Design and Development of Modular Limbs for Reconfigurable Robots on the Moon

Abstract: In this paper, we present the development of 4-DOF robot limbs, which we call Moonbots, designed to connect in various configurations with each other and wheel modules, enabling adaptation to different environments and tasks. These modular components are intended primarily for robotic systems in space exploration and construction on the Moon in our Moonshot project. Such modular robots add flexibility and versatility for space missions where resources are constrained. Each module is driven by a common actuator characterized by a high torque-to-speed ratio, supporting both precise control and dynamic motion when required. This unified actuator design simplifies development and maintenance across the different module types. The paper describes the hardware implementation, the mechanical design of the modules, and the overall software architecture used to control and coordinate them. Additionally, we evaluate the control performance of the actuator under various load conditions to characterize its suitability for modular robot applications. To demonstrate the adaptability of the system, we introduce nine functional configurations assembled from the same set of modules: 4DOF-limb, 8DOF-limb, vehicle, dragon, minimal, quadruped, cargo, cargo-minimal, and bike. These configurations reflect different locomotion strategies and task-specific behaviors, offering a practical foundation for further research in reconfigurable robotic systems.

• The paper introduces highly modular 4-DOF robotic limbs that serve as adaptable building blocks for reconfigurable lunar robots.
• It details a robust hardware and three-layered ROS 2-based control architecture, achieving precise joint responses and high torque performance.
• Experimental validation across nine distinct configurations demonstrates the system's versatility for lunar exploration and construction tasks.

Modular Robotic Limbs for Reconfigurable Lunar Robots

System Overview and Design Rationale

This paper presents the design and experimental validation of highly modular 4-degree-of-freedom (DOF) robotic limbs, specifically crafted for self-reconfigurable robots with application relevance to lunar exploration and construction in resource-constrained extraterrestrial environments (2601.04541). The modules, denoted as "Moonbots," serve as building blocks for diverse robotic morphologies, capable of functioning as limbs, manipulators, structural elements, or suspension/steering links depending on the task.

The structural design is centered on macro-modular and heterogeneous system principles, allowing for chain-like assembly or hybrid integration with specialized wheel modules. All actuation is standardized through a custom-developed, compact actuator with high torque-to-speed ratio (over 75 Nm torque, up to 26 rpm speed) and precise control, supporting both manipulation and dynamic locomotion. Mechanical modularity is achieved with symmetric roll–pitch–pitch–roll kinematic layouts and genderless gripper mechanisms, which enable robust inter-module connections and bidirectional limb association.

Figure 1: Developed 4-DOF modular limbs, with a palette arrangement, wheel module grasp, and dual-limb 8-DOF manipulator assembly.

Figure 2: Internal view of the limb and actuator components, detailing motor, harmonic drive, and control electronics integration.

Hardware and Software Architecture

Each limb integrates a BLDC outrunner motor paired with a harmonic drive, achieving minimal backlash and high torque density. The base electronics include a LattePanda SBC, dual DC-DC power converters, LiPo batteries, and CAN-daisy-chained ODrive motor controllers. The communication protocol utilizes CAN for intra-limb coordination and wireless networking for teleoperation and supervisory control.

Figure 3: Hardware architecture schematic for the limb module, illustrating the electronic and communication topology.

Control software is structured into three hierarchical abstraction layers:

• Low level: Direct hardware drivers utilizing ROS 2 and position/velocity/current feedback loops with nested FOC controllers at 8 kHz.
• Middle level: IK solvers and trajectory planners (including the Motion-Stack framework), enabling synchronized joint control and multi-limb clustering.
• High level: Teleoperation interfaces, supporting joystick, keyboard, and planned extensions for RL-based learning controllers.

  Figure 4: Three-layered ROS 2-based software architecture for multilimb control.

  

  Figure 5: Implemented ROS 2 nodes interfacing with joints, drivers, teleoperation, and IK routines.

Actuator Performance Evaluation

Quantitative assessment of actuator modules covered static load endurance and dynamic response profiling:

• In static tests, limb actuators maintained payloads up to 73 Nm with negligible positional drift (<0.05 revolutions at peak load), while remaining well below the rated continuous current limits.
• For dynamic response, three position control modes were benchmarked. The "trapezoidal trajectory" mode provided smooth, damped transitions without velocity spikes and was selected for operational deployment.

Figure 6: Actuator current profile and position tracking under incremental static loads.

Figure 7: Step input position and velocity responses across different control modes.

In assembled limb modules, all actuators demonstrated synchronized concurrence and efficient attainment of joint targets, including under composite movement and Cartesian pose objectives. Response times for full-limb movements via IK commands were sub-0.5 seconds, with smooth velocity curves and no observable overshoot.

Figure 8: Simultaneous multi-joint step response in fully assembled limb module.

Figure 9: Step response for actuator position control under different input control modes.

Experimental Reconfigurability and Morphology Demonstrations

The modular limbs and wheel modules were assembled into nine distinct configurations, validating the system's versatility and mechanical robustness under real-world field conditions (including sand and unstructured lunar analog terrain):

• Limb to Limb Connection: Sequential trajectory execution and IK alignment for dual-limb handshake, forming a redundant 8-DOF manipulator.

  Figure 10: Sequential actuation steps for limb-to-limb connection and mechanical handshake.

• Limb to Wheel, Vehicle, Dragon, Cargo Configurations: Rapid reconfiguration into single-limb + dual-wheel vehicles, multipurpose cargo carriers (with up to four wheels and four limbs), and hybrid manipulator-mobility morphologies.

  Figure 11: Limb to wheel connection sequence, assembling minimal mobile configuration.

  

  Figure 12: Vehicle configuration (suspension, steering, transformable modes).

  

  Figure 13: Dragon configuration, enabling mobile manipulation after pallet departure.

  

  Figure 14: Cargo configurations, showing single and dual-wheel modes, with payload demonstration.

• Spinbot, Quadruped, Bike Modes: Exploratory configurations for constrained-environment traversal, rough terrain negotiation, and narrow-passage mobility, some requiring further control sophistication via IMU-based balance algorithms.

  Figure 15: Spinbot traversing sand, quadruped morphology, bike mode for passage navigation.

Numerical Results and Strong Claims

• The actuator's payload tolerance proved robust, with successful maintenance of up to 73 Nm static load while remaining below rated current—a clear indication the hardware exceeds requirements for both manipulation and legged locomotion in severe lunar conditions.
• The vehicle mode achieved a top speed of approximately 1.0 m/s with full dual-wheel cargo payload (solar panel station mock-up), validating operational feasibility for substantial transport tasks.
• Reconfiguration operations were completed via teleoperation and preprogrammed joint-trajectory/IK commands, with discrete transitions (e.g., vehicle suspension-to-steering mode) performed without mechanical disconnect, strengthening claims for operational flexibility and minimization of downtime.
• The configuration modularity enables functionally reconfigurable robotics, facilitating adaptation to task requirements without system-level homogeneity constraints or structural redesign.

Theoretical and Practical Implications

This research augments the taxonomy of macro-modular robotics, showcasing physical symmetry, genderless connectors, actuator unification, and distributed control as critical features for real-world reconfigurable robot deployment in extreme and resource-sparse environments. The work demonstrates that modularity-driven adaptation outperforms extended joint range strategies for multiple mission profiles, including manipulation, mobility, rough terrain traversal, and payload transport.

From a control perspective, the abstraction over IK/trajectory synchrony and hardware interface scalability paves the way for the incorporation of learning-based and autonomous reconfiguration strategies, promising improved resilience and task autonomy in future lunar missions.

Speculation on Future Developments

Long-term research directions include autonomous self-assembly, sensing augmentation for real-time navigation and manipulation, and reinforcement learning integration to endow the robots with more adaptive behaviors. The modular design is expected to enable not only planetary exploration but also terrestrial operations in hazardous or unpredictable environments (disaster response, construction, logistics). Expansion of connector designs, module heterogeneity (specialized sensors/manipulators), and energy management will likely drive next-generation evolutions in modular robotics.

Conclusion

The presented system achieves a substantial advance in modular robotic flexibility for lunar exploration contexts, combining robust hardware—standardized high-torque actuators and symmetric limb modules—with a composable software stack and proven experimental field performance. Validated across nine operational configurations, this approach establishes a credible foundation for future research in autonomous, reconfigurable, multi-purpose robotic systems suitable for extraterrestrial construction and exploration, with extension potential to demanding terrestrial applications.