DE102005025536A1 - Mobile machine used as a hydraulic driven excavator comprises a unit for generating traveling and working movement, devices for measuring the position and/or the speed of working hinges and pressure sensors - Google Patents

Abstract

Mobile machine comprises a unit for generating traveling and working movement, devices for measuring the position and/or the speed of working hinges and pressure sensors arranged in drive elements and/or wire strain gauge sensors for measuring elastic bending and inertial sensors and/or optical measuring units for controlling the hinges with a combined active oscillation damping and/or contact power control for the digging force. An independent claim is also included for a method for moving earth and bulk material. Preferred Features: The unit for generating traveling and working movement is automatic or semi-automatic.

Description

The The invention relates to mobile working machines, in particular hydraulically powered earthmoving machinery, such as inlet loading machines (discontinuous loaders), backhoes (for example, general purpose excavators) and flat dredgers (for example, leveling machines), and a method of earth and bulk material movement to this Machinery is applied.

hydraulic Powered earthmoving machines are loaders and excavators that have one Diesel engine, a landing gear, a working equipment and a tool to Loading, shoveling or gripping. Draw the devices particular characterized by the fact that they have hydraulic drive elements dispose, mobile be used and for Earth and bulk material movements be used.

At the Operation of a mobile work machine continues the work cycle often composed of repetitive operations. For example a universal excavator with a bucket tool for mooring used a shaft, the following steps are performed: To fill of the grave vessel, lifting and lowering the boom, swinging back and forth of the superstructure and emptying the grave vessel. These procedures are essentially repeated until the desired Aushubprofil achieved is. An efficient and accurate working movement requires from the operator a high level Concentration and experience.

In the prior art, a variety of automation solutions for these cycles are known. For example, in the DE 199 45 967 proposed a method and an apparatus for controlling angles of a work machine with which even inexperienced operators can perform the excavation and loading operation easily and efficiently. Also in the EP 1 416 095 For example, a method for automatically moving a tool mounted on the articulated arm end on a terrain with a simplified control system for driving the tool will be described.

Furthermore, it is known that the control quality in these automated work cycles in contact situations of the tool, such as digging or gripping, may decrease. The contact forces can also lead to damage or even destruction of the implement, eg in a collision. This is from the EP 6 575 90 an automatic excavator control system for a backhoe is known, which has a Grablaststeuereinrichtung for detecting a Grablast with suitable sensors.

however is not considered in the known state of the art, that load oscillations occur during the said work steps, because the individual joints have an arm flexibility, the oil hydraulics is compressible to a certain extent, either over a landing gear spring-loaded suspension feature or have an elastic structure, e.g. at one with Filled with air pressure Tire is the case, or by supporting elements an additional twist elasticity of the vehicle chassis is caused. Especially for mobile vehicles are due to road bumps and fast load changes operational vibrations excited. On the other hand, vibrations also by external disturbances, like for example, by emptying the digging tool caused. Vibrations express themselves in rolling, pitching and lifting movements and can lead to machine damage. moreover Vibrations reduce the ease of use and are often by the driver as disturbing felt.

task Therefore, the present invention is a mobile work machine and to provide a method of earth and bulk material movement, with which it is possible to automate entire work cycles or parts thereof, taking into account occurring vibrations and / or contact forces become a higher one and more consistent productivity with less qualification of the operator, a higher Safety at work and less physical and psychological To reach the burden of people in the work process.

According to the invention succeeds the solution This object with the features of claims 1 and 6.

advantageous Embodiments of the mobile working machine according to the invention and the method of the invention for earth and bulk material movement are in the subclaims specified.

The The invention is explained in more detail below with reference to drawings. In show the drawings:

1 Block diagrams of the semi-automatic and fully automatic excavator operation according to the invention

2 - Presentation of an input loading machine

3 - Mechanical replacement model according to the invention for an excavator

4 - Representation of the external forces and forces acting on an excavator

According to the invention, the driving and working movement of a mobile work machine is generated with the aid of a path planner and implemented by means of a trajectory sequence controller, wherein the path planner has semi-automatic and automatic basic features ( 1 ). Examples of an automated construction project are: leveling with a bulldozer, in which case the path planner generates the synchronized movements of the chassis and working joints; Creating a shaft with an excavator; Bulk movements, so that eg a heap of earth is removed automatically. Examples of automated subtasks in earthmoving include: automatic digging, automatic movement between two or more locations, synchronized movement of the work implement (eg, a bucket following a particular path in the world coordinate system).

Becomes the path planner operated in semi-automatic mode, an operator can the speed and direction of an automated (part) task pretend or manually control the individual joint positions.

The in the said work machines resulting load oscillations be according to the invention with the help a state feedback active steamed, wherein the method of vibration damping with the above-described Trajektorienfolgeregler is combined. The vibrational states become measured with suitable sensors or with an observation algorithm estimated.

Lies a contact situation of the implement, such as digging or when gripping, the control quality of the Trajektorienfolgeregelung lose weight. In addition, you can Contact forces, e.g. in a collision, destroy the implement. That's why According to the invention a subordinate Force control used. Here are the drive actuators each with two oil flow units Mistake. For example, the differential cylinder of the stem cylinder a universal excavator with two 3/3-way valves operated. It results a multivariate system, so that in addition to the force control, the pressure level in the cylinder the two chamber pressures as additional Define controlled variable leaves. This is the oil pressure separately regulated in each cylinder chamber, resulting from the pressure difference gives the given cylinder force. The advantage is to be seen in that with a pure power regulation the grasping or digging targeted and safely done can be. When gripping, e.g. the necessary gripping and holding force exactly implemented by the force controller. On the other hand, too the trajectory sequence controller combined with the subordinate force controller be so disturbing contact forces in the sense of a forward connection and thus eliminate it in the sense of active compensation.

Farther it is also within the scope of the invention that when digging an observer to estimate the Digging force can be used. This estimated grave force will then become Compensation used so that the superimposed orbit control has good follow-up behavior. In addition, through the use of the observer safety-relevant situations are detected.

1. Modeling:

The design of the prime mover and the work equipment determine the number of hydraulically driven joint degrees of freedom. Cement dredgers usually have a pivoting superstructure, a movable boom in various design variants and an undercarriage with wheeled or tracked chassis ( 2 ). Recessed loading machines and flat excavators differ to the latter only in that they have no pivoting superstructure.

The inventive method for earth and bulk material movement is based on the basic idea that dy to simulate namic and static behavior of the mechanical and hydraulic system of the working machine in a physical model. First of all, a rigid body model is set up with the help of the Lagrangian formalism for the spatial movement of the joints:

M

Massenmatrix

q

Vektor der Gelenkwinkel und Gelenkpositionen

C

Kraft- und Momentenmatrix der Coriolis- und Zentrifugaleffekte

G

Matrix der Gravitationskräfte und Gravitationsmomente

u

Antriebskräfte und Antriebsmomente

Here are:

M

mass matrix

q

Vector of joint angles and joint positions

C

Force and moment matrix of the Coriolis and centrifugal effects

G

Matrix of gravitational forces and gravitational moments

u

Driving forces and drive torques

equation (1) describes the dynamic and static movement of the Joints in dependence the imprinted driving forces and drive torques, with the help of oil-hydraulic energy converters such as hydraulic motors or hydraulic cylinders are applied.

As already mentioned, load vibrations occur during operation of the work machines, since the individual joints have an arm flexibility and the oil hydraulics are compressible to a certain extent. Therefore, the mechanical system is modeled with rigid model bodies and discrete elasticities and damping ( 3 ). The equation of motion (1) is extended by the elasticity degrees of freedom δ , which describe eg horizontal, vertical and torsional vibration components of the individual drive joints:

δ

Vektor der Elastizitätsfreiheitsgrade

D

Vektor der Schwingungsdämpfung

K

Vektor der Federkonstanten der rückstellenden Kraft oder des Moments

Here are:

δ

Vector of elasticity degrees of freedom

D

Vector of vibration damping

K

Vector of spring constant of restoring force or moment

The individual elements of K can be illustrated as spring constant, which includes on the one hand the bending and on the other hand the compression phenomenon of the oil hydraulics. The restoring force is linearized by one operating point. The total spring constant of a joint can therefore be calculated as a series connection of the spring constants of the oil compression and spring constants of the arm flexibility:

c _i is the resulting spring constant of an exemplary joint with the index 'i'. The variables c _{i_bieg} and c _{i_Komp} are the spring constants of the arcing and oil compression. Depending on the system configuration, further elasticity phenomena may become more relevant. For example, a non-rigid chassis adds extra elasticity. Formula (3) is extended accordingly.

When using a hydraulic cylinder, the drive torque of the corresponding joint can be determined by taking into account the geometric articulation kinematics of the joint and cylinder with an articulation coefficient: u i (q i ) = F Zyli · φ i (q i ) (4)

u _i is the moment of an exemplary joint with the index 'i', F _Zyli is the cylinder force and φ _i (q _i ) is the coupling coefficient as a function of the joint angle. If a differential cylinder with only one delivery source is applied, the cylinder force results in:

p _cyli is the pressure in the cylinder (depending on the direction of movement piston or ring side), A _zyli is the cross-sectional area of the cylinder (depending on the direction of movement piston or ring side), β is the oil compressibility, V _zyli is the cylinder volume (depending on the direction of movement piston-cylinder. or ring side), Q _i is the flow rate of a lower-level flow controller and K _i is the proportionality constant, which describes the relationship between flow and control signal y _{i of} the controlled flow unit. Dynamic effects of the flow unit are neglected. The variables z _i , z. _i describe the position or speed of the cylinder rod and depend on the linkage kinematics.

If a differential cylinder is charged with two independent flow sources, the cylinder force results in:

Equation (6) differs from equation (5) only in that the ring and end face of the cylinder is subjected to two independent flow sources. In this case, the joint i considered by way of example comprises the two input _quantities y _{ring_i} and y _{stirn_i} .

A hydraulic drive motor is described by the following equations:

u _j is the moment of an exemplary joint with the index 'j', i _j is the gear ratio between engine speed and drive speed of the joint, V _j is the displacement of the hydraulic motors, Δp _j is the pressure drop across the hydraulic drive motor, β is the oil compressibility, φ. _Motj indicates the engine speed, Q _j is the flow rate in the hydraulic circuit and K _j is the proportionality constant that indicates the relationship between flow rate and the control voltage y _j of the regulated delivery flow unit. Dynamic effects of flow control are neglected. The moment _uj in equation (7) can be converted into the resultant drive force with the lever length ratio, as is necessary, for example, in the traction drive.

If the hydraulic motor is supplied with two independent flow sources, the moment will be:

Equation (8) differs from equation (7) only in that the intake passage and the exhaust passage of the engine are supplied with two independent flow sources. In this case, the joint, j 'considered as an example, comprises the two input _quantities y _{Ein_j} and y _{Aus_j} .

The equation of motion (1) and (2) describe a free movement of the prime mover. Contact situations, such as those that occur when digging, are not taken into account here. As force introduction points on the one hand, the hydraulic drive units for turning the excavator and the excavator arms and the forces occurring when the excavator bucket penetrates into the soil to be considered ( 4 ). For this reason, equation (1) is extended as follows:

If arm and oil hydraulic elasticities are considered, equation (2) becomes:

The variable u _contact ( q ) generally designates the external contact forces or contact torques acting on the drive machine, which occur, for example, during trenching.

• • Eindringwiderstand an der Grundschneide und an den Seitenwänden
• • Reibungswiderstand an den Seitenwänden und Seitenschneiden

To describe the external forces or moments a physical model for the different contact situations is deposited. As an example, the excavation force model is listed in more detail when a bucket is used as a working tool. It is assumed that the following components act on the elements of the grave tool:

• • Penetration resistance on the base cutting edge and on the side walls
• • Frictional resistance on the sidewalls and side cutting

The cutting force caused by the penetration resistance at the base cutter can be described by the following formula:

F_{ein_Schneide}

Schnittkraft an der Schneide

ϛ_f

Erdstoffdichte fest

h

Schnitttiefe (Spandicke)

b

Breite des Erdstoffkeils

g

Erdbeschleunigung

ϑ

Abbruchwinkel (Neigungswinkel der Abbruchfläche)

φ

Reibungswinkel im Erdstoff (innerer Reibungswinkel)

φ_St

Reibungswinkel Erdstoff-Schneide (Stahl)

α

Freiwinkel

γ

Schüttwinkel

Here are:

F _{ein_Schneide}

Cutting force at the cutting edge

_f

Earth density tight

H

Cutting depth (chip thickness)

b

Width of the earth wedge

G

acceleration of gravity

θ

Abbruchwinkel (inclination angle of the demolition area)

φ

Friction angle in the earth (internal friction angle)

φ _St

Friction angle of ground cutting edge (steel)

α

clearance angle

γ

angle of repose

When determining the frictional force caused by the side cutting edges, it is assumed that a shell-shaped element of earth material is created in front of the side cutting edge, which forms a bulky earth wedge directly in front of the face. The cutting force results from the condition of a passive earth pressure model:

F_{Ein_Seite}

Schnittkraft an der Seitenschneide

V

Volumen des Erdstoffkeils als Funktion der Schnitttiefe

δ

Schnittwinkel

Here are:

F _{one_page}

Cutting force at the side cutting edge

V

Volume of the wedge of the soil as a function of the depth of cut

δ

cutting angle

Frictional forces on the side walls and on the side cutters are proportionately divided into a coulomb friction and a viscous friction: F Reib_Seite_Schneide_coul = F N (H) μ C z F Reib_Seite_Schneide_visk = F N (H) μ V z · v (13)

F_{Reib_Seite_Schneide_coul}

coulombsche Reibungskraft

F_{Reib_Seite_Schneide_visk}

viskose Reibungskraft

F_N

Normalkraft aus dem passiv wirkenden Erddruck als Funktion der Schnitttiefe

μ_C

Reibungskoeffizient der Coulombreibung

μ_V

Reibungskoeffizient der viskosen Reibung

z

Anzahl der beteiligten Seitenwände

v

Schnittgeschwindigkeit

Here are:

F _{Reib_Seite_Schneide_coul}

coulombic frictional force

F _{Reib_Seite_Schneide_visk}

viscous frictional force

F _N

Normal force from the passive earth pressure as a function of the depth of cut

μ _C

Coefficient of friction of coulomb friction

μ _V

Coefficient of friction of viscous friction

z

Number of side walls involved

v

cutting speed

The entire digging power is from the addition of the above listed shares: F dig = F Ein_Schneide + F Ein_Seite + F Reib_Seite_Schneide_coul + F Reib_Seite_Schneide_visk (14)

For bulldozer shields and other types of tools, it is possible to derive corresponding excavation force models according to the above principles. Essentially, the digging force depends on the joint angle position and on parameters dependent on soil and tool type:

The vector P contains all parameters of the digging force model.

to Description of the external contact forces or moments may alternatively to an observer can be used, who appreciates the mentioned sizes. basis the state space representation (10) offers this. Additional to the contact force is over described a suitable static or dynamic model. Based then the equation (10) is extended. In the simplest case can the grave force as piecewise be accepted constantly. To model the digging force as accurately as possible can, the dynamic process of digging is physically described, represented by differential equations and then on coupled the equation of motion (10). The temporal course of the Grave left itself as a function of system states describe the extended system differential equation. The task of the observer lies therein, these required for the reconstruction conditions appreciate. This is done using the dynamic model with a suitable algorithm implemented on a machine, making a numerical simulation the system behavior is present. Subsequently, the estimation error determined, weighted with a suitable matrix and the simulation model as additional Input size switched on, so the estimation error asymptotically converged to zero. In this case, the mentioned estimation error from the difference between the measured and the numerical evaluation observed conditions educated. It is essential that not the complete system conditions are measured Need to become, but with the help of the described method only states are needed the metrologically available stand. Examples include measurements of the joint angle, the drive pressure and / or measurements of elastic degrees of freedom of movement. These be together with the system inputs in the observation tree processed, so the states be estimated to describe the grave or contact force can.

2. Automation and Control structure:

The The following described automation task is limited hydraulically driven wheeled excavators and carry out the earth-loading process by moving their work equipment by themselves. aim is it to relieve the operator and a fluctuating qualification one consistent productivity sure. This operation will be described below as semi-automatic Operation designated.

in the fully automatic operation, this concept is a parent Railway planning module extended. In this mode of operation will then Defined missions that independently coordinated the railway planning module movements implements.

• 2.1 Trajektorienfolgeregelung der Arbeitsgelenke
• 2.2 Trajektorienfolgeregelung mit aktiver Schwingungsdämpfung
• 2.3 Kraftregelung mit oder ohne zuschaltbaren Trajektorienfolgeregler auch bei Kontaktsituation (z.B. beim Graben)

The aim is to equip this device with the system described according to the invention in order to perform certain classes of earth movement automatically. Automated relocation of the excavator is not planned. The automation concept comprises three expansion stages:

• 2.1 Trajectory succession control of the working joints
• 2.2 Trajectory Sequence Control with Active Vibration Damping
• 2.3 Force control with or without switchable trajectory sequence controller even in contact situations (eg when digging)

The Reference trajectory of the working joints for the web control is inventively using calculated by a railroad planner. The module has two Operating modes: fully automatic operation and semi-automatic operation. In the latter mode, the operator controls the work tool manually with regard to a selectable Trajectory. The path planning module synchronizes the gameplay of the Drive joints such that the working tool the desired path takes the operator by hand to the desired Speed proportional signal (e.g., via joystick) of the tool pretends. From this information, the time indexed Joint positions and higher Derivatives such as joint speed, acceleration and pressure under consideration calculated from kinematic and dynamic limitations of the implement used. Possible Ambiguities of the solution thereby avoided by kinematic or dynamic constraints considered which are e.g. based on time or energy optimization are based. By way of example, the process of leveling with an excavator will be explained in more detail. The The task consists in the height profile of certain surface sections to straighten. It can be horizontal or oblique to be applied surfaces act. The operator uses his control command to enter the Speed and the direction of the tool. The railway planner in semi-automatic operation calculates the synchronized Reference joint angle of the jib that the loading shovel has an in height profile linear exerts radial movement. Will for the Duty a flat excavator (for example, leveling machines) is used for the Chassis (such as steering angle and driving speed) a reference track planned and calculated the temporally indicated boom position such that in the world coordinate system the desired height profile comes about.

in the Essentially, the railway planner differs for fully automatic operation for semi-automatic operation in that the movement of the tool automatically, without operator input, is planned. For this purpose is superimposed the desired one Position history of the tool based on the work task or extracted sequentially using a stored path (teach-in) and the required joint position calculated. Again, possible Ambiguities of the solution thereby avoided by kinematic or dynamic constraints considered become. In addition, the movement of the joints is under consideration calculated from maximum dynamic and kinematic limits. For this purpose, software modules for automatic trenching, handling and emptying developed. The functions can be individually or individually by the operator combined selected become.

Automatic digging:

specially for low skilled labor is the process of actually digging a time-consuming task as the working joints of the excavator arm synchronized with respect to avoiding filling losses need to be controlled. In addition, the location of the soil removal by the operator is often difficult to see. This requires a briefing by additional staff. From these establish it is advantageous to develop an automatic grave function. Based on the desired Digging depth, a path planning module calculates the time-indexed motion handle and spoon cylinder according to a maximum tear and breakout power. Furthermore, it must be independently recognized when the Shovel filled with soil is, so that the envelope movement can be initiated.

Automatic envelope:

The The envelope movement of the excavator consists of superimposed partial movements the excavator arm and the superstructure. So that the driver relieves and the dredging performance is increased, is the development of a function to the automated game play the pan and Rückschwenkenbewegung intended. This is the reference movement of the drive joints calculated with a path planning module. In the envelope operations it is a synchronized movement of the excavator arm and the spin mechanism.

The signal generated by the path planner Q _Ref includes in semi-automatic and fully automatic operation, the reference joint position and its derivative up to a desired order n:

2.1 Trajectory Succession the working joints:

als neuen Eingang v gesetzt und folgendermaßen berechnet:

According to the invention, the method for earth and bulk material movement includes a trajectory sequence controller. Here, the control signals of the hydraulic drive units are calculated in such a way that on the one hand follow the joint positions ground exactly the planned path Q _ref and on the other the work tool with respect to disturbances such as external forces or model errors, is stabilized. The control module relies on a flatness-based design method that is often described in the literature as a "computed torque method." In the equation of motion (1), the joint acceleration becomes

set as new input v and calculated as follows:

eingesetzt, lässt sich das für die Bewegung und Stabilisierung notwendige Moment bzw. Kraft u _Ref wie folgt berechnen:

The vectors K _D and K _P present the control parameters, which are determined, for example, with a pole specification. The poles of the pole specification polynomial are selected so that the system is stable, the control operates sufficiently fast and the manipulated variable limitations are not achieved with typically occurring control deviations. Is equation (17) in equation (1) with the condition

The moment or force u _Ref required for the movement and stabilization can be calculated as follows:

die inverse Anlenkkinematik. Hierbei ist F_{zyl_Refi} die Sollkraft des Zylinders an einem beispielhaft isoliert betrachteten Gelenk ,i'. Der dafür benötigte Öldruck wird mit der Druckgleichung von Formel (5) folgendermaßen approximiert:

The joint positions and the joint speed are either measured with a suitable sensor on the device or estimated by observers. The reference torque or the reference force u _Ref are provided by the hydraulic drive systems. If a cylinder drive is used, this is given by equation (4) and (5):

the inverse articulated kinematics. In this case, F _{zyl_Refi is} the nominal force of the cylinder on a joint, i 'considered as an example, isolated. The required oil pressure is approximated with the pressure equation of formula (5) as follows:

The claimed oil _flow Q _{Ref_i} is provided by equation (5) from a regulated subordinate _flow unit. Thus, the control signal with the aid of formula (5), (19) and (20) can be transformed

If a joint is operated with a hydraulic motor, the control signal of the subordinate flow unit is calculated from formula (7) and equivalent to equations (19) to (21), and results in a joint with the index 'j':

2.2 Trajectory Succession with active vibration damping:

Basically offer 2 basic design procedures: the decentralized variable-parameter Condition control and the flatness based trajectory tracking.

ever according to expression of nonlinearities And system conditioning can be beneficial to one way or another be. The controllers should be set according to defined criteria (eg pole specification) can be parameterized. For the Ground contact situation become switching rules of an impedance control developed. Depending on the system configuration can be clever formulation of the problem as MIMO system next to the position the force is used as a controlled variable become. Target of vibration damping is the active compensation of load or device vibration due to of mechanical and hydraulic elasticity phenomena occur. The basic idea is based on a feedback of the system states and an algorithm that calculates the drive manipulated variables from it and a non-vibrational trajectory tracking scheme of the referenced Rail guaranteed. For this purpose, mechanical deformations are measured with suitable sensors or observer supported based on measurable system states estimated. The deformation state variables can thereby consist of several shares, for example, by elasticities of the Boom sections, by distortion deformation of the chassis, by the sprung suspension suspension through Deformation phenomena of the suspension wheels, and by oil compression elasticities of the Actuators or the support caused. According to the invention these states with suitable sensors such as strain gages, ultrasonic sensors, Laser scanners, inertial sensors, etc. measured. Basically occurs the problem on that individual state variables are not measurable. accessible Measured variables are usually the axle positions. More problematic are the bends in the cantilever arms, which in principle according to the invention with the help an observer can be reconstructed. Other parameters are the pressure signals in the hydraulic control circuits. For cost reasons will be this also reconstructed observer-based.

2.2.1 Nonlinear Trajectory Sequence Control with subordinate status feedback

The starting point is the equation of motion (2). This is a so-called underactivated system because the order of system degrees of freedom (joint positions and degrees of elasticity) is greater than the number of drive elements. If the joint positions are considered as a so-called flat exit, the system is not differentially flat. The control design is realized according to the invention with an input / output linearization and a targeted damping of the residual dynamics. For this purpose, the model (2) after Isidori in the non-linear

Control normal form transformed:

Here, z _{1, i} with i = 1..n are the joint positions of the working machine. The index n corresponds to the number of drive units. The index r _i denotes the relative degree of the corresponding subsystem of (23). The vector n presents the states of zero dynamics, the last lines in equation (23) describing the dynamic behavior of the zero dynamics. The trajectory sequence control is based on the penultimate lines of the control normal form (23), since the system input variables appear there. These equations are transformed according to the manipulated variables u _i and the variable with the highest derivative order is replaced with an extended input. The signal of the extended input is calculated by means of a linear control law and a pilot control component. Deviations from the given position and, depending on the system configuration, also the path errors higher derivations are taken into account. By way of example, the following is the law of Trajektorienfolggegelung indicated when each subsystem has a relative degree of three:

Here, K _{i are} the control parameters that weight the position, speed and acceleration error. The dynamic trajectory tracking error can be described by the following equation:

The control parameters can be calculated with a pole specification of the desired control dynamics. The control signals u _i in equation (24) are converted into the corresponding control signal of the flow units according to the already explained principle (equation (19) to (22)).

The zero dynamics essentially describe the residual vibration remaining using equation (24). Since the zero momentum equations are coupled to the articulation states z, according to the invention, this residual dynamics (23) is selectively attenuated by the elasticity conditions are weighted with a control parameter and a control signal of the subordinate conveying power unit of the drive system is generated in terms of a negative feedback: Δy = - R · δ (26)

The vector Δ y denotes the control signal of the flow unit, R is the control parameters for setting the vibration damping and δ denotes the elastic degrees of freedom such as bend angle and / or compression deformation of the hydraulic oil.

The idea of damping the vibration behavior, which reflects the zero dynamics, with a linear state control is explained in more detail below with reference to a simple example. It is the task to dampen the horizontal bending vibration of a monobloc boom of an excavator with the help of a targeted rotational movement of the superstructure. The structure of the zero dynamics describes the following differential equation for the bend size -mlcosφ A · Φ .. D + m · ẅ H + b DL · ẇ H + c · w H = 0 (27) a weakly damped system. Here, m is the equivalent mass of the monoblock boom, which is located at a distance of 1 to the center of rotation. φ _A is the angle of elevation, φ _D is the angle of rotation of the upper carriage, w _h is the horizontal bend size, b _DL is the structural damping of the boom and c is the replacement spring constant of the boom deflection. The introduction of a subordinate linear state feedback leads to: Dy φD = -R φD w H (28)

R _φD is the freely selectable control parameter for vibration damping. The flow _signal Δy _φD leads to a proportional _flow : .DELTA.Q φD = K φD Dy φD (29)

Dynamic effects of the flow controller are neglected, but can generally be considered by a corresponding model adaptation. The DGL (27) is extended by a dynamic proportion of the flow controller. The rotational speed of the superstructure is proportional to the flow rate Δφ. D = ζ φD · .DELTA.Q φD (30)

With the proportionality constant ζ _φD , the acceleration _component thus becomes: Δφ .. D = -R φD K φD ζ φD · ẇ H (31)

Substituting equation (31) into equation (27) will do so -mlcosφ A · (Φ .. D + R φD K φD ζ φD · ẇ H ) + m · ẅ H + b DL · ẇ H + c · w H = 0 (32) and can be reshaped as follows: -mlcosφ A · Φ .. D + m · ẅ H + (b DL + mlcosφ A R φD K φD ζ φD ) · Ẇ H + c · w H = 0 (33)

The resulting attenuation can thus be changed with the control parameter R _φD : b Res (R φD ) = b DL + mlcosφ A R φD K φD ζ φD (34)

2.2.2 Decentralized control

und

geschätzt und mittels einer Vorwärtsaufschaltung kompensiert. In der Massenmatrix M(q,δ) wird die Nichtlinearität dadurch berücksichtigt, dass q,δ als veränderliche Systemparameter eingehen. Dadurch wird ein adaptives System und entkoppeltes System erreicht. Somit kann die Bewegungsgleichungen (2) in n Teilmodelle gespalten werden: ẋ i = A i(q,δ)x i + B i(q,δ)ui yi = C i(q,δ)x i i = 1..n (35) According to the invention, a decentralized control concept can also be implemented for vibration-damped path control. First, the nonlinear terms in the equation of motion (2)

and

estimated and compensated by means of a forward connection. In the mass matrix M ( q , δ ), nonlinearity is taken into account by taking q , δ as variable system parameters. This achieves an adaptive system and decoupled system. Thus, the equations of motion (2) can be split into n submodels: ẋ i = A i ( q . δ ) x i + B i ( q . δ u) i y i = C i ( q . δ ) x i i = 1..n (35)

Here, x _{i is} the state vector of the ith subsystem, which generally includes four states: joint position, velocity, deformation state, and strain rate. A _i , B _i , C _i are the system, input and output matrix of the ith subsystem. According to the invention, the vibration-damped trajectory control is realized with the aid of a robust adaptive state controller and an adaptive control law. Due to the feedback and the pilot control, Eq. (35) to: u i = - K i x i + S i w i ẋ i = ( A i ( q . δ ) - B i ( q . δ ) K i ) x i + B i ( q . δ ) S i w i y i = C i ( q . δ ) x i (36)

K _i is the matrix of the controller gains of the state controller with the entries k _1i , k _2i , k _3i , k _4i . S _i is the feedforward vector, which weights the setpoint trajectory w _{i of} the joint i, so that an accurate trajectory behavior is guaranteed. The dynamic behavior of the controlled system is determined by the position of the eigenvalues of the matrix A _R = A _i ( q , δ ) -B _i ( q , δ ) K _i , which are also poles of the transfer function in the frequency domain. The eigenvalues of the matrix can be determined by calculating the zeros with respect to the variable s of the characteristic polynomial p (s) from the determinant as follows: det (s I - A R ) ≡ 0 where p (s) = det (s I - A R ) = det (s I - A i ( q . δ ) + B i ( q . δ ) · K i ) (37)

wobei l die Systemordnung ist, die mit der Dimension des Zustandsvektors gleichzusetzen ist. Im Falle des Modells nach Gleichung (35) ist l = 4 und damit p(s): p(s) = (s – r1)(s – r2)(s – r3)(s – r4) = s4 + p3s3 + p2s2 + p1s + p0 (39) I is the unit matrix. By returning the state variables via the regulator matrix K _i . These eigenvalues can be shifted specifically to the control input. It is now required that the controller gains equation (37) take certain zeros, in order to influence specifically the dynamics of the system, which is reflected in the zeros of this polynomial. This results in a specification for this polynomial according to:

where l is the system order to be equated with the dimension of the state vector. In the case of the model according to equation (35), l = 4 and thus p (s): p (s) = (s - r 1 ) (s - r 2 ) (s - r 3 ) (s - r 4 ) = s 4 + p 3 s 3 + p 2 s 2 + p 1 s + p 0 (39)

The poles r _i are to be selected so that the system is stable, the control operates sufficiently fast with good damping and the manipulated variable limitation is not achieved with typical occurring control deviations.

The control gains can now by coefficient comparison of the polynomials equation (37) and (39) be determined.

The transfer function of the controlled subsystem becomes

W _i (s) denotes the output signal of the ith subsystem transformed into the image area. In order to calculate the pilot control _{amplifications} S _i (K _Vi0 to K _Vi4 ), the transfer function (41) is extended by the addition of the _reference variables:

An ideal system behavior with regard to the position, the velocity, the acceleration, the jerk and, if necessary, the derivative of the jerk results precisely if the transfer function of the system (42) of feedforward control and transfer function becomes identical to one. This leads to a linear system of equations, which can be solved in analytical form for the desired pilot control _gains K _Vi0 to K _Vi4 . The coefficients are also dependent on the known controller gains K _{i of} the state controller. Since the system structure (35) contains changing parameters, state controller and precontrol matrix are adaptively calculated and thus adapted to the variable system. The control signals u _i in equation (36) are calculated according to the already explained principle with the aid of equations (19) to (22) in the corresponding control signal of the flow units.

2.3 force control with or without switchable trajectory sequencer:

The Approximated driving equations (19) to (20) describe proportional behavior of desired force or nominal torque and that force Input signal of the delivery unit. This is just a control. model errors and model uncertainties of the drive units can therefore to force or Cause momentary errors. This can come with a bad behavior of the web regulation and with a slow vibration damping to make noticable. Especially affect in contact situations external forces the follow-up behavior. This makes itself e.g. noticeable when digging that the tool digs too slowly, or the tool jammed in the ground. On the other side is often in a contact situation, a track regulation undesirable. It becomes common demanded that the tool only have a defined force on the environment exerts so that e.g. safety guidelines are met.

For these reasons, a subordinate force regulator for the joints of the working machine to be controlled is inventively developed. Actuators are supplied with two delivery units on the implement. If it is a differential cylinder, each cylinder chamber is supplied separately with hydraulic oil. If it is an engine, each of the inlet and the outlet channel has a supply unit. The model of the drive units corresponds to equations (6) and (8) and represents a multi-variable system. The basis of the force control are two pressure regulators, which are derived below for a differential cylinder. The cylinder force of an exemplarily isolated drive joint with the index 'i' is given by equation (6): F refi = A 1i · p Refi_1 - A 2i · p Refi_2 (43)

Here, F _Refi denotes the target force of the cylinder, A _1i and A _2i are the end and annular surfaces of the cylinder, and p _{Refi_1} , p _{Refi_2} are the target pressures acting on the end and annular surfaces. Since there is an ambiguity of the resulting pressures to provide a desired force, an additional controlled variable is introduced. This can be, for example, a medium pressure, which can be calculated as follows:

die jeweils von der geforderten Kraft F_Refi und von einem gewünschten Mitteldruck p _Refi abhängen. Diese Drücke werden mit zwei Reglereinheiten, die auf der exakten Eingangs-/ Ausgangslinearisierung basieren, umgesetzt. Dazu wird die Druckdifferentialgleichung von Formel (6) nach dem gewünschten Ölstrom umgestellt: QRefi_1 = βVzyli_1(Zi(qi))·ṗRefi_1 + Azyli_1z .i(qi,q .i) QRefi_2 = βVzyli_2(Zi(qi))·ṗRefi_2 + Azyli_2z .i(qi,q .i) (46)und die Ableitung der Drücke ṗ_{Refi_1},ṗ_{Refi_2} als erweiterte Eingangsgrößen definiert: ṗRefi_1 = v1 ṗRefi_2 = v2 (47) It is p _Refi the specified medium pressure. If equation (44) is combined with equation (43), the reference pressures can be described

the desired each of the required force F _refi and by a medium pressure p _Depend on _Refi . These pressures are translated with two regulator units based on the exact input / output linearization. For this purpose, the pressure differential equation of formula (6) is changed to the desired oil flow: Q Refi_1 = βV zyli_1 (Z i (q i )) * P Refi_1 + A zyli_1 for example i (q i , q .i) Q Refi_2 = βV zyli_2 (Z i (q i )) * P Refi_2 + A zyli_2 for example i (q i , q .i) (46) and the derivative of the pressures ṗ _{Refi_1} , ṗ _{Refi_2 defined} as extended input variables: Ṗ Refi_1 = v 1 Ṗ Refi_2 = v 2 (47)

For the pressure profile to have a good follow-up behavior, the extended inputs are calculated as follows: v 1 = ṗ Refi_1 + R Drucki_1 (p i_1 - p Refi_1 v 2 = ṗ Refi_2 + R Drucki_2 (p i_2 - p Refi_2 ) (48)

Here are R _{Drucki_1} and R _{Drucki_2} the control parameters of the pressure regulator. The chamber pressures p _{i_1} and p _{i_2} are measured with suitable sensors or estimated with the help of an observer. Taking into account the proportional behavior between the input signal of the secondary flow regulator and the delivery volume (equation (6)), the output of the pressure regulator, taking into account equations (46) through (48), results in:

equation (49) presented the nonlinear subordinate force regulator. The derivation of a Moment control of a motor takes place in an analogous manner. The advantage is that a desired Force or a desired one Moment provided quickly and accurately at the work joint becomes.

The methods described in section 2.1 and 2.2 can be improved with the subordinate force control. In addition, disturbing forces can be specifically compensated. When digging with a wheeled excavator or when leveling with a caterpillar contact or digging forces occur, which have a negative impact when traversing the desired trajectory. In order to improve the follow-up behavior, the contact force is calculated with the help of the grave force model (15) or the grave force observer described in this type of operations. These estimated shares will then be forwarded to the Force controller switched and thereby compensated. Another advantage of lower-level force control is to increase the operational safety of the implement, since Equation (43) captures the force between the implement and the environment and can be used for monitoring purposes. If, for example, the loading shovel encounters an obstacle when digging, this leads directly to an increase in pressure in the drive element. This information can be used to interrupt the digging process. In addition, drive elements can be operated purely force-controlled. The advantage is that the device is not damaged even in the event of an unintentional collision since the force control is not feedback-free with regard to external force effects. In this regard, with the help of the choice of medium pressure p _Refi the power will be limited.

there the medium pressure level is chosen so that a maximum force to a manipulated variable limit of Force controller leads. Alternatively, the digging observer can also be used for surveillance purposes be used. Building on a knowledge base, the digging power dependent on the class of earth or bulk material includes, can with scheduled earth movements estimated force occurring become. Does the observer provide values that are outside this range? can safety-relevant situations such as the dredging of cables or manholes detected, and appropriate action be initiated.

Claims (11)

Mobile working machine, in particular a hydraulically driven earth-moving machine, comprising a motor, working equipment and a tool for loading, shoveling or gripping connected by working joints, characterized in that it comprises a path planner for generating its driving and working movement, suitable means for Detecting the position and / or speed of the working joints and at least one pressure sensor in the drive elements and / or strain gauge sensors for detecting the elastic bending and / or inertial and / or optical measuring means for a Trajektorenfolgeregelung the working joints with a combined active vibration damping and / or Contact force control for the digging force has. Mobile work machine according to claim 1, characterized that the path planner works in semi-automatic mode. Mobile work machine according to claim 1, characterized that the path planner works in automatic mode. Mobile work machine according to one of claims 1 to 3, characterized in that the Trajektorienfolgeregler a non-linear Trajektorienfolgeregler with a subordinate state feedback is. Mobile work machine according to one of claims 1 to 3, characterized in that the Trajektorienfolgeregler a decentralized Trajektorienfolgeregler with a subordinate state feedback is. Method for earth and bulk material movement with a mobile Work machine according to claim 1 to 5, characterized in that the dynamic and static behavior of the mechanical and static hydraulic system of the working machine in a physical Model, a reference trajectory for the working joints the working machine is calculated and the position and the position the working joints of the working machine are controlled exactly wherein in the physical model an active vibration damping and / or a subordinate contact force control is taken into account, wherein at the contact force control a maximum contact force is defined and when exceeded this threshold of the duty cycle is interrupted. Method according to claim 6, characterized that for determining the reference trajectory, the system states of Working joints by appropriate means, these means pressure sensors in the drive elements, strain gauge sensors for detecting the elastic Bending, inertial sensors or optical measuring devices or position sensors for the Can include working joints, recorded and returned to Trajektorienregeler and with the help of a Algorithm the actuator control variables for the working joints be determined from it. Method according to one of claims 6 or 7, characterized that the determined reference trajectory with a nonlinear Trajectory sequencer controlled with subordinate state feedback becomes. Method according to one of claims 6 or 7, characterized that the determined reference trajectory with a decentralized Trajektorienfolgeregler controlled with subordinate state feedback becomes. Method according to claim 6, characterized that for the contact force control the contact forces with a suitable static and / or dynamic model. Method according to claim 6, characterized that for the contact force control determines the contact forces by estimation become.