Method for adjusting an outer dimension of an autonomous mobile robot, control device, and autonomous mobile robot
The reconfigurable chassis mechanism for AMRs, controlled by a control device, addresses the challenge of adapting external dimensions for efficient navigation by optimizing stability and efficiency, enhancing agility and maneuverability.
Patent Information
- Application Number
- PCT/EP2025/050417
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2024-01-12
- Filing Date
- 2025-01-09
- Publication Date
- 2025-07-17
AI Technical Summary
Existing autonomous mobile robots (AMRs) with rigid chassis face limitations in adapting their external dimensions to navigate dynamically and efficiently through diverse environments, lacking methods to adjust their length and width dimensions effectively.
A reconfigurable chassis mechanism for AMRs, controlled by a control device, allows adaptive adjustment of external dimensions based on multi-objective optimization, balancing tipping safety, directional stability, and dimensional change efficiency, using actuators and sensors to adapt to environmental conditions.
Enhances agility, maneuverability, and stability of AMRs by optimizing external dimensions for safe and efficient navigation through complex environments, reducing the need for human intervention and minimizing wear on moving parts.
Smart Images

Figure EP2025050417_17072025_PF_FP_ABST
Abstract
Description
[0001] Description Method for adjusting an external dimension of an autonomous mobile robot, control device, and autonomous mobile robot. The invention relates to a method for adjusting an external dimension, in particular a longitudinal and / or transverse dimension, of an autonomous mobile robot (AMR), a control device configured and programmed to carry out the method, and an AMR having a reconfigurable chassis and the control device. An autonomous mobile robot (AMR) is a vehicle that can move autonomously and perform tasks in a self-controlled manner without external support. Such robots use sensors, cameras, and other devices to navigate and interact with their environment. This enables them to perform tasks such as transporting materials or equipment in a warehouse or production facility.provide support in search and rescue missions or assist in medical or rehabilitation treatments. Currently, chassis with four or eight wheels are primarily used in the industry for heavy, omnidirectional autonomous mobile robots, depending on the application. GB 2594598 A discloses an automatic track width adjustment system for an agricultural mobile robot. The robot comprises a frame defining a plane and a plurality of wheels, the system having an adjustment mechanism for changing the position of at least one of the wheels relative to the frame in a direction perpendicular to a direction of movement of the robot within the plane. However, the dimensions of the frame cannot be changed. Other mobile robots are also known from CN 110194216 A, CN 217345507 U, US 9346502 B2, US 9616948 B2, and US 11235821 B2. From Rayguru, MM, et al.,An output feedback-based robust saturated controller design for pavement sweeping self-reconfigurable robot. IEEE / ASME Transactions on Mechatronics, 2021. 26(3): pp. 1236-1247. For example, a sweeper with a reconfigurable chassis is known. Accordingly, some mobile robots include a reconfigurable or adaptive chassis so that an external dimension of the robot can be adapted, for example, according to the environment in which it moves. However, methods for adjusting the external dimension of the reconfigurable chassis, i.e., the robot, have so far been lacking. It is therefore the object of the invention to provide a technique for adjusting an external dimension, in particular a longitudinal and / or transverse dimension, of an autonomous mobile robot. This object is achieved by the method,The control device and the autonomous mobile robot according to the independent claims are solved. Advantageous embodiments are the subject of the dependent claims and / or are described below. A method according to the invention is used for adjusting an external dimension, in particular a longitudinal and / or a transverse dimension, of an autonomous mobile robot. Autonomous mobile robots (AMRs) were developed to reduce the limitations associated with rigid path following and minimize dependence on human intervention. AMRs have perception systems to assist navigation and independent control systems. Consequently, AMRs are capable of navigating to targets, perceiving and interpreting their environment, and proactively avoiding obstacles or hazards.that they encounter along their path. In addition, AMRs can perform actions to fulfill their assigned tasks. According to the invention, the AMR has a reconfigurable chassis that allows the external dimensions of the AMR to be changed. The installation of a reconfigurable chassis enables the development of a flexible AMR platform, characterized by its ability to adapt its dimensions to navigate through dynamic and complex environments. The AMR with the reconfigurable chassis can proactively adjust its external dimensions based on actuators and external constraints. For example, by using a smaller frame, the AMR can achieve greater agility and maneuverability, making it well-suited for transporting smaller objects. Conversely, a larger frame offers improved stability and the ability toto transport larger objects. This adaptability gives the AMR a higher degree of flexibility compared to other AMRs with static dimensions, enabling the AMR to move effectively through various scenarios and meet a wide range of mission objectives. Furthermore, the outer dimensions of the reconfigurable chassis, i.e., the AMR, are adjusted using a multi-objective optimization that includes tipping stability, directional stability, and dimensional change efficiency as objective functions. In multi-objective optimization, a problem solution is subject to several requirements that must be met in the best possible way. These requirements are referred to as objective functions and are typically conflicting.so that an optimum with respect to all objective functions cannot be achieved with a single solution. Nevertheless, when solving the multi-objective optimization using a known method, a good compromise between the individual objective functions can be found. Accordingly, the invention provides a technique for setting the optimal external dimension of the AMR. The objective functions can each be weighted. In particular, the tipping stability and the directional stability can be weighted equally or more heavily than the dimensional change efficiency. Consequently, the stability of the AMR is given priority over the dimensional change efficiency. Furthermore, as a constraint for setting the external dimension, a minimum adjustable dimension of the reconfigurable chassis and a minimum of a maximum adjustable dimension of the reconfigurable chassis and a maximum permissible dimension that depends on an environment can be specified.in which the autonomous mobile robot moves. Accordingly, the setting of the external dimension is limited to permissible values. In particular, the environment in which the AMR moves is taken into account. A safety distance can be taken into account when determining the maximum permissible dimension. The minimum adjustable dimension can, in particular, be a minimum adjustable longitudinal and / or a minimum adjustable transverse dimension. The maximum adjustable dimension can, in particular, be a maximum adjustable longitudinal and / or a maximum adjustable transverse dimension. The maximum permissible dimension can, in particular, be a maximum permissible longitudinal and / or a maximum permissible transverse dimension, which depend on the environment in which the autonomous mobile robot moves.Again, a safety margin can be taken into account when determining the maximum permissible longitudinal and / or the maximum permissible transverse dimension. The objective function for tipping safety can aim to minimize a negative value of a transverse dimension of the reconfigurable chassis. This also corresponds to minimizing a reciprocal value of the transverse dimension or maximizing the transverse dimension. Consequently, tipping safety for the AMR can be increased and tipping can be prevented. The objective function for directional stability can aim to minimize a longitudinal dimension of the reconfigurable chassis. Since the longitudinal dimension also corresponds to a wheelbase, maneuverability and thus directional stability can be improved. The objective function for dimensional change efficiency can aim toto minimize the amount of dimensional change. The amount of dimensional change can be the amount of longitudinal dimensional change and / or the amount of transverse dimensional change. Consequently, AMR efficiency can be achieved by reducing the time required for the dimensional change. Furthermore, wear on moving parts for reconfiguring the chassis can be reduced. Preferably, the objective function for the dimensional change efficiency can aim to minimize the amount of longitudinal dimensional change and the amount of transverse dimensional change. The amount of longitudinal dimensional change and the amount of transverse dimensional change can each be weighted. In this case, the amount of longitudinal dimensional change can be weighted equally or more heavily than the amount of transverse dimensional change. Consequently, a higher dimensional change efficiency can be achieved.since a longitudinal dimension change due to a change in the chassis in the direction of movement of the AMR is less complex. Furthermore, a control signal can be output to at least one actuator of the reconfigurable chassis for adjusting the outer dimension. Consequently, the reconfigurable chassis can be adjusted according to a result of the multi-objective optimization. A control device according to the invention is designed and programmed to carry out the method according to the invention. Consequently, the advantages of the method according to the invention can be achieved. An autonomous mobile robot according to the invention has a reconfigurable chassis that enables a change in the outer dimension of the autonomous mobile robot.the control device according to the invention. Consequently, the advantages of the method according to the invention can be achieved. Further measures improving the invention are described in more detail below, together with the description of a preferred embodiment of the invention, with reference to the figures, wherein identical or similar components are provided with the same reference numerals. Figure 1 shows a first schematic perspective view of an autonomous mobile robot in a state with minimal longitudinal and transverse dimensions of a chassis. Figure 2 shows a second schematic perspective view of the robot to illustrate a process for changing a longitudinal dimension of the chassis. Figure 3 shows a third schematic perspective view of the robot to illustrate a process for changing a transverse dimension of the chassis.Figure 4 shows a fourth schematic perspective view of the robot to illustrate a process for simultaneously changing the longitudinal and transverse dimensions of the chassis. Figure 5 shows a schematic perspective view of a corner module of the robot. Figure 6 shows a schematic exploded view of a chassis of the robot. Figure 7 shows a schematic perspective view of a guidance mechanism of the robot. Figure 8 shows a schematic representation of a model for explaining lateral load transfer. Figure 9 shows a schematic representation of a model for explaining forces acting on a wheel during cornering, as well as a Kamm circle. Figure 10 shows a schematic representation of a track following model. Figure 11 shows a schematic plan view of an environment in which the robot moves. Figure 12 shows a schematic perspective view of the environment in which the robot moves.Figure 13 shows a time diagram of the longitudinal dimension of the robot during movement in the environment. Figure 14 shows a time diagram of the transverse dimension of the robot during movement in the environment. Figure 15 shows schematic plan views of the environment and the robot at different times during a movement of the robot in the environment. Figures 1 to 4 show an exemplary autonomous mobile robot 1 (hereinafter referred to as AMR) comprising a chassis 22 on which four corner modules 10 are arranged and supported. A corner module 10 is arranged at each corner of the chassis 22, which is essentially rectangular in plan view. The chassis 22 of the AMR 1 is shown in more detail in Figure 6. The chassis 22 comprises four sliding modules, namely in the form of a front first sliding module 2, a rear second sliding module 3,an outer third sliding module 4 and an inner fourth sliding module 5. The outer sliding module 4 has a first leg 4a and a second leg 4b arranged parallel thereto, wherein the two legs 4a, 4b of the outer sliding module 4 are cross-connected to one another via a first support leg 4c at the end faces of the legs 4a, 4b. The support leg 4c is arranged perpendicular to the two legs 4a, 4b. In the two legs 4a, 4b of the outer sliding module 4, a first and second leg 5a, 5b of the inner sliding module 5 are mounted and axially guided via an associated linear guide mechanism 6, 9. The two legs 5a, 5b of the inner sliding module 5 are cross-connected to one another via a first support leg 5c and a second support leg 5d. The two support legs 5c, 5d are perpendicular to the two legs 5a,5b. The first linear guide mechanism 6 is operatively arranged between the first leg 4a of the first sliding module 4 and the first leg 5a of the second sliding module 5. The fourth linear guide mechanism 9 is operatively arranged between the second leg 4b of the first sliding module 4 and the second leg 5b of the second sliding module 5. A second linear guide mechanism 7 is arranged between the rear sliding module 3 and the support leg 4c of the outer sliding module 4, which supports the outer sliding module 4 relative to the rear sliding module 3, and vice versa, as well as axially guides the rear sliding module 3 relative to the outer sliding module 4. A third linear guide mechanism 8 is arranged between the front sliding module 2 and the first support leg 5c of the inner sliding module 5, which supports the inner sliding module 5 relative to the front sliding module 2, and vice versa.and guides the front sliding module 2 axially relative to the inner sliding module 5. All sliding modules 2-5 or their legs of the chassis 22 are arranged in a common plane parallel to a traveling surface of the AMR 1. The chassis 22 can also have means—not shown here—for receiving structural elements of the AMR 1. Figure 7 shows the first linear guide mechanism 6 as an example, with the other guide mechanisms 7, 8, 9 being essentially identical. The linear guide mechanism 6 shown here comprises a rack 18 and a pinion 19, which mesh with each other and implement a locking function. Furthermore, a guide block 21 is provided.on which a sliding block 20 is axially guided. Using the example of the first linear guide mechanism 6, the guide block 21 is arranged on the first leg 4a of the outer sliding module 4, and the sliding block 20 is arranged on the first leg 5a of the inner sliding module 5. When the sliding modules 4, 5 are displaced relative to one another by actuating the corner modules 10 and a desired axial relative position of the two first legs 4a, 5a is reached, this position can be fixed by locking the pinion 19 relative to the rack 18. This ensures the stability and strength of the chassis 22 and any structure arranged thereon (not shown here) required for the operation of the AMR 1. One of the corner modules 10 is arranged on each sliding module 2-5, in such a way thatthat in the top view of the chassis 22, a corner module 10 is arranged in each corner of the chassis 22. Figure 5 shows one of these corner modules 10 as an example, with all other corner modules 10 being essentially identical. The load of the AMR 1 is supported on the surface it is traveling on via the corner modules 10. Each corner module 10 comprises a support 11 on which the components of the corner module 10 are arranged and via which the corner module 10 is connected to the respective sliding module 2 - 5. Each corner module 10 comprises a rotatably driven and steerable wheel 12.a steering mechanism 13 with a steering motor 15 for adjusting a steering angle of the wheel 12 and a hub motor 17 (not shown here) within the wheel 12 for generating drive power to rotate the wheel 12. The steering motor 15 is operatively connected to the wheel 12 via a gear stage 16 designed as a spur gear stage. A smaller first gear 16a of the gear stage 16 is directly connected to a drive shaft (not shown here) of the steering motor 15 and transmits drive power to a larger second gear 16b meshing therewith, which in turn is connected in a rotationally fixed manner to a fork 23 on which the hub motor 17 is arranged and on which the wheel 12 is rotatably mounted. This allows the corner module 10 to rotate 360° around its vertical axis, allowing any desired steering angle of the wheel 12 to be set. In Figure 1, a control device 14 is indicated, which is designed toto control the corner modules 10 and to actuate the aforementioned locking mechanism on the guide mechanisms 6 - 9. The corner modules 10 and the respective guide mechanisms 6 - 9 thus correspond to actuators for adjusting an outer dimension of the reconfigurable chassis 22. Furthermore, the control device 14 is communicatively connected, for example, to sensors (not shown here) in order to receive information about the environment of the AMR 1. Furthermore, the control device 14 receives information about the operating states of the AMR 1. Thus, the control device 14 controls and regulates a drive of the AMR 1 as well as a reconfiguration of the chassis 22, which will be described below, depending on an environment or a spatial condition around the AMR 1. The control device 14 is designed to, on the basis of information about an environment lying in front of the AMR 1 in the direction of travel, control at least one of the corner modules 10,Preferably, at least two of the corner modules 10, preferably all corner modules 10, are controlled or driven in such a way that the outer dimensions of the AMR 1 are adapted to the environment in front of the AMR 1 in the direction of travel. Thus, the longitudinal and transverse dimensions of the AMR 1 are adapted, depending on the situation, to the available space in front of the AMR 1, so that the AMR 1 can move stably, safely, and efficiently. One or more corner modules 10 can be controlled in such a way that the AMR 1 either only lengthens or only widens, or both lengthens and widens. The external adjustment is achieved by releasing or releasing the locking mechanisms on the guide mechanisms 6-9 and adjusting the steering angles of the wheels 12 of the corner module 10 according to the desired adjustment of the longitudinal and / or transverse dimensions of the chassis 22. After adjusting the steering angle of the respective corner module 10, the wheels 12 are driven in such a way thatthat at least two operatively connected sliding modules 2 - 5 are displaced axially relative to one another, whereby the chassis 22 becomes either longer or shorter and / or either wider or narrower. Figure 1 shows the AMR 1 or the chassis 22 of the AMR 1 with minimal external dimensions, i.e., minimal longitudinal dimensions and minimal transverse dimensions. In this case, the AMR 1 is maximally compact and can move in the tightest of spaces. Figure 2 shows a state during or after a change in the longitudinal dimension l of the chassis 22. The corner modules 10 are driven directly by the associated hub motors 17 in order to change the longitudinal dimension l of the AMR 1. The corner modules 10 of the front and inner sliding modules 2, 5 and the corner modules 10 of the rear and outer sliding modules 3, 4 move in the opposite direction. When the longitudinal dimension l of the chassis 22 is changed, a wheelbase is changed.i.e., enlarged or reduced. It is conceivable that only the corner modules 10 of the front and inner sliding module 2, 5 or only the corner modules 10 of the rear and outer sliding module 3, 4 move relative to the respectively stationary corner modules 10. This can save energy because only two corner modules 10 are actuated at a time. Figure 3 shows a state during or after a change in the transverse dimension w of the chassis 22. The corner modules 10 are actuated such that the wheels 12 are rotated by 90 degrees compared to the illustration in Figure 2 before the hub motors 17 of the corner modules 10 increase or decrease the lateral distance between the corner modules 10 arranged next to one another in the longitudinal direction of the chassis 22. The corner modules 10 of the outer and inner sliding modules 4, 5 and the corner modules 10 of the front and rear sliding modules 2,3 in the opposite direction. When changing the transverse dimension w of the chassis 22, a track width is changed, i.e., the chassis 22 is adjusted wider or narrower. It is conceivable that only the corner modules 10 of the outer and inner sliding modules 4, 5 or only the corner modules 10 of the rear and front sliding modules 2, 3 move relative to the stationary corner modules 10. This can save energy, since only two corner modules 10 are actuated at a time. Figure 4 shows a state during or after a simultaneous change in the longitudinal dimension l and the transverse dimension w of the chassis 22. The corner modules 10 are actuated such that the wheels 12 are rotated by approximately 45 degrees compared to the illustration in Figure 2.Before the distances between the corner modules 10 are increased or decreased in the longitudinal and transverse directions by means of the corner modules 10 through the corresponding drive of the hub drives 17. Consequently, the corner modules 10 move toward or away from each other in a star-shaped pattern. When the transverse dimension w of the chassis 22 is changed, a track width is changed, i.e., the chassis 22 is made wider or narrower. At the same time, when the longitudinal dimension l of the chassis 22 is changed, a wheelbase is changed, i.e., the chassis 22 is lengthened or shortened. In all cases, a sensor system (not shown here) can be provided to determine and track the current distance between the corner modules 10.so that the outer dimensions of the robot can be adjusted to the desired dimensions as precisely as possible. The steps of detecting the environment or the spatial conditions and adjusting the longitudinal dimension l and / or the transverse dimension w of the chassis 22, i.e., the AMR 1, to the detected environment can, in principle, be repeated as often as desired. A method for adjusting the longitudinal dimension l and the transverse dimension w of the AMR 1 using the reconfigurable chassis 22 is described below. It should be noted that the control device 14 is configured and programmed to execute the method and to output corresponding control commands to the corner modules 10 and the guide mechanisms 6-9 as the actuators of the chassis 22 for adjusting the longitudinal dimension l and / or the transverse dimension w. In order for the AMR 1 to perform its task, which essentially consists of transporting materials or goods in an environment, such as a warehouse,To ensure that the AMR 1's requirements for transporting goods efficiently and reliably are met, certain framework conditions must be observed. These framework conditions concern three aspects: tipping stability, directional stability, and dimensional change efficiency. In particular, tipping of the AMR 1 when carrying a certain load should be prevented. This risk can be minimized by setting an appropriate transverse dimension w of the AMR 1. The appropriate transverse dimension w corresponds to an appropriate wheelbase of the robot 1. Furthermore, a well-designed AMR 1 should have good directional stability, enabling it to navigate smoothly, make precise turns, and avoid various obstacles in its path. Furthermore, an unnecessary change in the longitudinal dimension l and / or the transverse dimension w not only reduces the efficiency of the autonomous mobile robot 1 due to the time required,but also causes increased wear on moving components, such as the corner modules 10 or the guide mechanisms 6-9. Consequently, operations for extending or retracting the chassis 22 should be minimized as much as possible. Accordingly, the following equation (1) can be formulated for multi-objective optimization or multi-criteria optimization: ^ = min(^^^^ + ^^^^ + ^^^^) (1) where J1 is the objective function for tipping safety, J2 is the objective function for directional stability, J3 is the objective function for dimensional change efficiency, and W1 to W3 are weights for the respective objective functions. The objective function J1 for tipping safety is based on a model for lateral load transfer, which is shown in Figure 8. From the model, equations (2) for the moment equilibrium at the left and right wheels can be derived: ^ ^ − ^^ ℎ^ ^^ ^ ^ = ^^^ ^ ^, ^^ + ^ ^^^^ℎ^ = ^^^where hg is the height from the ground to the center of gravity (CG for center of gravity in Fig.8), ay is the lateral acceleration, m is the mass, w is the transverse dimension corresponding to the track width, and Fzl and Fzr are the vertical loads on the left and right wheels, respectively. Furthermore, LTR denotes the lateral load transfer ratio and can be expressed by the following equation (3): ^ ^^ = ^ ^ ^ ^ ^ (3) Equation (3) shows that LTR is indirectly proportional to the transverse dimension w of the AMR 1. Consequently, the objective function J1 for the stability against overturning can be defined by the following equation (4) to minimize the reciprocal of the transverse dimension w. This can also be achieved by minimizing the negative value of the transverse dimension w. As shown in Figure 9, the wheel forces are induced by the steering angle ^ and are responsible for directional stability. Furthermore, the wheel force Fx in the longitudinal direction and the wheel force Fy in the transverse direction are coupled and bounded by the Kamm circle. Consequently, the following equations (5) can be derived from the model and the Kamm circle: From the first equation, it can be seen that the tire force Fy in the lateral direction is proportional to the steering angle ^. The steering angle ^ is specified by the path tracking algorithm. For simplicity, the pure pursuit method is used here. In the kinematic model of the pure pursuit method shown in Figure 10, O is the current center of rotation, R is the current radius of rotation, P is the target point to be reached by the AMR, d is the forward distance between the target point P and the center of the rear wheel, l is the longitudinal dimension, and e is the lateral deviation, which indicates the smallest distance between the current vehicle direction and point P. Consequently, ^ is used to represent the angle between the target and the current direction to later find the steering angle ^. The following equation (6) can therefore be derived from the model shown in Figure 10: ^ ^^^ = ^ ^ = 2 ^ ^ sin ^ (6)Assuming that tan^ is approximately equal to ^, equation (6) results in a direct proportionality between the steering angle ^ and the longitudinal dimension l, i.e., the wheelbase. Minimizing the tire force Fy in the transverse direction can not only reduce the transverse friction force but also create more room for the longitudinal friction force, which has a positive effect on directional stability. Therefore, the objective function J2 for directional stability is expressed by minimizing Fy using the following equation (7): ^^ = min^^^^ (7) Since Fy, as stated in equation (5), is proportional to the steering angle ^, which in turn is proportional to the longitudinal dimension l, equation (7) can be reformulated as equation (8): ^^ = min(^) = min(^) (8) Accordingly, equation (8) describes that the smaller the longitudinal dimension l, i.e., the wheelbase, the higher the directional stability.Within the framework of dimensional change efficiency, the goal is to minimize a quantity of a linear dimension change Δ^ and a quantity of a transverse dimension change Δ^. The linear dimension change Δ^ and the transverse dimension change Δ^ are defined as the difference between the final length dimension and the initial length dimension, and as the difference between the final transverse dimension and the initial transverse dimension, respectively, according to the following equation (9): where l is the final longitudinal dimension, w is the final transverse dimension, l0 is the initial longitudinal dimension, and w0 is the initial transverse dimension. To minimize the longitudinal dimensional change Δl and the transverse dimensional change Δw, the following objective function J3 for the dimensional change efficiency is defined in equation (10): ^^ = min(^^Δ^ + ^^Δ^) (10) where ^1 is a weight for the longitudinal dimensional change Δl and ^2 is a weight for the transverse dimensional change Δw. ^1 is set smaller than ^2 (i.e., ^1 < ^2) so that the longitudinal dimensional change Δl is prioritized over the transverse dimensional change Δw. This approach is chosen because the longitudinal dimensional change Δl of the AMR 1 is less movement-intensive than a transverse dimensional change Δw. This is because the change in the longitudinal dimension l occurs along the direction of movement of the AMR 1.By inserting equation (9) into equation (10), the objective function J3 for the dimensional change efficiency can finally be expressed by the following equation (11):. In equation (11), the amounts are converted into quadratic terms to simplify formulation and computer-aided processing. For multi-objective optimization, the following constraints can also be specified: where lenv and wenv are the maximum permissible longitudinal and transverse dimensions due to the environment in which the AMR 1 moves, lmin and lmax are the limitations of the longitudinal dimension of the AMR 1, and wmin and wmax are the limitations of the transverse dimension of the AMR 1. The longitudinal and transverse dimensions l and w therefore range between the minimum longitudinal and transverse dimensions of the chassis (lmin / wmin) and the minimum of the maximum longitudinal and transverse dimensions of the chassis (lmax / wmax) and the maximum longitudinal and transverse dimensions permitted by the environment (lenv / wenv). From the above equations, the following equation (13) for multi-objective optimization can be formulated: ^ = min(^ ^ ^(−^) + ^^^ + ^^[^^(^ − ^^) + ^(^ − ^^)^]) (13) Consequently, equation (14) is a quadratic optimization problem. Numerous efficient numerical algorithms exist for its solution, e.g., Active Set, Interior Point, etc., ensuring real-time capability and feasibility of the method. The effectiveness of the method is evaluated below using a simulation in an environment shown in Figures 11 and 12. The environment consists of three shelves R1 to R3, arranged in parallel rows spaced from each other. An obstacle O1 is placed in front of the first shelf R1. Another oblique obstacle O2 is placed in front of the third shelf R3. The AMR 1 intends to move through the environment along the dashed line. An initial longitudinal dimension and an initial transverse dimension of the AMR 1 are set to 0.9 m and 0.62 m, respectively. The following table shows the variables for the multi-objective optimization that result in the environment shown in Figures 11 and 12.The values of wenv are smaller than those of the actual widths w1 and w2, since a tolerance of 0.05 m must be maintained on each side. Point P1 l0 0.9 m w0 0.62 mlmin 0.9 m wmin 0.62 mlmax 1.2 m wmax 0.92 mlenv 1.5 m wenv 0.9 mPoint 3 l0 1.05 m w0 0.9 mlmin 0.9 m wmin 0.62 mlmax 1.2 m wmax 0.92 mlenv 1.5 m wenv 0.8 mWeights W. 1 5 W2 10W3 10 ^1 5^2 1 distances w1 1 m w2 0.8 m w3 0.9 m Figures 13 and 14 show a temporal progression of the transverse dimension w and the longitudinal dimension l, respectively, when the AMR 1 moves in the environment shown in Figures 11 and 12. It can be seen that initially a change occurs at waypoint P1. At waypoint P2, due to the minimization of the amount of longitudinal dimension change and the amount of transverse dimension change in the objective function J3 for the dimensional change, no change occurs efficiently, although this would be possible. At waypoint P3, the transverse dimension w and the longitudinal dimension l are then reduced again, so that the AMR 1 can travel through the distance w3 between the second shelf R2 and the second obstacle O2.Figure 15 shows the following processes when the AMR 1 moves in the environment shown in Figures 11 and 12: (a) Arrival at waypoint P1, (b) Longitudinal dimension change at waypoint P1, (c) Transverse dimension change at waypoint P1, (d) Arrival at waypoint P2, (e) Arrival at waypoint P3, (f) Transverse dimension change and longitudinal dimension change at waypoint P3, (g) Omnidirectional movement to pass obstacle O2, (h) Arrival at endpoint P4. In the environment shown in Figures 11 and 12, the AMR 1 first moves to waypoint P1. It increases its width to 0.9 m and its length to 1.05 m and continues to exit the corridor at waypoint P2. The AMR 1 then rotates the four corner modules 10 by 90 degrees and moves horizontally toward waypoint P3 to utilize its flexibility. Since the opening with the distance w3 is too narrow for the AMR 1, it reduces its outer dimension by setting its transverse dimension w to 0.8 m and its longitudinal dimension l to 1.05 m.The AMR 1 then travels with omnidirectional turning to avoid obstacle O2. Consequently, the AMR 1 can enter the corridor between shelves R2 and R3, ultimately reaching the end point P4. This test scenario shows that, taking the multi-objective optimization solution into account, the AMR 1 increases its outer dimensions at waypoint P1 to improve tipping resistance and directional stability under the existing environmental conditions. At waypoint P3, the outer dimensions are not reduced to the minimum dimensions to continue to ensure high tipping resistance. On the other hand, a further increase in the outer dimensions at waypoint P2 is omitted. Consequently, in multi-objective optimization, tipping resistance and directional stability can be efficiently prioritized over dimensional changes by setting the weights accordingly.It should be noted that the description of the preferred embodiment used the robot shown in Figure 1. However, the present invention is not limited thereto and can be applied to any autonomous mobile robot that allows for a change in the outer dimension. The change in the outer dimension corresponds to a corresponding change in the track width and edge clearance of the autonomous mobile robot used.
[0002] List of reference symbols 1 Roboter2 Front sliding module 3 Rear sliding module 4 Outer sliding module 4a First leg of the outer sliding module 4b Second leg of the outer sliding module 4c Support leg of the outer sliding module 5 Inner sliding module 5a First leg of the inner sliding module 5b Second leg of the inner sliding module 5c First support leg of the inner sliding module 5d Second support leg of the inner sliding module 6 First linear guide mechanism 7 Second linear guide mechanism 8 Third linear guide mechanism 9 Fourth linear guide mechanism 10 Corner module 11 Träger 12 Rad 13 Steering mechanism14 Control device15 Steering motor16 Gear ratio16a First gear of the gear ratio16b Second gear of the gear ratio17 Hub motor18 Rack 19 Ritzel 20 Sliding block21 Guide block22 Chassis 23 Gabell Longitudinal dimension of the chassis w Transverse dimension of the chassis P1 Waypoint P2 Waypoint P3 Waypoint P4 Endpoint R1 First shelf R2 Second shelf R3 Third shelf O1 First obstacle O2 Second obstacle w1 Distance between the first shelf and the first obstacle w2 Distance between the second shelf and the third shelf w3 Distance between the second shelf and the second obstacle
Claims
AMENDED CLAIMS received by the International Bureau on 20 June 2025 (20.06.2025) 1. A method for adjusting an outer dimension, which has a longitudinal (I) and a transverse dimension (w), of an autonomous mobile robot (1) having a reconfigurable chassis (22) that allows a change in the outer dimension, characterized in that the outer dimension of the reconfigurable chassis (22) is adjusted using a multi-objective optimization that includes tipping safety, directional stability and dimensional change efficiency as objective functions.
2. Method according to claim 1, characterized in that the objective functions are each weighted, and the tipping safety and the directional stability are weighted equally or more strongly than the dimensional change efficiency.
3. Method according to claim 1 or 2, characterized in that as a secondary condition for setting the outer dimension, a minimum adjustable dimension, which has a minimum adjustable longitudinal dimension (Imin) and a minimum adjustable transverse dimension (wmin), of the reconfigurable chassis (22) and a minimum of a maximum adjustable dimension, which has a maximum adjustable longitudinal dimension (Imax) and a maximum adjustable transverse dimension (wmax), of the reconfigurable chassis (22) and a maximum permissible dimension, which has a maximum permissible longitudinal dimension (lenv) and a maximum permissible transverse dimension (w en v) which depends on an environment in which the autonomous mobile robot (1 ) moves.
4. Method according to one of claims 1 to 3, characterized in that the objective function for the tipping safety aims at minimizing a negative value of a transverse dimension (w) of the reconfigurable chassis (22).
5. Method according to one of claims 1 to 4, characterized in that AMENDED SHEET (ARTICLE 19) the objective function for directional stability aims to minimize a longitudinal dimension (I) of the reconfigurable chassis (22).
6. The method according to any one of claims 1 to 5, characterized in that the objective function for the dimensional change efficiency aims to minimize an amount of dimensional change comprising an amount of longitudinal dimensional change (AI) and an amount of transverse dimensional change (Aw).
7. The method according to claim 6, characterized in that the objective function for dimensional change efficiency aims to minimize an amount of longitudinal dimensional change (AI) and an amount of transverse dimensional change (Aw), the amount of longitudinal dimensional change (AI) and the amount of transverse dimensional change (Aw) are each weighted, and the amount of longitudinal dimensional change (AI) is weighted more heavily than the amount of transverse dimensional change (Aw).
8. Method according to one of claims 1 to 7, characterized in that a control signal is output to at least one actuator (10, 6 - 9) of the reconfigurable chassis (22) for adjusting the outer dimension.
9. Control device (14), characterized in that it is designed and programmed to carry out the method according to one of claims 1 to 8.
10. Autonomous mobile robot (1) comprising a reconfigurable chassis (22) enabling a change in the external dimension, comprising a longitudinal (I) and a transverse dimension (w), of the autonomous mobile robot (1), characterized in that the autonomous mobile robot (1) further comprises the control device (14) according to claim 9. AMENDED SHEET (ARTICLE 19)
Citation Information
Patent Citations
Motorized motor vehicle with variable body geometry
DE19756307A1
Traveling unit, non-transitory storage medium, and system
US20210253170A1
Adjustable transportation body and a method thereof
WO2022200335A1