Abstract: In this chapter, based on analyzing the characteristics of the working target, a new kind of autoclimbing robot is proposed, which is used for cleaning the spherical surface of the National Grand Theatre in China. The robot's mechanism and its unique aspects are presented in detail. A distributed controller based on a CAN bus is designed to meet the requirements of controlling the robot. The control system is divided into six parts, five CAN bus control nodes and a remote controller, which are designed and established based mainly on the microprocessor P80C592. Then the motion functions are described in detail. From the safety point of view, the process of climbing from one strip to another is very dangerous because the robot has to adapt to the outer shape of the building. Additionally, this robot has to be quite big in order to realize all necessary functions. Therefore, in designing this mechanism and its controls, an equilibrium between safety and size has to be reached. This is why the kinematics model of the climbing process is introduced. For system design and control purposes, the dynamics of the robot are calculated by applying the Lagrange equation. The force distribution of the front and rear supporting mechanisms is computed in a way that ensures the safety of the climbing process. After that, a new approach to path planning for wall-cleaning robots is presented considering movement security, cleaning efficiency and the percentage of cleaning coverage. The successful on-site test confirms the principles described above and the robot's ability to work on a spherical surface.
Introduction: The "3D" Challenge of Modern Architecture
Cleaning the glass and titanium facades of modern high-rise buildings is a "3D" job: dangerous, dirty, and difficult. While several wall-cleaning robots exist, most are designed for flat, uniform surfaces. The National Grand Theatre in Beijing, an elliptic half-shell composed of over 20,000 planks set at varying angles across 54 distinct strips, presents a unique challenge.
Existing robots lacked the dexterity to navigate these curves or couldn't safely transition between structural strips. This research introduces a specialized autonomous climbing robot designed to navigate this complex geometry while maintaining high cleaning efficiency and structural safety.
Mechanical Design and Distributed Control
The robot utilizes a dual-frame structure (main and auxiliary) capable of relative sliding. To protect the delicate glass and titanium planks, the robot does not grip the building directly. Instead, it interacts with "sliding-rods" mounted on tracks integrated into the theatre's structure.
Key mechanical features include:
- Abdominal Plate: An interlocking mechanism with wheels on the sliding-rods that acts as a fail-safe against "capsizing" (peeling off the curved wall).
- Indirect Gripping: Using sliding-rods avoids structural fatigue or damage caused by direct track grasping.
- Safety-Size Equilibrium: By using aluminum and polyethylene sliding-mates, the 3-meter robot maintains a manageable weight of 100kg, balancing functionality with the building's load-bearing limits.
To manage the robot's 25 joints (17 active), a distributed control system based on the Controller Area Network (CAN) bus was implemented. This architecture distributes the workload among five specialized nodes and a remote controller, ensuring high reliability and system extensibility.
Dynamic Modeling and Path Planning
The researchers employed Lagrange equations to model the robot’s climbing dynamics. This modeling allowed for precise calculation of force distribution, ensuring the robot neither tips away from the wall nor exceeds the 1500N bearing limit of the building's planks.
For navigation, a Complete Coverage Path Planning (CCPP) strategy was used. The building's surface was subdivided into six regions based on material and structural boundaries, allowing the robot to optimize horizontal cleaning sweeps and achieve over 95% coverage.
Key Results and Performance
The robot significantly outperformed previous "Sky Cleaner" models in on-site testing:
- Efficiency: Achieved a cleaning rate of 150 m²/hour, a fourfold increase over first-generation models.
- Speed: Demonstrated vertical climbing at 200 mm/s and sideways cleaning at 100 mm/s.
- Safety: Dynamic analysis confirmed that the maximum supporting forces (approx. 88.6 kg) remained well within the building's safety margins.
Conclusion
The successful deployment at the National Grand Theatre confirms the effectiveness of the sliding-rod interface and distributed CAN bus control. While highly specialized for this specific infrastructure, the robot provides a blueprint for future modular systems capable of maintaining complex, non-planar architectural surfaces.
Summary updated