Patent Application
20230166947
This application claims priority under 35 U.S.C. § 119(a) to Austria Application No. A50961/2021 filed Nov. 30, 2021, the disclosure of which is expressly incorporated by reference herein in its entirety.
The present invention relates to a method for controlling a lifting device that moves a load along a first direction of movement and along a second direction of movement within a specified working area of the lifting device in accordance with a specified trajectory from a starting point to an end point.
The planning of suitable trajectories is often an essential part of operational and problem-solving concepts in a wide variety of technical applications. Here, a trajectory is understood to be the temporal profile of the movement of a body along a curve, a path or a way, which in the case of a rigid body can be described, for example, by the position profile of its center of gravity. The common use of the term “trajectory” in control engineering as a temporal (reference) profile of state or output variables of a (technical) system to be controlled deviates from this definition, but does not contradict it. If a technical problem requires on the one hand the planning of a movement of a physical body, but on the other hand also the realization of the planned movement, for example by suitable control or regulation of drive elements provided for this purpose, the mentioned meanings of the term “trajectory” coincide, which is the case in many applications and also in the context of the following explanations.
Trajectory planning as described above is a basic sub-task in the automation of vehicles, such as passenger cars or trucks. With the help of map- or sensor-based data, trajectories, which primarily describe the movement of a vehicle on a road, are first planned and then implemented by the control of the vehicle as set point variables. In addition to convenience aspects, feasibility and collision-free nature of a trajectory are often important criteria when implementing this approach.
Another area in which the topic of trajectory planning is of great importance is the field of transport logistics. In order to increase the efficiency of logistics processes, among other things, rapid goods turnover is required in this division. This results in particular in requirements for rapid loading and unloading operations of cargo ships, as well as correspondingly rapid movement operations of lifting devices used for loading and unloading. Such lifting devices are also increasingly being operated in automated mode, wherein demands for fast movement operations are directly reflected in the trajectories to be implemented in (automated) operation.
Lifting devices are available in a wide variety of embodiments, which are accordingly used in a wide variety of applications. For example, there are tower cranes, which are mainly used for building construction and civil engineering, or mobile cranes, e.g., for the assembly of wind turbines. Furthermore, bridge cranes are used e.g., as indoor cranes in factory halls, or gantry cranes e.g., for the manipulation of transport containers at handling locations for the intermodal handling of goods. Goods to be transported are mainly stored in standardized containers, so-called ISO containers, which are equally suitable for the transport modes road, rail and water. Objects to be transported by a lifting device, such as a group of containers with goods contained therein, are hereinafter referred to simply as “load” or “loads”.
The design and mode of operation of gantry cranes in particular are well known and are described, for example, in US 2007/0289931 A1 on the basis of a “ship-to-shore crane” (STS crane). A gantry crane has a supporting structure, or gantry, on which a boom is arranged. Here, the gantry with wheels is movably arranged on a track, for example, and can be moved in one direction. The boom is fixed to the gantry and usually has a trolley that moves along the boom. To pick up a load, such as an ISO container, the trolley is usually connected by four ropes to a load receiving element known as a spreader. To pick up and manipulate a container, the spreader can be raised or lowered by winches. The spreader can also be adapted to containers of different sizes.
Typically, cranes are operated by a crane operator, who usually controls the movement of a crane from a cabin. Such a cabin can be arranged on a gantry or on a movable trolley. For precise and collision-free manipulation of loads, high demands are placed on the crane operator, which is why a training period of at least one year is usually necessary. In particular, the rapid movement of loads with small pendulum movements represents a highly complex activity, which is why several years of practical experience is usually required to successfully perform the job of a crane operator. The job of a crane operator is often made more difficult by a high level of physical stress, due among other things to the high level of concentration required in conjunction with sitting for long periods with the gaze tilted downward. For these reasons, among others, it is a declared goal of automation technology to at least partially automate the complex manipulation procedures of a lifting device and thus simplify them for the crane operator. Inseparably linked to the automated operation of lifting device is the planning of suitable trajectories to be implemented by the lifting device in automated operation.
One issue that has received little attention to date, but is becoming increasingly important, is the replanning of existing trajectories. An example relevant to practice, in which the requirement for trajectory replanning increasingly occurs, is the STS (“Ship-to-Shore”) cranes already mentioned. As mentioned, the primary purpose of STS cranes is to load containers from cargo ships onto trucks for onward transportation or onto so-called “automated guided vehicles”. It is often the case that a truck has not yet arrived in an area that can be reached by the lifting device or crane, but the movement of the load must already be started for reasons of efficiency. For obvious reasons, the actual position of a truck for load pickup can only be recorded and subsequently taken into account after its arrival. In order to be able to start the movement of a container or a load even before the truck arrives, it is necessary in such cases to first assume a suitable target position for the planning of a first trajectory and to adjust the assumed target position to the true target position later in the course of replanning.
The same applies to obstacles that must be avoided by the load during the movement of the load. An obstacle here can be another ship, a stack of containers, for example on a ship or on land, or any other obstacle. In this respect, it can be advantageous for efficiency reasons to start moving a container or a load even before the dimensions of an obstacle are known, for example. Also, an obstacle may occur only during the movement of a load. In such cases, it is also necessary to make assumptions first, particularly regarding the number and shape of obstacles to be considered. Based on these assumptions, an initial trajectory can be planned, which is to be adapted to the actual obstacles in the course of a later replanning. Thus, it is necessary to replan a trajectory depending on changed obstacles.
Although the topic of trajectory planning for lifting device is addressed in the prior art, issues regarding trajectory replanning, particularly due to changing obstacles, have hardly been mentioned in the relevant literature so far.
For example, EP 3461783 B1 describes the determination of a trajectory for the movement of a load by a lifting device. For this purpose, a geometric path on the one hand and a dynamic path on the other hand are calculated, which are combined in a further step to generate a trajectory. A disadvantage of this approach is the large computational effort associated with it. A (real-time capable) replanning during the operation of a lifting device is thus only possible to a limited extent.
CN 111170153 A, on the other hand, describes a method for replanning a given trajectory on the basis of obstacles detected by measurement. The necessary measurement requirements represent a significant complication for the implementation of this method. Furthermore, CN 111170153 A does not describe how obstacles that change during operation can be taken into account and, moreover, does not make any statements about the real-time capability of the method described.
In addition, the known prior art shows a plurality of further disadvantages, such as the usually given necessity that no changes of direction along the given movement directions must result from a replanning. A further essential point is the real-time capability of concepts for trajectory replanning. Since the described replanning usually has to be done during the operation of a lifting device, the computational effort associated with replanning must not interfere with the operation of a lifting device. This aspect is also not considered in the prior art.
Accordingly, it is an object of the present invention to provide an improved trajectory replanning method for a lifting device that enables efficient replanning of a trajectory along which a load is moved, taking into account a variable obstacle.
This object is solved by the present invention. In this case, the lifting device is assumed to be one that moves a load along a first movement direction and along a second movement direction within a specified working area of the lifting device in accordance with a specified trajectory from a starting point to an end point.
According to the invention, for this lifting device, at a command time point during the movement of the load, a new obstacle for the movement of the load, which is arranged between the position taken on by the load at the command time point and the end point, is specified for this lifting device and/or an existing obstacle, which is arranged between the position taken on by the load at the command time point and the end point, is changed to a new obstacle for the movement of the load, wherein, taking into account specified kinematic constraints of the lifting device, at least one individual movement is planned for the first movement direction and for the second movement direction respectively, which movement determines the further movement of the load along the respective movement direction from the command time onwards, wherein for each movement direction at least one individual movement ends in a projection of the end point onto the respective movement direction, and wherein the planned individual movements are executed in accordance with a specified movement sequence in order to move the load further along the first movement direction and along the second movement direction in accordance with a replanning trajectory resulting from the movement sequence of the planned individual movements without colliding with the obstacle. By planning individual movements, the numerical complexity is significantly reduced herein compared to methods known from the prior art, enabling efficient and flexible replanning even during operation of a lifting device.
In an advantageous manner, the individual movements planned for the first movement direction and those planned for the second movement direction are planned as independent individual movements, which means that no mutual dependencies need to be taken into account and the actual replanning task can be simplified even further.
In a further advantageous embodiment of the present invention, the position profiles of the load specified by the planned individual movements are specified as position profiles that can be continuously differentiated in time at least four times. In this manner, swinging of the moving load can be avoided in a remarkable way due to the principle.
In a further advantageous embodiment of the present invention, the position profiles specified by the individual movements are filtered by a filter with a predeterminable time constant in order to generate the temporally at least four times continuous differentiability. In this manner, it can be ensured in a particularly advantageous manner that predefined kinematic constraints of the lifting device are adhered to. The time constant of the filter can here be made dependent on the geometry of the lifting device in an advantageous manner to enable precise matching to the given lifting device.
In a further advantageous embodiment of the present invention, a second individual movement is planned for the second movement direction to move the load around the newly specified and/or the modified obstacle. It should be noted here that in order to go around an obstacle, it is usually necessary to combine the individual movement planned for going around with other individual movements planned along the other movement directions, i.e., to suitably combine horizontal and vertical movements. In this case, it may be necessary to coordinate or adapt the new individual movement to any other individual movements that may already exist. In the described manner, the load can be moved around the obstacle and collisions of the load with the specified obstacle can be avoided.
In a further advantageous embodiment of the present invention, the sequence of movements mentioned provides for an alternating start and, if necessary, overlapping execution of the planned individual movements along the first movement direction and of individual movements along the second movement direction, as a result of which, on the one hand, obstacles can be avoided and, on the other hand, time can be saved to a partly significant extent.
Furthermore, prior to the execution of at least one planned individual movement, the method according to the invention allows checking whether the expected replanning trajectory leads to a collision of the load with an obstacle specified in the working area. This check can be performed in an advantageous manner by comparing the collision times at which the respective individual movements reach projections onto the corresponding movement direction of a keypoint derived from the specified obstacle, wherein a numerical root finding method can be used to determine these collision times at which the respective individual movements reach projections onto the corresponding movement direction of a keypoint derived from the specified obstacle. In this manner, obstacles in the working area of the lifting device can be included in the replanning according to the invention, and possible collisions can be reacted to in good time. It should be noted that the procedure according to the invention allows the described check to be performed in a particularly efficient manner.
If a possible collision is detected by the described anticipatory calculation, at least one individual movement cannot be executed and instead, along the movement direction for which the at least one non-executed individual movement has been planned, a specified braking operation can be executed for at least the duration of a specified minimum braking time. Thus, on the one hand, collisions can be avoided in an advantageous manner, and on the other hand, a new starting situation is created in the described manner, from which a renewed replanning of individual movements is made possible.
According to the preceding explanations, a new individual movement can be planned for the corresponding movement direction in an advantageous manner instead of the at least one individual movement that has not been executed, wherein a replanning trajectory to be expected by the at least one new individual movement is determined and it is checked whether the replanning trajectory to be expected leads to a collision of the load with an obstacle specified in the working area, and wherein the new individual movement is executed if no collision is detected, or the specified braking operation is continued for at least one further minimum braking time, and a replanning of an individual movement as well as a check of the replanning trajectory, which is to be expected by the newly planned individual movement, is performed again if a new collision with an obstacle is detected. In this manner, the method according to the invention can be continued without collision, even if a first replanning would have led to a collision, and it can be ensured that there is no collision of the transported load with obstacles during the entire movement of the load.
Further, in support of the present invention, the position of the load can be measured and the measurement of the position of the load can be used in performing the individual movements. The use of measurement data is usually particularly advantageous for the implementation of planned individual movements by control technology.
Furthermore, it should be noted that the method according to the invention is by no means limited to movements in a plane, and that movements in (three-dimensional) space can also be planned by the present invention. In the case of more than two dimensions, the properties of the inventions at issue have an even more advantageous effect. Here, the independence of the planned individual movements should be especially emphasized, in particular since three interdependent coordinates would further increase the complexity to be mastered in replanning—compared to the already challenging 2D case.
Other exemplary embodiments and advantages of the present invention may be ascertained by reviewing the present disclosure and the accompanying drawing.
In the following, the present invention is described in greater detail with reference to
The particulars shown herein are by way of example and for purposes of illustrative discussion of the embodiments of the present invention only and are presented in the cause of providing what is believed to be the most useful and readily understood description of the principles and conceptual aspects of the present invention. In this regard, no attempt is made to show structural details of the present invention in more detail than is necessary for the fundamental understanding of the present invention, the description taken with the drawings making apparent to those skilled in the art how the several forms of the present invention may be embodied in practice.
A running element 5 is usually arranged on this boom 4, which is movable in the longitudinal direction of the boom 4, i.e., in the X-direction in the example shown. Accordingly, a running element 5 can be mounted in guides by rollers. A running element drive, not shown, is provided for the running element 5 in order to move the running element 5 in the X-direction. The running element drive can be provided on the running element 5 itself, but can also be arranged on the boom 4 or on the supporting structure 3. The running element 5 is usually connected by retaining elements 6 to a load receiving element 7 for receiving a load 8. In the case of a container crane 2, the load 8 is typically a container 9, in most cases an ISO container 20, 40 or 45 feet long, 8 feet wide and 8 feet, 6 inches high.
The retaining elements 6 are usually in the form of ropes, wherein in most cases four retaining elements 6 are arranged on the running element 5, but more or fewer retaining elements 6 can also be provided, in the simplest case also only a single retaining element 6. To accommodate a load 8, such as a container 9, the distance between the running element 5 and the load receiving element 7 can be adjusted by a lifting drive (not shown), thus in the Y-direction in
In the situation shown in
In order to move a load 8 along a trajectory T from the starting point A to the end point E, the lifting device 1 has a crane controller 16 with a computing unit 12, for example in the form of an electronic control unit in the form of suitable hardware and/or software, by which the movement of the load 8 along the respective movement direction X, Y, Z can be controlled. If a starting point A and/or end point E is defined manually by a user, for example, the starting point A and/or end point E can be transmitted to the computing unit 12, for example, via a suitable interface. A starting point A and/or an end point E can also be determined in the computing unit 12. Possible implementations of a computing unit 12 include microprocessor-based hardware, microcontrollers, and integrated circuits (ASIC, FPGA). The crane controller 16 communicates with the drives provided for this purpose, such as a running element drive or a winch, in order to control and/or regulate the movement of the load 8, and is typically designed to be able to detect the current position of the running element 5, the position of the load receiving element 7 and, if necessary, the position of the supporting structure 3.
In this sense, in the situation shown in
From a trajectory T along which the load 8 is moved, a rapid transfer of the load 8 from the starting point A to the end point E is usually required. As explained at the beginning, during such a transfer it may become necessary for several reasons to adapt a given trajectory T during the movement of the load 8 to changing circumstances in the area of the lifting device 1, i.e., to replan the trajectory T. This problem is solved by the method according to the invention. How the method according to the invention can be used to replan a trajectory T taking into account a changed or newly specified obstacle 11n is specifically shown in
Here, three situations are plotted along the time axis t, which can occur during the application of the method according to the invention during the movement of a load 8 by a lifting device 1 along a first and along a second movement direction, in this case along the movement directions X and Y. In the situation shown in
At time point tK, hereinafter also referred to as command time point tK, a new obstacle 11n is now specified in place of the previous obstacle 11a. The command time point tK can be specified by an operator, such as a crane operator, and can therefore be selected as freely as possible. However, the command time point tK can also be generated internally in the computing unit 12. Also, several successive command time points tK for a repeated change of an obstacle 11 are conceivable. At command time point tK, the load 8 is at position Pz(tK). According to the invention, the replanning of the movement of the load 8 is started immediately after the new obstacle 11n has been specified. This is done according to the previous explanations by projections of the position Pz(tK) taken on by the load 8 at the command time point tK and of the end point E onto the movement directions X, Y, as well as by projections of keypoints S1n, S2n, which are derived from the new obstacle 11n, onto the movement directions X, Y. Preferably, the keypoints S1n, S2n are corner points of the prohibition zone V1n as shown. For the prohibition zone V1a, keypoints S1a, S2A can of course also be specified. As will be shown below, keypoints such as S1n, S2n can be used as intermediate targets, particularly when planning movements along the Y-axis, to avoid collisions with a prohibition zone V1n and thus with the obstacle 11n.
The projections of the position Pz(tK) taken on by the load 8 at the command time point tK, of the end point E and of the keypoints S1n, S2n on the X-axis are referred to as PZ′, E′, S1n′ and S2n′, while the projections on the Y-axis are referred to as PZ″, E″, S1n″ and S2n″. According to the invention, the projections PZ′, E′, S1n′, S2n′ and the projections PZ″, E″, S1n″, S2n″ are considered separately. For each of these groups of projections, individual movements TE are planned to connect the respective projections PZ′, E′, S1n′, S2n′ and PZ″, E″, S1n″, S2n″ along the corresponding movement direction X, Y, respectively. An individual movement TE preferably defines a profile of position, velocity and acceleration along the respective movement direction. For example, by an individual movement TEX for the movement direction X, the projections PZ′, E′, S1n′, Sn2′ along the movement direction X are connected. For each of the movement directions X. Y there is at least one individual movement TE. Individual movements related to movement directions such as X, Y are referred to here as TEX, TEY.
In order to avoid obstacles 11a, 11n that are located between the position PZ(tK) taken on by the load 8 at the command time point tK and the end point E, several individual movements TEY can be planned, in particular for the second movement direction Y. Thus, individual movements TEY in the form of lifting movements (“Hoist Up”) can be provided for the load 8 along the movement direction Y on the one hand, and individual movements TEY in the form of lowering movements (“Hoist Down”) can be provided on the other hand. To avoid obstacles 11a, 11n, lifting movements (“Hoist Up”) are usually performed first, and lowering movements (“Hoist Down”) are only performed after a suitable horizontal movement along the X-axis (“Move Trolley”) to move the load 8 around an obstacle 11n or to lift the load 8 over an obstacle 11n. It should be noted that in order to move a load 8 around an obstacle 11n, individual movements TE along the movement directions X, Y must be combined.
In the situation shown in
In this context, it should be noted that along the X-axis, a single individual movement TEX from the projection PZ′(tk) to the projection E′ is usually sufficient to ensure a collision-free movement of the load 8 into the end point E. It should also be noted at this point that the present invention also covers the case where an obstacle 11a previously assumed for planning a trajectory T subsequently turns out not to exist, i.e. the obstacle 11a disappears. In such a case, a single individual movement TEY along the Y-axis is sufficient to transfer the load 8 from its position PZ(tK) to the end point E.
After planning the individual movements TE, the planned individual movements TE are executed according to a specified sequence, hereinafter referred to as “movement sequence”, to move the load 8 along the first movement direction X and along the second movement direction Y. Thus, a replanning trajectory TU results from the planned individual movements TE. The replanning trajectory TU is shown by the solid line in
Individual movements TE, as may occur in the method according to the invention, are shown in
By separating the new overall movement to be planned into a plurality of individual movements TE to be planned instead, the (re-)planning task to be solved is significantly simplified compared to methods known from the prior art. Thus, the complex problem of planning at least one two-dimensional movement is reduced to planning several only one-dimensional movements. Especially this circumstance proves to be advantageous in the practical implementation, since the planning of one-dimensional movements is a well-known problem in control and automation engineering, which has already been solved in various ways. In the concrete planning of the individual TE movements, a plurality of known approaches can thus be used. If, for example, a transfer of the load 8 to the end point E is to be ensured by time-optimal individual movements TE, the bang-bang controller approach, for example, which is sufficiently known from control engineering literature, can be used, using maximum accelerations along the movement directions X, Y that can be realized by the given drives of the lifting device 1.
In
In the example shown in
σ(x) stands herein for the sigma function well known from mathematics, which takes on the value one for arguments greater than or equal to zero (x≥0), and is equal to zero otherwise. vx0 stands herein for the initial velocity at the beginning of the individual movement. Accordingly, the planning of individual movements TE, on the basis of which a position profile sa, a velocity profile vx and an acceleration profile ax are determined, can be performed by defining functions as shown above, for example by specifying time points Ta1 and Ta2. In an advantageous manner, boundary conditions can also be specified for this purpose, such as sx(tE)=E′, vx(tE)=vtE. In particular, the velocity vtE can be chosen equal to zero, but also different from zero.
According to the above explanations, for planning a time-optimal individual movement TE, acceleration phases with maximum positive acceleration a+max, with maximum negative acceleration a−max, as well as with vanishing acceleration, i.e. an acceleration ax=0, must be specified. In the procedure shown in
Kinematic and/or geometric limit values can, for example, be stored in the computing unit 12 or specified to the computing unit 12. Further kinematic limit values are preferably a maximum speed vTmax and/or a maximum acceleration aTmax of the running element 5, or a maximum speed vHmax and a maximum acceleration aHmax of the lift drive in Y-direction. A geometric limitation can be given, for example, by a maximum deflection angle ΘXYmax of the load receiving element 7 in the plane EXY. In the case of a third movement direction of the lifting device in the Z direction, a maximum speed vSmax and a maximum acceleration aSmax of the supporting structure 3 can additionally be specified as kinematic limit values and a maximum deflection angle ΘZYmax of the load receiving element 7 in the plane EZY can be specified as a geometric limit value. However, in addition to planning time-optimal individual movements TE, other approaches to planning individual movements TE can also be used.
In many cases, it proves advantageous in this context to plan the individual movements TEx, TEy (and if necessary TEz) along the axes X, Y (and if necessary Z) independently of one another. This means that there is no relationship between the planned individual movements TEx, TEy (or TEz) on the respective axes, which could be expressed mathematically, for example, and which, in the case of a change in the position Pz of the load 8 along a first movement direction, for example the X-axis, would cause a change in the position Pz of the load 8 also along a second movement direction, for example the Y-axis.
The planned individual movements TE are subsequently executed either simultaneously or offset in time. Staggering planned individual movements TE may be necessary, particularly in the case of prohibited areas Vi or obstacles 11 described above. For example, it may be necessary to delay the start of a lowering movement in the Y-direction in order to drive around an obstacle 11. An individual movement TEY planned in the Y-direction would therefore only be started with a time delay after the individual movement planned in the X-direction has already been started. The specification of which individual movement TE is to be executed first and which individual movement TE is to be executed later, if necessary, is expressed in the context of the present invention by a so-called movement sequence, which can be specified, for example, to the crane controller 16 or to the computing unit 12 provided in the crane controller 16 via a suitable interface. A movement sequence can thus be stored in the computing unit 12. A movement sequence thus defines the order in which the planned individual TE movements are to be executed. For example, a movement sequence might be: “First horizontal movement in X-direction, only then vertical movement in Y-direction.” A movement sequence can also be in the form of a table in which TE start times are assigned to the planned individual movements at which they are started. A movement sequence can also comprise time intervals that define waiting times that must at least be provided between the start times of successive individual movements TE.
A central problem in the manipulation of loads 8 by lifting devices 1, which in many cases can either be solved, but usually at least greatly reduced, by suitable trajectory planning, is the problem of load swing. In particular, rapid movement operations of lifting devices can often excite undesirable swinging and/or swaying movements of transported loads 8 or of load receiving elements. Swinging of loads 8 can in turn delay handling and manipulation procedures, since loads such as containers often cannot be placed at all in such cases, or at least not with sufficient precision, and it is necessary to wait until a given swinging has subsided again. In
In the present context, it is shown that so-called flat trajectories, i.e., trajectories which can be continuously differentiated a sufficient number of times depending on the concrete design of the lifting device, are effective against (load) swinging and/or pendulum movements. This approach is based on the fact that a lifting device 1 as shown in
Flat outputs and their derivatives allow, in the case of a flat system such as the lifting device 1 under consideration, the representation (also “parameterization”) of the internal state or system variables of the flat system. The internal state or system variables of a flat system can thus be represented as functions of the flat outputs and their derivatives, which, of course, the person skilled in the art of control engineering knows about. For the suppression of swinging of a load 8 transported by a lifting device 1, the fact that (in an ideal consideration) a control system constructed with the aid of a flat output cannot excite any swinging or pendulum movements of the load is of importance at this point.
In the case of the lifting device 1 considered here, the position profiles of the load 8 along the movement directions X, Y represent flat outputs of the system “lifting device” 1. Here it is shown that for the suppression of swinging and/or pendulum movements it is sufficient to specify the position profiles sx(t), sy(t) determined by the planned individual movements TE as four times continuously time-differentiable position profiles sx(t), sy(t). In this manner, swinging and/or pendulum movements can be suppressed without having to rely on measurements of the load position PZ. It should be mentioned at this point that the use of trajectories that can be continuously differentiated even more often can ensure even smoother movement profiles that are even gentler on the drives.
To ensure that the position profiles sx(t), sy(t) defined by the specified individual movements TE are actually fourfold continuously differentiable, the position profiles sx(t), sy(t) can be filtered before they are executed. However, the described fourfold differentiability can also be achieved in a manner other than filtering, for example by polynomial approaches for the specified position, velocity and/or acceleration profiles. The filtering of individual movements TE, which are available as scalar progressions, is much easier than the filtering of multidimensional profiles. This is another reason why the use of individual movements TE represents a significant advantage in the control of lifting devices 1.
The described approach for generating flat trajectories by a suitable filtering or polynomial approach can also be used in a particularly advantageous manner to comply with the kinematic constraints for the lifting device drives mentioned further above as well as the geometric limitation of the deflection angle of the load. For example, the partial movements of the load 8 are related to the movement of the running element 5 via the known pendulum equations:
The variables xT, vT, aT here stand for the movements of the running element 5, xL and yL on the other hand for the coordinates of the load 8, whose corresponding derivatives are noted as {dot over (x)}L, ,
, xL(4) and so on. According to the above, the given equations represent a flat parameterization of the state variables xT, vT, aT as a function of the flat outputs xL, yL. The given relationship makes it clear that specifications regarding the profiles of the coordination xL and yL have a direct effect on the movement of the running element 5.
To implement the described filtering to generate flat trajectories, for example, a suitable filter F can be used, such as a moving-average filter (MA filter). A schematic representation of such filtering is shown in
The filter time τfilt of a filter F implemented as an MA filter can be selected here in a particularly advantageous manner as a function of the geometry of the lifting device 1 and the pendulum equations. By selecting larger time constants, the previously planned individual movements can be filtered even more strongly and, for example, even more conservative progressions can be specified for the running element. By multiple filtering, individual movements TE can be generated in an extremely advantageous manner, by which kinematic constraints specified for the drives of the lifting device 1 are taken into account, such as speed, acceleration, jerk limitations.
How prohibition zones Vi can be defined in concrete terms is further shown in more detail in
In a situation such as that shown in
A first possibility in this respect is to execute the planned individual movements TE at intervals from one another, i.e., to wait after executing an individual movement TE and only continue with the next individual movement TE corresponding to the movement sequence after a specified waiting time.
Specifically, a movement sequence can specify, for example, that in the situation shown in
In the course of determining the movement sequence in which the planned individual movements are to be executed, it is often particularly advantageous to additionally check whether a collision of the load 8 with a prohibition zone Vi may occur in the profile of the resulting replanning trajectory TU.
How to proceed in this respect is shown in
In the X-direction, however, it can be seen that the individual movement TExa would lead to a collision with the obstacle 11. The individual movement TExa planned at the beginning is shown in
In a particularly elegant manner, a check as to whether a collision with an obstacle is to be expected can be performed in the following manner: Before the execution of at least one planned individual movement TE, so-called collision times tiX, tiY are determined and compared with one another. The collision times tiX, tiY are to be understood as the respective time points at which the load 8 would reach a projection Si′ or Si″ of a keypoint Si derived from a prohibition zone Vi placed around an obstacle 11 on the respective movement direction X, Y by the respective planned individual movements TEX, TEY.
As mentioned, keypoints Si are preferably corner points of prohibition zones Vi. In order to check whether an expected replanning trajectory TU would lead to a collision of the load 8 with an obstacle 11 specified in the working area 15, it can be checked in the situation shown in
In the procedure described, particular attention must be paid to the Y-coordinate of the load position PZ, i.e., the projection PZ″ of the load position PZ on the Y-axis. It must be taken into account whether the Y-coordinate PZ″ of the load position PZ at the beginning of an individual movement TEY in the Y-direction is larger or smaller than the projection Si″ of the next keypoint Si in the Y-direction onto the Y-axis. In many practically relevant cases, the load position PZ at the beginning of a newly planned individual movement TE is given by the load position PZ at command time point tk, i.e. PZ(tk). If the projection Si″ of the keypoint Si closest in Y-direction to the Y-axis is smaller than the projection PZ″ of the load position PZ to the Y-axis, the condition for a collision is fulfilled exactly if vice versa tiY<tiX applies.
Individual movements TEX, TEY are often not available in analytical form. In an advantageous manner, numerical root finding methods can also be used in such cases to determine the collision times tiX, tiY, such as the well-known bisection method or Newton's method.
In practical implementation, it is often advantageous to combine both analytical and numerical methods in order to determine the collision times tiX, tiY. Such a case is described below. The acceleration profiles ax and ay are initially specified as piecewise constant as described. For these piecewise constant acceleration profiles ax and ay, the corresponding position profiles sx and sy can easily be calculated analytically. For the resulting analytical descriptions of the position profiles sx(t) and sy(t), the collision times tiX, tiY can be calculated with little effort, e.g. by determining the zeros tx0, ty0 of the functions fx(t)=sx(t)−Si′, fy(t)=sy(t)−Si′. For these zeros tx0, ty0, accordingly fx(tx0)=fy(ty0)=0 is valid by definition. Of course, the formulas fx(t) and fy(t) can also take into account the height and width of the load 8, which are usually given by the height H7 of the load receiving element 7 plus the height H9 of a container 9 in the Y-direction, as well as half the width B7 of the load receiving element 7 (typically corresponding to half the width B9 of the container 9). Depending on whether the load 8 approaches the next keypoint Si from the left, right, top or bottom, the function fx(t), for example, can be increased or decreased by half the width B9.
If the individual movements TEX, TEY are filtered, for example by an MA filter described at the beginning, the position profiles sx(t) and sy(t) in particular are filtered to form filtered position profiles {tilde over (s)}x(t) and {tilde over (s)}y(t). For such filtered position profiles {tilde over (s)}x(t) and {tilde over (s)}y(t), the property of an MA filter can now be utilized, according to which the zeros tx0, ty0 determined from the function fx(t) shown above can be shifted by filtering by a maximum of the filter time τfilt of the MA filter used for filtering. Accordingly, for the case of twofold filtering, a time interval of length 2·τfilt can be placed around the previously determined zeros tx0, ty0, with the zeros tx0, ty0 at the beginning of the interval at the left interval boundary. In such a time interval, a bisection procedure can now be used to search for zeros of the functions {tilde over (f)}x(t)={tilde over (s)}x(t)−Si′ and {tilde over (f)}y(t)={tilde over (s)}y(t)−Si″. These zeros then correspond to the sought collision times tiX, tiY for the filtered position profiles {tilde over (s)}x(t) and {tilde over (s)}y(t).
Besides the described comparison of two collision times tiX, tiY, another approach for collision checking can be mentioned. Thus, initially only the collision time tiy can be calculated, which indicates when a projection of a given keypoint Si″ onto the Y-axis is reached along the Y-axis by a planned individual movement TEY in Y-direction. This collision time tiy can also be used to determine the position sx(tiy), which describes the position on the X-axis specified by the planned individual movement TEx in the X-direction at the collision time tiy.
The check whether a collision with an obstacle 11 is to be expected due to planned individual movements TE can then be performed by comparing the position sx(tiy) with the projection Si′ of the keypoint Si on the X-axis. In the specific case shown in
On the other hand, to check for a collision, it is also possible to first determine the collision time tix, to use this collision time tix to determine the corresponding position on the Y-axis sy(tix), and to compare the corresponding position on the Y-axis sy(tix) with a projection Si″ of a keypoint Si on the Y-axis. In this sense, in the case shown in
If one of the above-mentioned advance calculations before the execution of at least one individual movement TE shows that a collision is to be expected, the individual movement TE causing the collision is not executed and instead a specified braking operation is performed along the movement direction X, Y for which the at least one individual movement not executed is planned, in
After the minimum braking time TSt, at least one individual movement TE that has not been executed is replanned to a new individual movement TE in a replanning step in an advantageous manner. In a further checking step with the at least one new individual movement TE, it can be checked again whether the replanning trajectory TU to be expected on the basis of the new individual movement TE leads to a collision. If no collision is detected, the newly planned individual movement TE can be executed.
If, however, a collision of the load 8 with an obstacle 11 is predicted again in the preceding verification step, a specified braking operation can advantageously be performed again for at least one further minimum braking time TSt. After that, a replanning step and a review step following this replanning step can be performed again.
The load 8 can also be brought to a complete standstill by repeatedly performing the braking operation described above. Further planning of individual movements TE then occurs starting from the standstill of the load 8. The possibility of using the method according to the invention to plan corresponding individual movements TE starting from the standstill of the load 8 and thus a trajectory T resulting from these individual movements TE directly opens up the option of also planning fundamentally new trajectories T for the movement of a load 8 using the method according to the invention. If the load 8 is at standstill at the start of a movement operation and no trajectory T is yet known for its movement from the starting point A to the end point E, the standstill at the starting point A can be assumed/specified as the end of a repeatedly executed braking operation and, by applying the method according to the invention, a first trajectory T for transferring the load 8 from standstill at the starting point A to the end point E can also be planned.
Similar to the previously described determination of the collision times tix, tiy, a time interval can also be specified for a braking operation in which the load 8 can be brought to a complete standstill. Specifically, the limits tk (command time point) and tk+TTE can be specified for this time interval, where TTE describes the duration of a planned individual movement TE. Within this time interval, the load 8 comes to a standstill when a braking operation is repeatedly executed, wherein the specific time point can be determined as before, for example, by a bisection method.
Although the present invention has so far been described in terms of a trajectory T in the X-Y plane EXY, it is also possible to extend it to a three-dimensional trajectory T in space. The basic method for replanning even three-dimensional trajectories T on the basis of preferably independent individual movements remains unchanged. If the load 8 is additionally moved along a third movement direction Z by the lifting device 1, at least one individual movement TEZ is also planned for the third movement direction Z in order to replan the trajectory T in accordance with the method according to the invention, which is executed in accordance with a specified movement sequence in addition to the individual movements TEX, TEY planned for the first movement direction X and for the second movement direction Y.
For the application of the present invention in the planning of three-dimensional trajectories, it is in many cases advantageous to specify a 3D working area for constraining the movement of the load 8 instead of a two-dimensional working area 15. In contrast to a purely two-dimensional consideration, however, it may also be necessary in the 3D case to take into account the extent of an obstacle 11 as well as the extent of a load 8 in the Z direction. This may be necessary, for example, in assembly halls where obstacles 11 arranged on the walls project into the space of the assembly hall. In such cases, situations may arise in which an obstacle 11 can also be avoided by suitable lowering movements, but this can also be implemented by the method according to the invention.
Such a case is shown in
In many practical applications, it can be advantageous in this context to use optical measuring systems to detect the load position PZ or also obstacles 11, as described for example in EP 3 653 562 A1. Optical measuring systems include suitable camera systems in particular, but also laser scanners. For the purpose of monitoring obstacles 11 that change with respect to their position in the working area 15, a camera can be mounted/installed directly on the trolley of a lifting device 1. Specifically, an optical detection of obstacles can in many cases provide information about how high and at which spatial positions obstacles 11 to be avoided are located. If a change in the position of an obstacle 11 is detected, a new prohibition zone V can be derived from it and the replanning of a trajectory for the movement of a load can be performed as described. In addition to camera and/or laser scanning systems, other approaches to detecting obstacles are also conceivable. In many cases, automation or logistics systems implemented in software also provide information about obstacles 11 to be considered.
It is noted that the foregoing examples have been provided merely for the purpose of explanation and are in no way to be construed as limiting of the present invention. While the present invention has been described with reference to an exemplary embodiment, it is understood that the words which have been used herein are words of description and illustration, rather than words of limitation. Changes may be made, within the purview of the appended claims, as presently stated and as amended, without departing from the scope and spirit of the present invention in its aspects. Although the present invention has been described herein with reference to particular means, materials and embodiments, the present invention is not intended to be limited to the particulars disclosed herein; rather, the present invention extends to all functionally equivalent structures, methods and uses, such as are within the scope of the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| A50961/2021 | Nov 2021 | AT | national |
