Balanced configuration transformation algorithm for shape-shifting modular robots

Abstract

The invention relates to the balanced configuration transformation of lattice-type shape-shifting modular robots. In particular, by constructing an intermediate structure with mechanical balance between the initial and target configurations, the invention ensures the preservation of the robot's structural integrity at every step of the configuration transformation by presenting the movement plan that generates the target configuration at the level of the modules that are the fundamental building blocks of the robot.

Description

[0001] BALANCED CONFIGURATION TRANSFORMATION ALGORITHM FOR SHAPE- SHIFTING MODULAR ROBOTS

[0002] TECHNICAL FIELD

[0003] The invention relates to the balanced configuration transformation of lattice-type shapeshifting modular robots.

[0004] In particular, by constructing an intermediate structure with mechanical balance between the initial and target configurations, the invention ensures the preservation of the robot’s structural integrity at every step of the configuration transformation by presenting the movement plan that generates the target configuration at the level of the modules that are the fundamental building blocks of the robot.

[0005] PRIOR ART

[0006] In the state of the art, many configuration change algorithms have been proposed for lattice-type shape-shifting modular robots. Two prominent examples of these algorithms are Melt Sort Grow [1] and Robomotion [2] algorithms. In the Melt Sort Grow algorithm, the implementation volume is quite large as the intermediate structure is in the form of lines. Furthermore, because this algorithm does not take into account physical constraints such as gravity in the movement of the modules, it cannot be adapted to real-world problems. On the other hand, the support columns method developed in the Robomotion algorithm considers mechanical balance. However, since the modules move through the tunnels used in this method based on the sliding-cube model principle, a system with high physical friction and energy loss has been introduced.

[0007] [1] D. Feshbach and C. Sung, Reconfiguring non-convex holes in pivoting modular cube robots, IEEE Robotics and Automation Letters, 6 (4) (2021 ) 6701 -6708. https: / / doi.Org / 10.1 109 / LRA.2021 .3095030

[0008] [2] S. Slee, J. Reif, (2010). Robomotion: Scalable, Physically Stable Locomotion for Self-reconfigurable Robots. In: D. Hsu, V. Isler, J.C. Latombe, M.C. Lin, (eds) Algorithmic Foundations of Robotics IX. Springer Tracts in Advanced Robotics, vol 68.

[0009] Springer, Berlin, Heidelberg, https: / / doi.org / 10.1007 / 978-3-642-17452-0_8

[0010] OBJECT OF THE INVENTION

[0011] The object of the invention is to provide mechanical balance by developing a plusshaped intermediate structure instead of support columns in the balanced configuration transformation of lattice-type shape-shifting modular robots. Thus, the problem of the large volume required by the line-shaped intermediate structure applied in the prior art has been solved. Again, instead of the sliding cube model applied in the state of the art, a pivoting cube model is used, resulting in a highly energy efficient and simple system.

[0012] Other objects of the invention are as follows:

[0013] • Creating a simple and efficient algorithm,

[0014] • The developed algorithm, with the mechanical balance it provides, not allowing the structure to topple to the ground and any module to break away from the structure,

[0015] • Reducing the configuration space to a very low level by not allowing misconfigurations,

[0016] • Improving the task completion capability of the lattice-type shape-shifting robot system through balanced shape-shifting operation,

[0017] • Thanks to the pivoting cube model, preventing energy loss due to friction force.

[0018] LIST OF FIGURES

[0019] Fig. 1 a. First configuration created for lattice-type modular robots

[0020] Fig. 1 b. Second configuration created for lattice-type modular robots

[0021] Fig. 2. Illustration of a modular robot system with three degrees of freedom

[0022] Fig. 3a. Exploded illustration of a module and the neighboring modules to which it can dock

[0023] Fig. 3b. Assembled illustration of a module and the neighboring modules to which it can dock Fig. 4. Schematic representation of the elements contained in a self-sufficient module

[0024] Fig. 5. Schematic representation of the elements required to communicate with a module

[0025] Fig. 6. Block diagram of a shape-shifting modular robotic system

[0026] Fig. 7. Operational flow of the method subject to the invention

[0027] Fig. 8. Operational architecture of the method subject to the invention

[0028] Fig. 9. Stages of determining the order of module selection

[0029] Fig. 10. A flow diagram of the method subject to the invention

[0030] References for the numbering given in the figures:

[0031] 1.1. Quadruped robot configuration

[0032] 1 .2. Snake-like robot configuration

[0033] 2.1 . A modular robot system with three degrees of freedom

[0034] 2.2. A module capable of pivoting along three different axes

[0035] 2.3. A module capable of pivoting along three different axes

[0036] 2.4. A module capable of pivoting along three different axes

[0037] 3.1 . A module capable of docking

[0038] 4.1 . Example module

[0039] 4.2. Power source

[0040] 4.3. Communication system

[0041] 4.4. Actuators

[0042] 4.5. Docking

[0043] 4.6. Memory

[0044] 4.7. Data signal flow (I / O system)

[0045] 4.8. Control unit

[0046] 5.1 . A module capable of communication

[0047] 5.2. Local processor

[0048] 5.3. Sensors and actuators

[0049] 5.4. Local status

[0050] 5.5. Central controller command

[0051] 5.6. Status indication signal

[0052] 6.1 . A modular robotic system

[0053] 6.2. Central control unit

[0054] 6.3. Identical modules 7.1 . Operational flow of the method

[0055] 7.2. Initial configuration

[0056] 7.3. Target configuration

[0057] 7.4. Intermediate configuration

[0058] 8.1 . Operational architecture of the method subject to the invention

[0059] 8.2. Inputs

[0060] 8.3. Central controller

[0061] 8.4. Outputs

[0062] 8.5. Initial configuration

[0063] 8.6. Target configuration

[0064] 8.7. Balance control

[0065] 8.8. Module selection order

[0066] 8.9. Module movement plan

[0067] 8.10. Configuration transformation plan

[0068] 9.1 . Stages of determining the order of module selection

[0069] 9.2. Identification of active modules in the layer

[0070] 9.3. Calculation of the distance of active modules to the center module

[0071] 9.4. Determination of selection rings from outside to inside according to distances

[0072] 9.5. Creation of a list with the selection order of active modules

[0073] 10.1 . A flow diagram of the method subject to the invention

[0074] DETAILED DESCRIPTION OF THE INVENTION

[0075] The steps of the method subject to the invention are as follows:

[0076] - Start the algorithm.

[0077] - Enter start and end configurations into the system.

[0078] - Check that whether the entered configurations are balanced.

[0079] - If it is balanced, start the module selection process. (If it is not balanced, go back to step 2)

[0080] - Depending on the module selection, create a shape-shifting plan.

[0081] - Construct a balanced intermediate structure by applying the shape-shifting plan.

[0082] - Finish the algorithm. The algorithm of the invention includes method steps to maintain the mechanical balance of shape-shifting modular robots during configuration transformation. The method involves the coordinated relocation of modules capable of forming different configurations by physically connecting or disconnecting from each other, controlled from a central unit. These modules have the ability to move individually in line with the commands from the central unit thanks to their internal actuator, battery, and electronic card. The method is applicable to all cubic modular robotic systems with a lattice-type structure.

[0083] The physical combination of modules with the ability to act independently enables various configurations to be created, enabling a system to perform multiple tasks. By changing the shape of the robotic system, different tasks can be performed, such as passing through narrow passages, obstacle jumping, object transportation, and bridge building. In addition to these tasks, shape-shifting modular robots are preferred in conditions where human intervention is difficult, such as firefighting, working in areas exposed to radiation, search and rescue, and field research. For example, in the task of identifying the survivors trapped under the rubble and delivering aid, a modular robot can detect the gaps under the rubble, adapt its shape according to these gaps and fulfill the task quickly. Another example can be found in the defense industry. Modules that can move independently can separate from the main structure and perform scanning in different directions and provide instant data flow to the main structure. Multiple modules come together to overcome obstacles encountered in rough terrain. Thus, it is possible to perform a faster and more efficient field survey work than traditional fixed-structure robots. Additionally, camouflage of the modules in terrain conditions will be easier.

[0084] Figs. 1a and 1b show two different configurations for lattice-type modular robots. The quadruped (1.1 ) and snake-like (1.2) robots are built by assembling autonomous cubic modules. These autonomous modules are capable of connecting and disconnecting from each other to form different robotic configurations. If all modules in the robotic system are identical to each other, the system is defined as homogeneous. Otherwise, a heterogeneous system emerges. The homogeneity of the system is advantageous in terms of ease of operation. Homogeneity is analyzed under three different headings: physical, hardware, and software. Physical homogeneity means that all modules are dimensionally identical to each other. In hardware homogeneity, all hardware in the modules is expected to be the same. Software homogeneity is achieved when all modules are programmed in the same way. As homogeneity increases in a robotic system, production costs decrease and the replacement of defective modules with new modules becomes easier. In the method subject to the invention, an algorithm is proposed to be used in a physically, hardware, and software homogeneous robotic system.

[0085] In order to increase task diversity and reduce computational burden, group movement of autonomous modules (meta-module) and parallel pivot movement approaches can be applied. Within a large modular robotic system, a group of modules can be defined as a gripper meta-module and object manipulation can be performed by this meta- module. Thus, thanks to a specialized meta-module within the system, it is possible to avoid the cost of creating a holder of modules every time object manipulation is required. On the other hand, the parallel pivot approach, which allows multiple modules to move simultaneously, reduces the computational burden while providing flexibility in task completion. For example, a robotic system assigned a field survey task drives its modules, which are separated from the main structure, into the field individually and sends a command to each module to scan a different area. Modules that scan different areas simultaneously can transmit information to the main structure instantly. Samples collected by the modules can be combined during docking to the main structure. The proposed algorithm can be easily adapted to these approaches.

[0086] An autonomous module includes components such as a controller, processor, software element, communication element, power supply, sensors, actuators and connector elements. The combination of self-sufficient modules creates a shape-shifting modular robot. In the modular robot configuration, the position of the modules can be changed autonomously and the modules can communicate among themselves.

[0087] Fig. 2 shows a modular robot system (2.1) with three degrees of freedom. In this robot system (2.1), the modules are connected to each other in a lattice-type configuration. The intersecting edges of two modules can be selected as the pivot axis. Fig. 2 shows the modules (2.2, 2.3, 2.4) that can pivot in three different axes. Each module with the appropriate border has the ability to pivot in all three axes. Pivot movement can take place in two different ways, at 90 or 180 degree angles. The angle at which the module pivots is determined by the neighboring module. In the directions shown, modules numbered 2.2 and 2.4 can pivot at 90 degrees and module numbered 2.3 can pivot at 180 degrees. The configuration transformation of the robotic system (2.1 ) is realized through meaningful sequences of pivot movements. In the method subject to the invention, two types of pivot movements can also be used.

[0088] Figs. 3a and 3b show a module (3.1) and the neighboring modules it can dock with. As can be seen, a module can be docked with up to six modules in total at the same time. In the prior art, mechanical and magnetic systems are usually proposed for docking. While mechanical docking mechanisms provide high bond strength, they increase the complexity of the modular robot system. On the other hand, although magnetic docking systems are easy to implement, the bond strength they provide is relatively weak. Energy consumption and the need for continuous energy supply are taken into account when designing docking mechanisms. The energy consumption of mechanical docking mechanisms is lower than that of magnetic docking mechanisms using electromagnets. For mechanisms using electromagnets, a continuous power supply is required. On the other hand, docking mechanisms using permanent magnets do not require a continuous power supply, nor do they consume energy. In these systems, however, the power provided by the actuator must overcome the bond strength. The method subject to the invention can work in any type of docking mechanism.

[0089] The surfaces where the modules are connected to each other can be used for docking as well as for tasks such as communication and energy transfer between modules. This transmission can be achieved through metal pins placed on the surfaces. Energy sharing can be critical if a module's battery runs out. Modules can communicate directly or via wireless connection. Different communication elements can be used for close (inter-module) and remote (between module and central controller) wireless communication. While infrared elements can be used for close communication, wireless link area (Wi-Fi) elements are preferred for remote communication.

[0090] In shape-shifting modular robots, it is desirable to keep the module size at levels that are as low as possible. If the actuator power can be kept constant while the module size is reduced, the power per module can be increased. However, miniaturizing the module size is subject to some restrictions. The power consumption of the module internal parts is one of the main size constraints. On the other hand, performance expectations, desired module weight, available actuator, and battery sizes are other important constraints.

[0091] Fig. 4 shows the elements contained in an example self-sufficient module (4.1 ). This example module (4.1 ) has a battery (4.2) to supply energy for all its operations, an I / O system (4.7) to regulate signal and data flow within the module, and a controller (4.8) as high-level regulators. The communication system (4.3), actuators (4.4), docking mechanism (4.5), and memory unit (4.6) controlled by these regulators are located in the module (4.1).

[0092] The example module (4.1 ), whose elements are shown, can be connected with other modules to form a large structure, or it can be part of a swarm of robots moving together without connection. The method of the invention has been developed for modules that can be connected to each other and form lattice-type structures.

[0093] In the method of the invention, the number of modules that the robotic system has is at a countable level. Module weight and energy consumption can be reduced to optimum level in terms of hardware. In order to find the optimum level in software, efficient algorithms must be used.

[0094] The configuration transformation of a lattice-type robotic system is possible by applying the movements of the modules from one position to another in a meaningful sequence. Maintaining the balance of the main structure is an important consideration when the modules are displaced. If the main structure is not balanced, the robotic system will collapse and the whole process will have to be restarted. The method subject to the invention ensures that the main structure balance is maintained while the modules are displaced, and that the configuration transformation can be carried out safely without allowing any module to be detached from the robotic system.

[0095] The power supply (4.2) element shown in the example module (4.1 ) can be a combination of one or more batteries, or it can be a voltage regulator or power distributor that regulates the power supplied from outside and makes it available within the module. In the method subject to the invention, since the spatial positions of the modules will change frequently while a large number of pivot movements are performed sequentially, it would be appropriate to place the batteries in the module without any external power cable input.

[0096] The communication system (4.3) includes various means of communication to ensure close or remote connections. For inter-module communication, optical means are the most common form of communication. Infrared means are often used for short distance connectivity. On the other hand, Wi-Fi modules and radar means are among the leading long-range communication means. The method subject to the invention uses a long distance connection between the central controller (4.8) and the modules. Since all commands come from the controller, there is no need for inter-module communication.

[0097] As shown on the example module (4.1), the main component enabling the movement of the modules are the actuators (4.4). Although various types of actuators are used for module movement in the field of modular robot systems, the most common actuators used in the prior art for lattice-type modular robots are commercially available servo or brushless DC motors. In addition, the prior art also includes unique actuators based on electromagnet thrusters. Depending on the area of use of the module, the torque that the actuator must provide may vary. For example, in a system developed for the space environment, the thrust generated by electromagnets would be sufficient, while in systems developed for gravity-dominated environments, the high torque generated by brushless motors would be sufficient. The method subject to the invention is suitable for working with all kinds of actuators that provide pivot movement.

[0098] The control unit (4.8) contains a low-power microprocessor that executes programs, performs data analysis and control functions within the module. The control unit (4.8) can contain more than one microprocessor. These microprocessors are contained in one or more microchips and can be placed on one or more printed circuit boards. For example, the control unit (4.8) can act as a central processor unit that detects, processes, and reacts appropriately to external control commands, or it can act as a computing unit that performs simple mathematical operations.

[0099] The memory element (4.6) can contain cache, volatile memory or non-volatile memory types separately or together. The memory (4.6) and I / O system (4.7) elements can be integrated together or individually into the control unit (4.8). The I / O system (4.7) can contain a sub I / O system (4.7) for elements inside the module

[0100] (4.1 ). The I / O system (4.7) can include I / O ports, communication elements, and sensors. Communication elements may have additional equipment such as antennas and Global Positioning System (GPS). Chemical sensors, image processing cameras, sound detecting microphones, distance sensing infrared sensors are some of the main sensors that can be used for module (4.1 ).

[0101] The ports used in the I / O system (4.7) can be divided into two categories: communication and non-communication ports. Data can flow through ports used for communication, while inter-module power sharing can be done through ports not used for communication.

[0102] Fig. 5 shows the elements required to communicate with a communication module

[0103] (5.1 ). Fig. 6 shows the block diagram of a shape-shifting modular robotic system (6.1 ). The modules (6.3) in the robotic system (6.1 ) are docked to each other and can communicate with each other and with a central control unit (6.2). Modules (6.3) are identical and contain the communication elements shown in Fig. 5. The inter-module communication network of a modular robotic system presents the network topology of that system. If the modules (6.3) are interconnected, the network topology also provides the configuration information of the robotic system. When a module (6.3) is displaced within the robotic system (6.1 ), the network topology dynamically updates itself. The dynamic communication network is important for the robotic system (6.1 ) to react to unexpected situations and helps the modules (6.3) to jointly perform tasks such as configuration transformation, displacement, and object manipulation. The method subject to the invention is capable of performing the aforementioned tasks thanks to the communication network managed by the central control unit (6.2).

[0104] The robotic system (6.1 ) can adopt any of the centralized, decentralized and distributed control approaches while performing the specified tasks. In the centralized control approach, all commands are transmitted to the modules by the central control unit

[0105] (6.2). In decentralized and distributed control approaches, there is no need for a central control unit (6.2), and the control function is carried out through an inter-module communication network. The method subject to the invention adopts a centralized control approach. The most important object of the method subject to the invention is to maintain the physical balance of a modular robotic system during shape-shifting. The effect of gravity is taken into account when analyzing physical balance. Thus, the method subject to the invention can be used in real world problems. A modular robotic system realizes shape-shifting through the movements of its modules. As each module is displaced on the configuration, it avoids movements that would disturb the balance of the configuration. The periodic contraction and expansion steps used by jellyfish during their movement helped in the development of the method subject to the invention. In this respect, the method is inspired by nature. Fig. 7 shows the operational flow (7.1 ) of the method. In the initial configuration (7.2) the robotic system is in the expanded position. Thanks to the sequential module movements, a balanced intermediate configuration (7.4) in the shape of a "plus" is built and the robotic system contracts. Finally, from the balanced intermediate structure (7.4), the target configuration (7.3) is reached by sequential module movements and the robotic system returns to the expanded position. Thanks to the balanced intermediate configuration, the physical balance of the robotic system is maintained at every stage of the shape-shifting process, preventing possible tipping situations. In a system consisting of a large number of modules, in the event of tipping, the task cannot be completed successfully and setbacks such as loss of time and module failure will be encountered.

[0106] The operational architecture of the method (8.1) subject to the invention is shown in Fig. 8. The inputs (8.2) of the method are the initial (8.5) and final (8.6) configurations. The central controller (8.3) checks whether the inputs are balanced (8.7). After confirming that the inputs are balanced, the order in which the modules (6.3) will be selected is decided (8.8). If the inputs are not balanced, the controller (8.3) requests physically balanced inputs from the user. Once the order of selecting module (6.3) is determined, a movement plan is created for each module (6.3) (8.9). The output (8.4) of the method is the configuration transformation plan (8.10) of the robotic system. This plan is implemented according to the operational flow (7.1) shown in Fig. 7. The method subject to the invention operates in the forward direction for the initial (7.2) and intermediate (7.4) configurations and in the reverse direction for the target (7.3) and intermediate (7.4) configurations. To clarify, the shape-shifting plan to reach the intermediate configuration (7.4) from the target configuration (7.3) is inverted and added to the end of the shape-shifting plan to reach the intermediate configuration (7.4) from the initial configuration (7.2), and the total shape-shifting plan (8.10) is obtained. The method subject to the invention utilizes the geometric properties of the configuration to determine the order of module selection. The height of the configuration is determined and each layer is numbered. The horizontal dimensions of the configuration are determined and the center point is determined. Taking the center point as a midpoint reference, a "plus" shape is created according to the width. This "plus" shape is the core of the intermediate configuration and is maintained during the shape-shifting process. The modules in the core are named passive modules. Passive modules are immobile and their task is to maintain balance. The modules outside the core in each layer are the active modules and it is the movement of these active modules that enables the shape-shifting process. Fig. 9 shows the stages of determining the order of module selection (9.1). Module selection is done from the upper layer to the lower layers. The first step is to identify the active modules in each layer (9.2). The distances of the active modules from the center of the core are then determined (9.3). Selection rings are determined from outside to inside according to the distance values (9.4). Module selection within the selection rings is determined in a counterclockwise direction. Finally, the general module selection order is determined from top to bottom and from outside to inside in a counterclockwise direction (9.5). According to the module selection order, each module is sent in turn to the position specified in the intermediate configuration. The shortest distance between the start and target locations of the module is determined using the Wide Priority Search algorithm.

[0107] In the method subject to the invention, the movement command (5.5) is sent to the modules (5.1) by the central control unit (6.2). The local processor (5.2), which receives the command, pivots in the appropriate direction by means of the actuators (5.3) in the module and reaches the target point. After each pivot movement, the module (5.1 ) determines its status by means of its sensors (5.3) and reports this status to the local status element (5.4). The local processor (5.2), which receives the status information from the local state element (5.4), reports its status (5.6) to the central control unit after each pivot movement. Thus, the method subject to the invention can be implemented with the help of coordination between the central control unit (6.2) and the module (5.1).

[0108] The flow diagram of the method subject to the invention is shown in Fig. 10. In this flow diagram, all the components of the method described in detail above are interrelated. The above-mentioned method subject to the invention is applicable to different real- world problems due to the balanced configuration transformation it provides. Various tasks such as providing diversity that allows a robotic hand to grasp different objects, displacement of a modular robotic system by shape-shifting, dynamic bridge construction for crossing obstacles can be solved by the method subject to the invention. In certain tasks, the method subject to the invention can be improved with meta-module and parallel pivot movement approaches.

Claims

CLAIMS1 . A balanced configuration transformation algorithm developed for modules that host a central control unit, interconnect with each other and form lattice-type structures, characterized in that it comprises the process steps of:- Taking the initial (8.5) and final (8.6) configurations of the robotic system as inputs of the algorithm (8.2),- Checking that whether the inputs are balanced by the central controller (8.3),- After confirming that the inputs are balanced, deciding on the order in which the modules will be selected,- If the inputs are not balanced, requesting from the user for inputs that are physically balanced by the controller (8.3),- Creating a movement plan for each module after determining the order of module selection,- Sending each module in turn to the position specified in the intermediate configuration according to the module selection order.

2. The step of deciding in which order the modules according to claim 1 will be selected, characterized by the stages of:- Determining the height of the configuration and numbering each layer,- Determining the horizontal dimensions of the configuration and determining the center point,- Creating a "plus" shape with respect to the width, taking the center point as a midpoint reference, and defining this "plus" shape as the core of the intermediate configuration,- Naming the modules in the core as passive modules,- Naming the cores outside the modules in each layer active modules,- Identifying the active modules in each layer, from the upper layer to the lower layers,- Then determining the distance of the active modules from the center of the core,- Determining the selection rings from outside to inside according to distance values,- Determining the module selection in the selection rings in a counterclockwise direction - Finally, determining the general module selection order from top to bottom and from outside to inside in a counterclockwise direction.