Transformable Modular Robots: A CPG-Based Approach to Independent and Collective Locomotion
Jiayu Ding, Rohit Jakkula, Tom Xiao, Zhenyu Gan · 2025
Read Transformable Modular Robots: A CPG-Based Approach to Independent and Collective Locomotion by Jiayu Ding, Rohit Jakkula, Tom Xiao, Zhenyu Gan (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-13674.txt.
Introduction
Jiayu Ding and co-authors build small robots that can join up, split apart, and move either alone or as a team. Each unit has its own motor, battery, and control, so it can roll on its own or lock with others for group travel.
A two level control scheme based on rhythm patterns runs the motion. One level steers each unit, and a higher level keeps units in step with each other. Tests in a physics simulator and with real hardware cover one unit alone and pairs working together. The results back the design for jobs like search and rescue or field sensing.
Catalog record
| CREATOR | Jiayu Ding, Rohit Jakkula, Tom Xiao, Zhenyu Gan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-13674
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Transformable Modular Robots: A CPG-Based Approach to Independent and Collective Locomotion Jiayu Ding, Rohit Jakkula, Tom Xiao, Zhenyu Gan Abstract Modular robotics offers a promising approach for developing versatile and adaptive robotic systems capable of autonomous reconfiguration. This paper presents a novel modular robotic system in which each module is equipped with independent actuation, battery power, and control, enabling both individual mobility and coordinated locomotion. The system employs a hierarchical Central Pattern Generator (CPG) framework, where a low-level CPG governs the motion of individual modules, while a high-level CPG facilitates inter-module synchronization, allowing for seamless transitions between independent and collective behaviors. To validate the proposed system, we conduct both simulations in MuJoCo and hardware experiments, evaluating the system’s locomotion capabilities under various configurations. We first assess the fundamental motion of a single module, followed by two-module cooperative locomotion. The results demonstrate the effectiveness of the CPG-based control framework in achieving robust, flexible, and scalable locomotion. The proposed modular architecture has potential applications in search-and-rescue operations, environmental monitoring, and autonomous exploration, where adaptability and reconfigurability are essential for mission success. I Introduction Modular robotics has emerged as a promising approach for developing versatile robotic systems that can dynamically adapt to varying task requirements and environmental constraints [ 1 , 2 , 3 ] . Unlike conventional single-unit robots with rigid structures and predefined functionalities, modular robots consist of multiple self-contained units that are capable of independent movement and cooperative operation to achieve collective behaviors. These systems demonstrate significant potential in applications requiring adaptability, scalability, and robustness, such as search-and-rescue [ 4 ] , planetary exploration [ 5 ] , and autonomous manufacturing and inspection [ 6 ] , where the ability to reconfigure and recover from failures is critical for mission success. A key advantage of modular robots is their ability to autonomously reconfigure and self-assemble into different morphologies based on task demands. As illustrated in Fig. 1 , a single module is capable of independent operation and wireless communication with other modules, enabling the formation of larger robotic structures with enhanced capabilities. By dynamically assembling into different configurations, modular robots can be adapted for various applications, including search-and-rescue missions, package transportation, and industrial manipulation. Existing self-reconfigurable systems such as PolyBot [ 7 ] , M-TRAN [ 8 ] , and SMORES-EP [ 9 ] have demonstrated significant progress in achieving these capabilities. However, challenges remain in optimizing control strategies that allow seamless transitions between independent and coordinated behaviors. Figure 1: Overview of the proposed modular robot. (A) A single module operates independently and communicates wirelessly with other modules. By assembling into more complex structures, the system can be adapted for various applications, including (B) search-and-rescue missions, (C) package transportation, and (D) industrial manipulation tasks. Incorporating biologically inspired control mechanisms , particularly Central Pattern Generators (CPGs) , has significantly improved modular robotic locomotion. CPGs, neural networks capable of generating rhythmic patterns, have been widely used in controlling legged and serpentine robots [ 10 , 11 , 12 , 13 , 14 ] . Unlike traditional trajectory-based controllers, CPG-based approaches inherently provide robustness against disturbances and adaptability to environmental variations. Recent studies have extended CPG models to modular robots, demonstrating their effectiveness in achieving synchronized locomotion across multiple connected modules [ 15 , 16 ] . Another critical aspect of modular robotics is decentralized control and wireless communication , which enhances system robustness and scalability. Unlike centralized architectures that rely on continuous wired computation and data exchange, decentralized approaches allow individual modules to function autonomously while coordinating with others as needed. This autonomy improves system resilience; if a module malfunctions or becomes damaged, another module can be reprogrammed to assume its role, ensuring continuity in task execution. Such fault tolerance is difficult to achieve in monolithic robot designs, where a single point of failure can lead to complete system failure [ 17 , 18 , 1 ] . Research in multi-agent communication has explored protocols such as MQTT and UDP to enable real-time synchronization among modules, providing a scalable framework for distributed modular robotic systems. Building upon these advancements, this paper presents a novel modular robotic system with independent actuation, battery power, and control, allowing each module to function autonomously or as part of a larger coordinated structure. A hierarchical CPG-based control framework is employed, where a low-level CPG governs the motion of individual modules, while a high-level CPG enables inter-module synchronization. This structure allows smooth transitions between independent and collective behaviors, facilitating efficient adaptation to diverse environments. The proposed system is validated through MuJoCo [ 19 ] simulations and real-world experiments, showcasing locomotion capabilities across different modular configurations, including single-module movement and two-module cooperative locomotion. Additionally, a preliminary experiment with a four-module configuration is presented in the discussion section. The remainder of this paper is organized as follows. Section II describes the mechanical design, control architecture, and networking of the modular system. Section III presents experimental validation and simulation results. Finally, Section IV discusses the conclusions and future directions of this research. II Methods This section describes the design and CPG-based controller implementation of the proposed modular robotic system. Each module is designed to operate independently while also supporting coordinated locomotion through a local Wi-Fi network. This section consists of the mechanical design, actuation system, and embedded control architecture. II-A Modular Robot Design and Networking The design of the proposed module adheres to the principles of modularity, ease of assembly, and rapid deployment. As shown in Fig. 2 (A), the j -th module is modeled as a floating-base system with six degrees of freedom (DoF) at its geometric center, allowing three translational and three rotational motions, represented as \mathbf{\boldsymbol{q}}_{c}^{j}=[x,y,z,q_{x},q_{y},q_{z}]^{T} . Each module incorporates five actuated joints with angles \mathbf{\boldsymbol{q}}_{b}^{j}=[q_{1},q_{2},q_{3},q_{4},q_{5}]^{T} , where each joint operates within the range [-\nicefrac{{3}}{{4}}\pi,\nicefrac{{3}}{{4}}\pi]\penalty\ [rad] , providing the flexibility needed for both independent locomotion and cooperative transformations. The positive direction is counterclockwise, with zero angles defined when the robot is in its flat configuration, as illustrated in Fig. 2 (A). The structural arrangement allows the module to execute a variety of motion primitives, including bending, rotation, and extension, which are essential for both autonomous locomotion and reconfiguration when multiple modules connect and interact. The joint orientations follow a bio-inspired kinematic configuration, balancing flexibility and stability: • q_{1} and q_{5} facilitate rolling movements about q_{x} , • q_{3} enables lateral pitching motion about q_{y} , • q_{2} and q_{4} provide primary yaw motion about q_{z} . Each joint is actuated by an RDS3235 High Torque Digital Servo [ 20 ] , a coreless motor servo with all-metal gears, ensuring high durability and precision. The servos operate at a stall torque of 35\text{\,}\mathrm{k}\mathrm{g}\,\mathrm{c}\mathrm{m} at 7.4\text{\,}\mathrm{V} , providing sufficient actuation strength for both individual locomotion and cooperative transformations. The module’s mechanical structure consists of two lightweight frames, 3D-printed using PLA filament, offering a balance between strength, flexibility, and low weight. These frames house the controller board and a 7.4\text{\,}\mathrm{V} , 3000\text{\,}\mathrm{m}\mathrm{A}\mathrm{h} , 15C 2S Li-ion battery pack, ensuring a compact, modular design while maintaining structural integrity. To enable rapid and secure mechanical coupling, each module is equipped with easy-locking male and female connectors on both ends. These connectors allow for quick attachment and detachment, enabling multiple modules to be assembled