METHOD FOR CONTROLLING AN ARTICULATED ARM WITH A MOBILE REMOTE CONTROL UNIT LOCATED SPATIALLY DISTANT THEREFROM, AND SUCTION EXCAVATOR

Abstract

A method for controlling an articulated arm with a mobile remote control unit located spatially distant therefrom employs a machine coordinate system which is linked to the articulated arm, and an input coordinate system which is linked to the remote control unit. A deviation between the spatial orientation of the input coordinate system relative to the spatial orientation of the machine coordinate system is determined. A target movement direction and target movement speed of the end piece of the articulated arm in the input coordinate system are detected via control elements of the remote-control unit. The target movement direction is transformed into a transformed movement direction using the determined deviation. The transformed movement direction and the movement speed are transmitted to an articulated arm control unit for controlling a drive unit of the articulated arm.

Claims

1. A method for controlling an articulated arm with a mobile remote control unit spatially remote therefrom, comprising the following steps: defining a machine coordinate system (M.sub.W) which is linked to the articulated arm so that the position of at least one end piece at the free end of the articulated arm can be determined in this machine coordinate system; defining an input coordinate system (M.sub.I) which is linked to the remote control unit; determining a three-dimensional deviation between the 3D spatial orientation of the input coordinate system and the 3D spatial orientation of the machine coordinate system; detecting a target movement direction and target movement speed of the end piece of the articulated arm in the input coordinate system input via operating elements of the remote control unit; transforming the target movement direction into a transformed movement direction using the determined 3D deviation between the input coordinate system (M.sub.I) and the machine coordinate system (M.sub.M); transmitting the transformed movement direction and the movement speed to an articulated arm control unit, and controlling at least one drive unit of the articulated arm to move the end piece to the specified target position.

2. The method according to claim 1, wherein the input coordinate system of the remote control unit is defined by determining a reference plane, wherein the detected target movement direction is corrected in order to compensate for a deviation between the position of the vertical axis of the remote control unit and the reference axis located on the reference plane.

3. The method according to claim 2, wherein the reference plane is determined on the basis of a gravitation vector, wherein the detected target movement direction is corrected in order to compensate for a deviation between the position of the vertical axis of the remote control unit and the gravitational axis.

4. The method according to claim 1, wherein the target movement direction and the target movement speed of the end piece are detected as a target movement vector in the input coordinate system; and that the target movement vector is transformed into a transformed movement vector using the determined deviation between the input coordinate system and the machine coordinate system; and that the transformed movement vector is transmitted to the articulated arm control unit.

5. The method according to claim 4, wherein spherical coordinates of the target position are calculated based on the transformed movement vector in the machine coordinate system.

6. The method according to claim 1, wherein the transformation from the input coordinate system to the machine coordinate system is carried out taking into account the determined three-dimensional deviation only if the three-dimensional deviation between the 3D spatial orientation of the input coordinate system and the 3D spatial orientation of the machine coordinate system is confirmed by measurements with at least two independent measuring systems.

7. The method according to claim 1, wherein to define the input coordinate system, its orientation in relation to the machine coordinate system is measured, preferably using one or more measuring systems from the following list: optical measuring systems with which passively or actively illuminated markers can be detected; inertial sensors that can be used to determine the gravitation vector and the Earth's magnetic field.

8. The method according to claim 1, wherein in order to define the input coordinate system, the relative rotation about the gravitational axis of the input coordinate system and the machine coordinate system is derived from a position measurement of at least three points, preferably using one or more measuring systems from the following list: laser-based position measuring systems, preferably with pulsed light; optical measuring systems with which passively or actively illuminated markers can be detected; stereo cameras that are preferably arranged on the machine unit carrying the articulated arm or on stationary construction site furnishings; one or more cameras that are arranged on the remote control unit or on stationary construction site furnishings.

9. The method according to claim 1, wherein position measurements are carried out to determine the position of the end piece, preferably using one or more measuring systems from the following list: optical measuring systems with which passively or actively illuminated markers can be detected; laser-based position measuring systems, preferably with pulsed light; stereo cameras that are preferably arranged on the machine unit carrying the articulated arm; one or more cameras that are arranged on the remote control unit; a mechanical measuring system on the articulated arm, preferably with rotary encoders; inertial sensors on the joints of the articulated arm and/or on the remote control unit.

10. The method according to claim 1, wherein the Earth's magnetic field is measured and taken into account to define the input coordinate system and the machine coordinate system.

11. The method according to claim 1, wherein the articulated arm is a component of one of the following devices: suction excavator; concrete pump; sewer cleaning machine; sewer inspection machine; drill; lifting platform.

12. A suction excavator with a vehicle frame, a fan unit for generating a suction flow for picking up material, a filter unit, a material collection container for collecting the picked-up material, a multi-link articulated hose carrier, and with a mobile remote control unit for controlling the movement of the articulated hose carrier, characterized in that a control unit of the suction excavator and the remote control unit are configured to carry out a method of claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Further details, advantages and developments of the invention are apparent from the following description of preferred embodiments, with reference to the drawing. Shown are:

[0062] FIG. 1 is a symbolic representation of a suction excavator that is being operated on a construction site by a user with a remote control unit;

[0063] FIG. 2 is a symbolic first representation of an articulated arm and a remote control unit for carrying out a method according to the invention for controlling the articulated arm;

[0064] FIG. 3 is a symbolic second representation of the articulated arm to illustrate the position of an end piece relative to a root joint;

[0065] FIG. 4 is a graph representation of possible relations between a dynamic input coordinate system of the remote control unit and a stationary machine coordinate system of the articulated arm with the position of the end piece;

[0066] FIG. 5 is a flow chart of a process chain for calculating all angles of the articulated arm from a target movement vector in the dynamic input coordinate system;

[0067] FIG. 6 is a symbolic third representation of the articulated arm to illustrate the division of angles between the links of the articulated arm;

[0068] FIG. 7 is an illustration of the concatenation of vectors to calculate a point and its distance from the root joint.

DETAILED DESCRIPTION OF THE INVENTION

[0069] FIG. 1 illustrates a typical application situation in which a suction excavator 10 with an articulated arm 01 is used on a construction site. A stationary machine coordinate system M.sub.M of the suction excavator 10, an end effector coordinate system M.sub.E of an end effector 04, and a dynamic input coordinate system M.sub.I of a remote control unit 02 are rotated three-dimensionally with respect to each other and also deviate from a world coordinate system M.sub.W based on the gravitational vector and the Earth's magnetic field. An alternative reference coordinate system can be obtained by measuring three reference points P.sub.1, P.sub.2, P.sub.3 which are attached to construction site furnishings such as construction fences.

[0070] FIG. 2 shows a schematic diagram of the articulated arm 01, which in the embodiment considered as an example below is an articulated hose carrier of a suction excavator (FIG. 1). The articulated arm 01 has a plurality of articulated arm links L.sub.n that are connected to each other via joints J.sub.n. The remote control unit 02 is provided spatially separated from the articulated arm 01, with which unit a user 03 can control the desired movements of the articulated arm 01. In telematics applications, this can also be out of sight and out of direct range, i.e. at any distance.

[0071] The remote control unit 02 and an articulated arm control (not shown) cooperate to carry out the method according to the invention for controlling the articulated arm. Ultimately, the goal is to move the end effector 04 (also called the end piece), which is located at the free end of the articulated arm 01, to a desired target position in order to perform a work task there. In the case of suction excavators, this work task usually involves collecting material, e.g. excavated soil, using the negative pressure generated by a fan unit of the suction excavator, and transporting the material through a suction hose carried by the articulated arm into a material collection container.

[0072] In FIGS. 1, 2, 3, 6, 7, coordinate system symbols are shown for easier understanding, wherein coordinate system symbols without arrowheads only represent an orientation but not a relevant position.

[0073] The articulated arm links L.sub.n rotate around the joints J.sub.n. The orientation of the articulated arm 01 is determined in the stationary machine coordinate system M.sub.M, while the orientation of the remote control unit 02 is defined in the dynamic input coordinate system M.sub.I.

[0074] FIG. 3 also shows the basic structure of the articulated arm 01 according to FIG. 2. The angle ranges shown here serve primarily to indicate the position of the end piece 04. The end piece 04 is attached to the last joint J.sub.E and can also be understood as an end effector whose position P.sub.E lies at the last joint J.sub.E. The movement of the position P.sub.E of the end effector is shown in FIG. 3 in spherical coordinates (, , r) relative to the root joint J.sub.0 or also to the first joint J.sub.1 which can only be rotated about the Z-axis on the suction excavator relative to the root joint J.sub.0 (no change in angle between J.sub.0 and J.sub.1). The following explanations of the execution of the method also refer to this type of representation.

[0075] The following assumptions are made for the functional implementation of the method for controlling the articulated arm in the embodiment of the articulated hose carrier of a suction excavator: [0076] a) The articulated arm 01 consists exclusively of one-dimensional rotation joints J.sub.n, where all joints J.sub.1 to J.sub.i are identically oriented, and an additional rotation with an axis of rotation rotated by 90 is only possible at the root joint J.sub.0 or J.sub.1. [0077] b) The movement of the end effector P.sub.E relative to the root joint J.sub.0 or J.sub.1 can be defined in spherical coordinates (, , r), wherein the azimuth angle is determined exclusively by the angle of the root joint J.sub.0 at the suspension of the arm, and the angles .sub.n of all other joints J.sub.1 to J.sub.i jointly determine the length (or the spherical radius r) and the polar angle (see FIG. 3). [0078] c) The ratios of the individual joint angles .sub.2 to .sub.i are predefined by weights w and offsets o (e.g. uniformly distributed) so that the radius r, i.e. the distance of the end effector position P.sub.E to the root joint J.sub.1, can be determined by specifying a single reference angle :

[00001]${}_n=w_n*+o_n$ [0079] where in the following we assume an equal distribution of the angles .sub.2 to .sub.5, i.e.: w.sub.n=1 and o.sub.n=0. [0080] d) The orientation of the input coordinate system M.sub.I of the remote control unit and of the machine coordinate system M.sub.M are jointly defined in a higher-level coordinate system (here world coordinate system M.sub.W) (see FIG. 4a). Alternatively, M.sub.I can also be defined in M.sub.M (FIG. 4b), or M.sub.M can be defined in M.sub.I (FIG. 4c). In addition, the position of the end effector P.sub.E has to be defined in the machine coordinate system M.sub.M. The following descriptions are based on a spatial structure (FIG. 4a). M.sub.W does not have to indicate an original position; a reference frame for orientation is sufficient, e.g. based on gravity and the north pole of the Earth's magnetic field (see FIG. 2 or FIG. 3). Alternatively, a reference coordinate system can be determined by measuring at least three reference points (see P.sub.1-P.sub.3 in FIG. 1).

[0081] FIG. 4 shows possible relations between the input coordinate system M.sub.I and the machine coordinate system M.sub.M with the position of the end effector P.sub.E as a graph.

[0082] As already explained above, the controlling of at least one drive unit of the articulated arm 01 for moving the end piece 04 or end effector P.sub.E to a target position specified by the transformed movement sector using previously known controlling can be done as described for example in DE 10 2016 106 427 A1. Such controlling can also be called inverse kinematics since it always controls the individual joints depending on the target position of the end piece. A possible technical implementation of this inverse kinematics in an articulated hose carrier of a suction excavator can be carried out as follows: [0083] 1. The control commands from the remote control unit are first processed algorithmically to indirectly manipulate the oil pressure in the hydraulic cylinders to move the articulated arm links, resulting in controlled movements of the end effector. [0084] 2. The articulated arm consists exclusively of one-dimensional rotation joints, wherein all joints are identically oriented, and only the root joint J.sub.0 has an axis of rotation rotated by 90. [0085] 3. The movement of the articulated arm can be defined in spherical coordinates, wherein the azimuth angle is determined exclusively by the angle of the root joint J.sub.0 at the suspension of the arm, and the angles of all other joints J.sub.n together determine the length (or the spherical radius r) and the polar angle . [0086] 4. The ratios of individual joint angles are predefined (e.g. evenly distributed) so that the desired arm length r can be determined by specifying a single angle value. [0087] 5. The angles of the articulated arm links are detected simultaneously using different sensors and measuring methods in order to eliminate the given systematic measurement errors. Preferably these are two or more of the following sensors: [0088] a. angle sensors in the joints of the multi-link articulated arm; [0089] b. linear position sensors in the hydraulic cylinders; [0090] c. inertial sensors for measuring the gravitation vector; [0091] d. camera-based or laser-based sensors for the absolute measurement of position and orientation of the individual articulated arm links, including the end effector, relative to an external measuring station, e.g.: [0092] i. on the machine, [0093] ii. mobile on a stand or integrated into construction site furnishings such as fences, [0094] iii. mobile on the remote control unit. [0095] 6. The relative orientation of the articulated arm and the remote control unit is detected by a combination of sensors in order to eliminate systematic measurement errors here as well. These are preferably: [0096] a. 3D inertial sensors in or on the remote control unit and on the articulated arm; [0097] b. redundant 3D inertial sensors with the greatest possible distance and immovable mechanical connection to detect and evaluate interference effects of local magnetic fields on the electronic compasses. [0098] c. camera-based or laser-based sensors for absolute measurement of the orientation of the remote control unit, articulated arm and end effector relative to each other or relative to an external measuring station, e.g.: [0099] i. on the machine, [0100] ii. mobile on a stand or integrated into construction site furnishings such as fences, [0101] iii. mobile on the remote control unit. [0102] 7. The position of the end effector is detected simultaneously using two measuring methods to detect systematic measurement errors. These are preferably: [0103] a. mechanical measurement of the end effector based on the orientation of all links of the articulated arm; [0104] b. camera-based or laser-based sensors for absolute measurement of the orientation of the remote control unit, articulated arm and end effector relative to each other or relative to an external measuring station, e.g.: [0105] i. on the machine, [0106] ii. mobile on a stand or integrated into construction site furnishings such as fences, [0107] iii. mobile on the remote control unit.

[0108] FIG. 5 shows a flow chart of the process chain for calculating all angles .sub.n of the articulated arm 01 from a target movement vector {right arrow over (V.sub.I)} detected in the remote control unit 02 in the input coordinate system M.sub.I. The control commands of the remote control unit 02 are processed in the sequence shown in FIG. 5 in order to ascertain all target angles .sub.n of the joints J.sub.n so that a controlled movement of the end effector P.sub.E along a transformed movement vector results. In so doing, the target movement vector is transformed into the transformed movement vector using the previously determined deviation between the input coordinate system M.sub.I and the machine coordinate system M.sub.M. One possibility of this transformation is explained in detail below for the case of mapping M.sub.M and M.sub.I in a common reference coordinate system M.sub.W (see FIG. 4a): [0109] I. Leveling (optional): the target movement vector is given in the input coordinate system M.sub.I of the remote control unit. Before the transfer (transformation) of into the machine coordinate system M.sub.M, the input coordinate system M.sub.I is aligned or leveled according to the previously ascertained gravitation vector so that only the rotation of the remote control unit 02 about the gravitational axis needs to be taken into account. For this purpose, a new leveled input coordinate system M.sub.I-U is constructed in the following sub-steps: [0110] 1. First, it is checked whether the input coordinate system M.sub.I is inclined by less than 90 to the gravitation vector , i.e. that the scalar product of a unit vector along the z-axis of the input coordinate system =(0, 0, 1) with the inverse of the normalized gravitation vector =()/ in a common world coordinate system M.sub.W is less than zero, i.e. the two point in different directions:

[00002]$\stackrel{}{G}*\left(M_I*\right)<0$ [0111] (under the assumption that is already defined in the world coordinate system M.sub.W).

[0112] Otherwise, the remote control unit is tilted downwards, and no clear interpretation of the input vector is possible. In this case, the control of the articulated arm should be interrupted. [0113] 2. If the precondition *(M.sub.I*)<0 is fulfilled, the axes of the leveled input coordinate system M.sub.I-U are constructed by calculating cross products between the x- or y-axis of the input coordinate system and the gravitation vector (in the common world coordinate system M.sub.W) (here using the example of the y-axis, i.e. a unit vector along the y-axis ).

[00003]$=\left(M_I*\stackrel{}{y_I}\right){\hat{G}}^{-1}={\stackrel{}{G}}^{-1}={\stackrel{}{G}}^{-1}$ [0114] 3. To accordingly level the target movement vector , it is simply expressed with identical values in the leveled input coordinate system M.sub.I-U.

[00004]$\stackrel{.Math.}{V_{I-U}}=\stackrel{.Math.}{V_I}$ [0115] II. Input transformation: The input vector V.sub.I or the leveled input vector V.sub.I-U can now be expressed in the machine coordinate system using the following calculation rule:

[00005]$\stackrel{.Math.}{V_{I-M}}=M_M^{-1}*\left(M_{I-U}*\stackrel{.Math.}{V_{I-U}}\right)$ [0116] III. New target position: if the current position of the end effector is known as a point P.sub.E in the machine coordinate system M.sub.M, the new target position P.sub.E can be calculated by being moved along the transformed movement vector in the machine coordinate system.

[00006]$P_E^{}=P_E+\stackrel{.Math.}{V_{I-M}}$ [0117] IV. Spherical coordinates: the target position of the end effector must be converted into spherical coordinates

[00007]$r=\sqrt{x^2+y^2+z^2}$$=\{\begin{array}{c}{\cos }^{-1}\frac{x}{\sqrt{x^2+y^2}}\mathrm{for}y0\\ 2*-{\cos }^{-1}\frac{x}{\sqrt{x^2+y^2}}\mathrm{for}y<0\end{array}$$={\tan }^{-1}\frac{z}{\sqrt{x^2+y^2}}$

[0118] The orientation of the machine coordinate system must be taken into account, and the resulting angle values must be shifted by a multiple of /2 as necessary. Alternatively, all three values of the spherical coordinates can also be ascertained by vector calculations. The radius r, or the distance of the target position P.sub.E from the root joint J.sub.1, is the length of the vector between the two points.

[00008]$r=.Math.\stackrel{.Math.}{J_1{P}_E^{}}.Math.$

[0119] The pivot angle .sub.0= is the scalar product of a unit vector along a reference axis in the machine coordinate system M.sub.M (e.g. the x-axis in FIGS. 2 and 3) and the normalized projection of onto the horizontal plane of the machine coordinate system M.sub.M (e.g. the x/y plane in FIGS. 2 and 3). For the projection of onto the desired plane, the vector component of the dimension to be ignored (e.g. z) can be set to zero. The projection can be written using cross products, e.g.:

[00009]$=\left(\left(\right)\right)*-1*$ [0120] is the scalar product of a unit vector along a reference axis in the machine coordinate system M.sub.M (e.g. the z-axis in FIGS. 2 and 3) and the normalized vector in the machine coordinate system M.sub.M.

[00010]$=*$

[0121] The pivot angle is already given as a result of this method step:

[00011]${}_0=$ [0122] If P.sub.E=J.sub.E, the vector divides the angle .sub.1 into the components .sub.1a and .sub.1b as well as .sub.E into .sub.Ea and .sub.Eb (see FIG. 6), where the following holds:

[00012]${}_{1a}=-$${}_{1b}={}_1-{}_{1a}$ [0123] V. 2D inverse kinematics: the calculation of the angles .sub.1 bis .sub.i can be solved in a two-dimensional coordinate system since all joints J.sub.1 to J.sub.i lie on the same plane and rotate around parallel axes. The sizes of the angles .sub.1 bis .sub.i together with the lengths of the adjacent links L.sub.1 to L.sub.i define the length of the vector . What is sought are the angles .sub.1 bis .sub.i with which the following applies:

[00013]$.Math.\stackrel{.Math.}{J_1{J}_E}.Math.=.Math.\stackrel{.Math.}{J_1{P}_E^{}}.Math.$

[0124] An analytical solution only exists in special cases. As a generic solution path for a virtually arbitrary number of links, varying ratios of the angles .sub.2 bis .sub.i, and different lengths of the adjacent links L.sub.1 to L.sub.i, the following possible solution path is described: [0125] 1. The geometric relationships of the length-relevant links L.sub.1 to L.sub.i are expressed in isolation in an independent 2D coordinate system, where L.sub.1 is aligned to the x-axis (since .sub.1 has no influence on the vector length ; see FIGS. 6 and 7). [0126] 2. Each of the length-relevant links L.sub.1 to L.sub.i is now expressed as a 2D vector in this coordinate system and rotated corresponding to the angle .sub.n with .sub.n=.sub.n (see FIG. 7). For L.sub.1, .sub.1=0 applies here since L.sub.1 is aligned along the x-axis.

[0127] The coordinates of the vectors to are each calculated as follows (with l.sub.1 to l.sub.i as the lengths of the links L.sub.1 to L.sub.i):

[00014]$x_{L_2}=l\cos \left({}_1\right)$$y_{L_1}=l\sin \left({}_1\right)$$x_{L_2}=l\cos \left({}_1+{}_2\right)$$y_{L_2}=l\sin \left({}_1+{}_2\right)$$.Math.$$.Math.$$x_{L_i}=l\cos \left({}_1+{}_2+.Math.+{}_i\right)$$y_{L_i}=l\sin \left({}_1+{}_2+.Math.+{}_i\right)$ [0128] 3. From the concatenation of the resulting 2D vectors to , a point P.sub.R results (see FIG. 7).

[00015]$P_R=\stackrel{.Math.}{L_1}+\stackrel{.Math.}{L_2}+.Math.+\stackrel{.Math.}{L_l}$ [0129] 4. The next task is to find the appropriate values .sub.n for which the distance corresponds to the target distance . All values .sub.n are defined by a common reference angle , because .sub.n=.sub.n and .sub.n differ from only by predefined weights w.sub.n and offsets o.sub.n.

[00016]${}_n=w_n*+o_n$

[0130] To search for the right , a binary search algorithm is used. In addition to the global parameter limits .sub.min and .sub.max, here local limitations .sub.n-min and .sub.n-max also have to be taken into account. As needed, local weights w.sub.n and offsets on enable an optimization of the range of motion of the entire articulated arm. [0131] 5. From the found value of , taking into account the local weights w.sub.n and offsets o.sub.n, all angle values from .sub.2 to .sub.i can now be derived. .sub.1 is composed of .sub.1a, which was already found in step IV during the translation into spherical coordinates, and .sub.1b. The latter is the angle, i.e. the scalar product between the normalized vector and a unit vector along the x-axis of the auxiliary coordinate system used here:

[00017]${}_{1b}=*\stackrel{}{x}$ [0132] VI. Alignment of end piece: to calculate the last remaining angle .sub.E, after ascertaining and the individual vectors to , a part is also already calculable:

[00018]${}_{Ea}=*$

[0133] .sub.Eb can be described as a scalar product of and a vector in the target orientation of the last link . Since the latter is defined relative to a vector in the reference coordinate system, e.g. the gravitation vector , we use its normalized representation in the machine coordinate system M.sub.M as a reference:

[00019]${}_{Eb}=*\left(\stackrel{}{G}*M_M^{-1}\right)$

[0134] Desired deviations of the orientation of from the gravitation vector can then be directly offset against the ascertained angle .sub.E.

[0135] FIG. 6 shows the division of .sub.1 into .sub.1a and .sub.1b and of .sub.E into a .sub.Ea and .sub.Eb by the vector . The length of the vector is determined by the lengths of the links L.sub.1, L.sub.2, L.sub.3 and the enclosed angles .sub.2 and .sub.3.

[0136] FIG. 7 shows the concatenation of the vectors to for the calculation of a point P.sub.R and its distance from the root joint J.sub.1.