency and the inherent physical damping of the device, whereby the effect of discretization is eliminated and stability is ensured. This has the advantage that the law of switching control can be defined in discrete time.

wobei ẽ_m(k) = x_m(k) - x_s(k - T_b) der um T_b verzögerte Positionsfehler zwischen der aktuellen Masterposition x_m und der Slaveposition x_s ist.The control of the slave reference position is dependent on the following relationships ( 2 ) to stabilize the teleoperation system. Switching between the three cases takes place in such a way that the slave reference on the master side imitates the delayed slave position on the slave side as precisely as possible without making the teleoperation system unstable: $${\tilde{x}}_s\left(k\right)=\{\begin{array}{l}\mathrm{F.}all\left(i\right):\hfill \\ {\tilde{x}}_s\left(k-1\right),\hfill \\ \text{for}\left|{\tilde{e}}_m\left(k\right)\right|<\left|e_m\left(k\right)\right|\hfill \\ \&\&\left|{\tilde{e}}_m\left(k\right)\right|>\left|{\tilde{e}}_m\left(k-1\right)\right|\hfill \\ \alpha =0\hfill \\ \mathrm{F.}all\left(ii\right):\hfill \\ {\tilde{x}}_s\left(k-1\right)+x_m\left(k\right)-x_m\left(k-1\right),\hfill \\ \text{for}\left|{\tilde{e}}_m\left(k\right)\right|<\left|e_m\left(k\right)\right|\hfill \\ \&\&\left|{\tilde{e}}_m\left(k\right)\right|\le \left|{\tilde{e}}_m\left(k-1\right)\right|\hfill \\ \alpha =\text{adaptively set based on generative energy}\hfill \\ \mathrm{F.}all\left(iii\right):\hfill \\ x_s\left(k-T_b\right),\hfill \\ \text{for}\left|{\tilde{e}}_m\left(k\right)\right|\ge \left|e_m\left(k\right)\right|\hfill \\ \alpha =0\hfill \end{array}$$

where ẽ _m (k) = x _m (k) - x _s (k - T _b ) is the _{position error delayed by T b} between the current master position x _m and the slave position x _s .

Case (i):

The method detects whether the human operator is introducing energy into the master-slave reference subsystem by activating the coupling spring k _m by moving the master device, | ẽ _m (k) | > | ẽ _m (k - 1) | ( 3a) , extends. It also detects whether the delayed slave position is trying to feed in energy in addition to the master, | ẽ _m (k) | <| e _m (k) |. In both cases, the controller keeps the slave reference position at its previous position | x̃ _s (k) | = | x̃ _s (k-1) |. This means that the potential energy limits of the coupling spring are only increased, if at all, by the operator by moving the master device.

Case (ii):

The procedure detects whether there is a distance between the delayed slave position and the slave reference position, | ẽ _m (k) | <| e _m (k) |, gives or not while the master holds its previous position or moves to the delayed slave position, | | ẽ _m (k) | ≤ | ẽ _m (k-1) | ( 3b , 3c ), moved to. This triggers the controller, which moves the slave reference position by the same order of magnitude and in the same direction as the master position. In other words, in case (ii) a saturated spring with a constant extension between the master and the slave reference is introduced (see 3c ). A saturated or constant force is exerted on the master-side input device in the direction of the slave position until the reference slave position exceeds the delayed slave position, lẽ _m (k) | = | e _m (k) |, reached. When the input device is used, the restoring force described above represents haptic feedback in the form of force feedback for the human operator of the device.

wobei ƒ_mc(k) die Kraftausgabe der Master-Steuerung ist, ẋ_m(k) die Geschwindigkeit der Master-Vorrichtung und T_S die Abtastzeit. Wenn E_beob(k) ≥ 0 für jedes k, dann bedeutet dies, dass der Ein-Port dissipativ ist. Wenn zu irgendeinem Zeitpunkt E_beob(k) < 0, dann erzeugt der Ein-Port Energie und die Menge der erzeugten Energie ist -E_beob(k), was zur Instabilität des Systems beitragen kann. Der Passivitätscontroller (PC) hat die Form eines zeitveränderlichen Elements a(k) in einer Reihenschaltung mit Impedanzkausalität, um nur die erforderliche Energiemenge zu dissipieren [4]. $$\alpha \left(k\right)=\{\begin{array}{rr}\hfill -\frac{E_{beob}\left(k\right)}{T_s{\dot{x}}_m{\left(k\right)}^2},& \hfill wenn{E}_{beob}\left(k\right)<0\\ \hfill \mathrm{0,}& \hfill wenn{E}_{beob}\left(k\right)\ge 0\end{array}$$

The introduction of a saturated spring is necessary so that the master position reaches the slave reference position, which in turn follows the delayed slave position. In a preferred embodiment, the “Time Domain Passivity Approach” (TDPA) is used as a tool to use the Passivity Observer (PO) to obtain the additional energy that is supplied to the master-slave reference subsystem due to the saturated spring , to be monitored and discharged via the passivity controller (PC). The master-slave reference subsystem can be viewed analogously to a one-port system, in which the PO may or may not be negative at a certain point in time, depending on the operating conditions and the peculiarities of the dynamics of the one-port element. Assuming an initial energy storage of zero, the power conjugate variables ƒ _mc (k) and ẋ _m (k) are used to monitor the energy flow: $${\mathrm{E.}}_{beOb}\left(k\right)={\mathrm{E.}}_{beOb}\left(k-1\right)+\left[{ƒ}_{mc}\left(k\right){\dot{x}}_m\left(k\right)+\alpha \left(k-1\right){\dot{x}}_m{\left(k-1\right)}^2\right]T_s$$

where ƒ _mc (k) is the force output of the master controller, ẋ _m (k) is the speed of the master device, and T _{S is} the sampling time. If E _obs (k) ≥ 0 for every k, then it means that the one-port is dissipative. If at any point in time E _obs (k) <0, then the one-port is generating energy and the amount of energy produced is -E _obs (k), which can contribute to the instability of the system. The passivity controller (PC) has the form of a time-varying element a (k) connected in series with impedance causality in order to dissipate only the required amount of energy [4]. $$\alpha \left(k\right)=\{\begin{array}{rr}\hfill -\frac{{\mathrm{E.}}_{beOb}\left(k\right)}{T_s{\dot{x}}_m{\left(k\right)}^2},& \hfill wenn{\mathrm{E.}}_{beOb}\left(k\right)<0\\ \hfill \mathrm{0,}& \hfill wenn{\mathrm{E.}}_{beOb}\left(k\right)\ge 0\end{array}$$

Using the above-described relationship, no additional energy is fed into the master proxy subsystem by the action of the saturated spring.

Case (iii):

The procedure determines whether the condition | ẽ _m (k) | ≥ | e _m (k) |. If so, then the control enables lẽ _m (k) | = | e _m (k) | by choosing the reference slave position equal to the delayed slave position | x̃ _s (k) | = | x _s (kT _b ) |. In this state, | x _s (k) | holds either its previous position or moving towards the master ( 3d ).

The master proxy subsystem described by equation (1) is asymptotically stable using the laws of switching control described in equation (2). The slave control follows the delayed master reference using a stable trajectory slave control, and therefore the positions of the master, local proxy and slave converge. The human operator feeds energy into the master-proxy coupling while interacting with the master robot. However, if the operator releases the master robot, the proposed control monotonically reduces the energy due to the dissipation from the inherent physical damping of the master robot and the adaptive damping element of the passivity controller (PC).

In the following, with reference to the 4th until 6th an experimental exemplary setup for presenting the invention and demonstrating the effectiveness of the control method according to the invention is described.

ms auf. Es wurde eine Positions-Positions-Steuerarchitektur mit einer Kopplungsverstärkung von k_m = 1,5 N/mm, b_m = 0,01 Ns/mm auf der Master-Seite und k_s = 1,5 N/mm, b_s = 0,01 Ns/mm auf der Slave-Seite implementiert. Die Verstärkungen waren so abgestimmt, dass bei keiner Verzögerung die durch die Diskretisierung erzeugte Energie abgebaut wurde und der Teleoperator daher passiv war.the 4th Figure 3 shows two rotating devices, each with one degree of freedom, used as an example of a teleoperation system according to the invention. Both rotary devices have a sampling rate ƒ _s = 1 kHz with the associated sampling time $T_s=\frac{1}{{ƒ}_s}=1$

ms on. A position-position control architecture was used with a coupling gain of k _m = 1.5 N / mm, b _m = 0.01 Ns / mm on the master side and k _s = 1.5 N / mm, b _s = 0.01 Ns / mm implemented on the slave side. The amplifications were coordinated in such a way that the energy generated by the discretization was not dissipated in the event of any delay and the teleoperator was therefore passive.

Experimental results:

The experiments for the proposed control were carried out in such a way that the human operator uses the rotary device on the master side as an input device according to the invention and maneuvers so that the robot device on the slave side, formed by the second rotary device, comes into contact with a hard wall (over 150 N / mm) which is located at 0.85 rad. Even for a well-coordinated coupling control, such an interaction would become unstable if a “large” time delay were introduced. This is over 5 shows where the system becomes unstable if the operator disrupts and releases the master, with a return time as a modeled time delay of the communication channels of 200 ms.

The experiment was repeated by implementing the control method according to the invention, the human operator having specified 4 contacts with the hard wall as the position to be assumed for the slave-side rotating device. the 6th shows the position progressions for the master position (master), the time-delayed slave position (slave delay) and the slave reference position (slave proxy), the time delay of the modeled communication channels between the master and slave side in the 6 a) and b) is on Trt = 200 ms and for the 6 c) and d ) selected to Trt = 1000 ms. For T _rt = 200 ms, the delayed slave position and the reference slave position are almost identical and there is no position drift between the master and the slave side. 6b shows that the master and slave have similar torque values when the user comes into contact with the hard wall. Even with higher delays of T _rt = 1000 ms, the interaction remains stable, but the slave reference position on the master side does not exactly follow the delayed slave position, as in 6c you can see. There are some cases where the slave reference keeps its previous position while ignoring the delayed slave position (t = 20-22 s, 24-25.5 s, 28.5-30 s). However, there is no positional drift between the master and slave side, and the master and slave torques generated are almost identical if the human operator specifies a position at which the slave-side rotary device comes into contact with the hard wall. This further shows that the passivity controller (PC) does not trigger when the operator specifies a position to be assumed in which the slave-side rotating device collides with a static environment, such as a hard wall, which increases transparency.

Credentials:

1. [1] Robert J Anderson and Mark W Spong. Bilateral control of teleoperators with time delay. IEEE Transactions on Automatic control, 34 (5): 494.501, 1989 .
2. [2] Günter Niemeyer and J-JE Slotine. Stable adaptive teleoperation. IEEE Journal of oceanic engineering, 16 (1): 152.162, 1991 .
3. [3] Michel Franken, Stefano Stramigioli, Sarthak Misra, Cristian Secchi, and Alessandro Macchelli. Bilateral telemanipulation with time delays: A two-layer approach combining passivity and transparency. IEEE transactions on robotics, 27 (4): 741.756, 2011 .
4. [4] Jee-Hwan Ryu, Jordi Artigas, and Carsten Preusche. A passive bilateral control scheme for a teleoperator with time-varying communication delay. Mechatronics, 20 (7): 812.823, 2010 .
5. [5] Dylan P Losey, Andrew Erwin, Craig G McDonald, Fabrizio Sergi, and Marcia K O'Malley. A time-domain approach to control of series elastic actuators: Adaptive torque and passivity-based impedance control. IEEE / ASME Transactions on Mechatronics, 21 (4): 2085.2096,2016.
6. [6] Hyunwook Lee, Jinoh Lee, Jee-Hwan Ryu, and Sehoon Oh. Relaxing the conservatism of passivity condition for impedance controlled series elastic actuators. In 2019 IEEE / RSJ International Conference on Intelligent Robots and Systems (IROS), pages 7610.7615. IEEE, 2019 .
7. [7] Marco De Stefano, Hrishik Mishra, Ribin Balachandran, Roberto Lampariello, Christian Ott, and Cristian Secchi. Multi-rate tracking control for a space robot on a controlled satellite: A passivity-based strategy. IEEE Robotics and Automation Letters, 4 (2): 1319.1326, April 2019 .
8. [8th] Marco De Stefano, Ribin Balachandran, and Cristian Secchi. A passivity-based approach for simulating satellite dynamics with robots: Discrete-time integration and time-delay compensation. IEEE Transactions on Robotics, 36 (1): 189.203, Feb 2020 .
9. [9] Alexander Dietrich, Xuwei Wu, Kristin Bussmann, Christian Ott, Alin Albu-Schäffer, and Stefano Stramigioli. Passive hierarchical impedance control via energy tanks. IEEE Robotics and automation letters, 2 (2): 522.529, 2016 .
10. [10] Andre Coelho, Harsimran Singh, Konstantin Kondak, and Christian Ott. Whole-body bilateral teleoperation of a redundant aerial manipulator. In 2020 IEEE International Conference on Robotics and Automation (ICRA), 2020 .
11. [11] Ribin Balachandran, Mikael Jorda, Jordi Artigas, Jee-Hwan Ryu, and Oussama Khatib. Passivity-based stability in explicit force control of robots. In 2017 IEEE International Conference on Robotics and Automation (ICRA), pages 386.393. IEEE, 2017 .
12. [12] Mikael Jorda, Ribin Balachandran, Jee-Hwan Ryu, and Oussama Khatib. New passivity observers for improved robot force control. In 2017 IEEE / RSJ International Conference on Intelligent Robots and Systems (IROS), pages 2177.2184. IEEE, 2017 .
13. [13] Dongjun Lee and Ke Huang. Passive-set-position-modulation framework for interactive robotic systems. IEEE Transactions on Robotics, 26 (2): 354.369, 2010.
14. [14] J Edward Colgate and Gerd G Schenkel. Passivity of a class of sampled data systems: Application to haptic interfaces. Journal of robotic systems, 14 (1): 37.47, 1997.
15. [15] Emmanuel Nu.no, Luis Basa.nez, Romeo Ortega, and Mark W Spong. Position tracking for non-linear teleoperators with variable time delay. The International Journal of Robotics Research, 28 (7): 895.910, 2009 .
16. [16] Dale A Lawrence. Stability and transparency in bilateral teleoperation. IEEE transactions on robotics and automation, 9 (5): 624.637, 1993 .
17. [17] Jordi Artigas, Jee-Hwan Ryu, and Carsten Preusche. Time domain passivity control for position-position teleoperation architectures. Presence: Teleoperators and Virtual Environments, 19 (5): 482.497, 2010 .
18. [18] Ribin Balachandran, Jordi Artigas, Usman Mehmood, and Jee-Hwan Ryu. Performance comparison of wave variable transformation and time domain passivity approaches for time-delayed teleoperation: Preliminary results. In 2016 IEEE / RSJ International Conference on Intelligent Robots and Systems (IROS), pages 410.417. IEEE, 2016 .
19. [19] Jordi Artigas, Jee-Hwan Ryu, and Carsten Preusche. Position drift compensation in time domain passivity based teleoperation. In 2010 IEEE / RSJ International Conference on Intelligent Robots and Systems, pages 4250.4256. IEEE, 2010 .
20. [20] Vinay Chawda, Ha Van Quang, Marcia K O'Malley, and Jee-Hwan Ryu. Compensating position drift in time domain passivity approach based teleoperation. In Haptics Symposium (HAPTICS), 2014 IEEE, pages 195.202. IEEE, 2014 .
21. [21] Andre Coelho, Harsimran Singh, Tin Muskardin, Ribin Balachandran, and Konstantin Kondak. Smooth position-drift compensation for time domain passivity approach based teleoperation. In Intelligent Robots and Systems (IROS), 2015 IEEE / RSJ International Conference on. IEEE, 2018 .
22. [22] Andre Coelho, Chistian Ott, Harsimran Singh, Fernando Lizarralde, and Konstantin Kondak. Multi-dof time domain passivity approach based drift compensation for telemanipulation. In 2019 19th International Conference on Advanced Robotics (ICAR), pages 695.701, Dec 2019 .
23. [23] Vinay Chawda and Marcia K O'Malley. Position synchronization in bilateral teleoperation under time-varying communication delays. Ieee / asme transactions on mechatronics, 20 (1): 245.253, 2015 .
24. [24] Ali Jazayeri and Mahdi Tavakoli. Absolute stability analysis of sampled-data scaled bilateral teleoperation systems. Control Engineering Practice, 21 (8): 1053.1064, 2013 .
25. [25] Jake J Abbott and Allison M Okamura. Analysis of virtual fixture contact stability for telemanipulation. In Proceedings 2003 IEEE / RSJ International Conference on Intelligent Robots and Systems (IROS 2003) (Cat. No. 03CH37453), volume 3, pages 2699.2706. IEEE, 2003 .
26. [26] Hyun Chul Cho and Jong Hyeon Park. Stable bilateral teleoperation under a time delay using a robust impedance control. Mechatronics, 15 (5): 611.625, 2005 .
27. [27] Aghil Jafari, Muhammad Nabeel, Harsimran Singh, and Jee-Hwan Ryu. Stable and transparent teleoperation over communication timedelay: Observer-based input-to-state stable approach. In 2016 IEEE Haptics Symposium (HAPTICS), pages 235.240. IEEE, 2016 .

Claims (7)

Control method for a time-delayed bilateral teleoperation system, comprising a master side (m) with a haptic input device for specifying a position to be _{assumed x m} and a slave side (s) with a robot device for _{assuming the position x m} specified on the master side, the master side with the slave side for data exchange via a time-delayed forward and return communication channel (f, b), the method comprising the steps: - definition of a virtual time-delayed slave reference position x̃ _s on the master side, which via a master-side virtual spring device with a virtual spring stiffness k _m is connected to the master position x _m for force and position feedback, Definition of a virtual time-delayed master reference position x _md on the slave side, which is connected to the slave position x _s via a slave-side virtual spring device with a virtual spring stiffness k _s , - specification of the position x _{m to be taken} via the input device on the master side, transfer of the position to be taken Position x _m via the forward communication channel on the slave side and specification of the transferred position x _m as master reference position x̃ _m on the slave side, - transmission of the slave position x _s reached with the robot device to the master side via the return communication channel and selective specification of the slave position x _s via a control as position x̃ _{s of} the slave reference on the master side, - whereby the control _{only passes on the reached slave position x s} selectively as position x̃ _s for the slave reference on the master side as long as it can be ensured that the bilateral teleoperation system asymptoti sch is stable, characterized by the specification of the position x̃ _s for the slave reference depending on the following relationship: $$\begin{array}{l}{\tilde{x}}_s\left(k\right)=\{\begin{array}{l}\mathrm{F.}all\left(i\right):\hfill \\ {\tilde{x}}_s\left(k-1\right),\hfill \\ \text{for}\left|{\tilde{e}}_m\left(k\right)\right|<\left|e_m\left(k\right)\right|\hfill \\ \&\&\left|{\tilde{e}}_m\left(k\right)\right|>\left|{\tilde{e}}_m\left(k-1\right)\right|\hfill \\ \alpha =0\hfill \\ \mathrm{F.}all\left(ii\right):\hfill \\ {\tilde{x}}_s\left(k-1\right)+x_m\left(k\right)-x_m\left(k-1\right),\hfill \\ \text{for}\left|{\tilde{e}}_m\left(k\right)\right|<\left|e_m\left(k\right)\right|\hfill \\ \&\&\left|{\tilde{e}}_m\left(k\right)\right|\le \left|{\tilde{e}}_m\left(k-1\right)\right|\hfill \\ \alpha =\text{adaptively set based on generative energy}\hfill \\ \mathrm{F.}all\left(iii\right):\hfill \\ x_s\left(k-T_b\right),\hfill \\ \text{for}\left|{\tilde{e}}_m\left(k\right)\right|\ge \left|e_m\left(k\right)\right|\hfill \\ \alpha =0,\hfill \end{array}\\ \end{array}$$

where ẽ _m (k) = x _m (k) - x _s (k - T _b ) represents the _{position error delayed by T b} between the current master position x _m (k) and the delayed slave position x _s (k - T _b ), with T _b as the time delay of the return communication channel (b), with t = k * T _S , where k = 1, 2, 3, ..., ne N * and with T _S as the sampling rate of the position detection. Procedure according to Claim 1 , characterized by the passive parameterization of the virtual spring stiffness of the master-side spring device k _m and / or the virtual spring stiffness of the slave-side spring device k _s with regard to the parameters of the control, the sampling frequency and the inherent physical damping of the bilateral teleoperation system. Method according to one of the Claims 1 or 2 , characterized by the specification of the parameter α in case (ii) as a time-variable element a (k) with impedance causality based on the equation: $$\alpha \left(k\right)=\{\begin{array}{rr}\hfill -\frac{{\mathrm{E.}}_{beOb}\left(k\right)}{T_s{\dot{x}}_m{\left(k\right)}^2},& \hfill wenn{\mathrm{E.}}_{beOb}\left(k\right)<0\\ \hfill \mathrm{0,}& \hfill wenn{\mathrm{E.}}_{beOb}\left(k\right)\ge 0\end{array},$$

with $${\mathrm{E.}}_{beOb}\left(k\right)={\mathrm{E.}}_{beOb}\left(k-1\right)+\left[{ƒ}_{mc}\left(k\right){\dot{x}}_m\left(k\right)+\alpha \left(k-1\right){\dot{x}}_m{\left(k-1\right)}^2\right]T_s,$$

with f _mc as the power output of the master control. Method according to one of the preceding claims, characterized by the specification of the position x _m to be assumed by a human operator via the input device. Controller for a time-delayed teleoperation system, comprising at least one data processing unit, the data processing unit for carrying out the method according to one of the Claims 1 until 4th configured. Bilateral teleoperation system, comprising a master page (m) with a haptic input device for specifying a position x _m to be assumed; a slave side (s) with a robot device for taking up the position x _m specified on the master side; at least one outward and return communication channel (f, b) for data exchange between the master and slave side; and at least one controller according to Claim 5 . Use of a controller according to Claim 5 or a teleoperation system according to Claim 6 to carry out the control method according to one of the Claims 1 until 4th .

Images