Facade Inspection: Design, Prototyping, and Testing of a Hybrid Cable-Driven Parallel Robot

In the field of architecture, early detection of damage or degradation of building facades has become increasingly vital due to the need for continuous monitoring of structural integrity. Traditional methods, such as visual inspections, are being supplemented by technological advancements, especially in robotics, which offer innovative solutions for more efficient and precise inspections. This work focuses on the development of a five degree of freedom hybrid cable-driven parallel robot designed for vertical facade inspections. A detailed robot’s design and CAD modeling, with a particular focus on a torque transmission mechanism that simplifies the motion of two cables using a single motor is presented. Two degrees of freedom are driven by cables, while the remaining three are driven by a Sarrus-type mechanism and a pan-tilt mechanism. The inverse kinematics models are also developed. A prototype is presented, involving additive manufacturing. A control system for tracking a zig-zag trajectory, commonly used in inspection tasks, was experimentally validated.

💡 Research Summary

The paper presents the design, prototyping, and experimental validation of a five‑degree‑of‑freedom (5‑DOF) hybrid cable‑driven parallel robot (HCDPR) intended for vertical façade inspection. Recognizing the limitations of traditional visual inspections—subjectivity, high labor costs, and safety risks—the authors propose a robotic solution that leverages the low inertia, large workspace, and lightweight nature of cable‑driven mechanisms.

The robot’s architecture combines two cable‑driven translational DOFs with three additional DOFs realized by a Sarrus‑type linear actuator (providing motion along the z‑axis) and a pan‑tilt unit for orientation control. A key innovation is the dual‑cable transmission system: a single motor drives two cables via a crossed‑belt (flat belt) arrangement. The crossed belt forces the driven pulley to rotate in the opposite direction of the drive pulley, thereby synchronizing the two cables while halving the number of actuators required. The belt contact angle is derived analytically from pulley spacing and diameter, yielding an efficiency of approximately 98 % and reduced vibration and noise compared with gear‑based drives.

The mobile platform, which carries the end‑effector, is constructed from acrylic plates reinforced with 14 mm diameter aluminum bars. Finite‑element analysis in SolidWorks predicts a maximum deformation of 0.57 mm under the expected payload, confirming sufficient stiffness for inspection tasks. The platform also integrates an Arduino Uno and a sensor mounting bracket, enabling on‑board data acquisition and closed‑loop control.

Inverse kinematics are derived by expressing the positions of the four drive pulleys and the eight cable attachment points relative to a local reference frame. Cable lengths are computed from Euclidean distances between pulley centers and attachment points, and unit vectors are obtained to describe cable direction. These geometric relationships are then linked to motor rotation angles through a linear mapping, providing the necessary command signals for trajectory tracking.

The prototype is fabricated using a V‑Slot aluminum extrusion frame (20 × 20 mm profile, overall dimensions 1.04 m × 1.04 m × 0.34 m) and Dynamixel MX‑106T servomotors. Structural components such as pulleys, spacers, and the anti‑slip belt crossing mechanism are 3‑D printed in PLA. Braided nylon cables (0.45 mm diameter) with a 50 kg load rating are employed. The anti‑slip system incorporates flat bearings and 6 mm Seeger rings to ensure smooth belt motion with minimal friction.

Experimental validation uses an OptiTrack optical motion‑capture system together with a Time‑of‑Flight (ToF) distance sensor to record the robot’s motion while it follows a zig‑zag inspection path—a common pattern for façade scanning. The recorded trajectory closely matches the planned path, and motor position data align with the predicted cable length variations. Minor deviations are attributed to transient cable tension fluctuations, indicating a need for more sophisticated tension‑balancing (tensegrity‑based) control strategies.

The authors conclude that the HCDPR successfully demonstrates a viable, low‑cost, and safe alternative to manual façade inspection. Its modular cable‑driven architecture offers scalability, while the dual‑cable transmission reduces weight and energy consumption. Future work will focus on improving tension consistency, reducing platform vibration, and integrating additional non‑destructive evaluation sensors (e.g., ultrasonic, thermographic) to create a fully autonomous inspection system capable of delivering quantitative structural health data.