Development of a 3 in Sewer Pipe Inspection Robot with an Articulated Differential Mechanism using X-shaped Linkages

Technical Summary: Development of a 3-in Sewer Pipe Inspection Robot with an Articulated Differential Mechanism

Problem Statement
The aging of sewer pipelines in Japan has created an urgent need for inspection, particularly for narrow inner-diameter pipes (100 mm or less) where over 39,200 km of infrastructure is older than 50 years. Existing solutions face significant limitations: push-in cameras struggle with high cable bending stiffness in multiple bends, while pneumatic peristaltic robots are too slow and require bulky compressors. Furthermore, existing self-propelled robots, such as those using planetary gear structures, pose evacuation risks during emergencies due to their complex, constantly pressing designs. The authors' previous iteration, Xbot-1, addressed emergency evacuation via a wire-driven mechanism but suffered from insufficient traction force over long distances and a tendency to get stuck when traversing pipe joints (steps).

Methodology and Mechanical Design
The paper proposes "Xbot-2," an upgraded 3-inch sewer pipe inspection robot featuring an articulated differential mechanism. The robot consists of a front camera/LED, a wire-driven unit, and multiple propulsion units connected by an outer casing.

• Articulated Differential Mechanism: The core innovation involves connecting multiple propulsion units, each containing four drive wheels mounted on X-shaped linkages. An aramid wire runs through the casing; winding it shortens the linkage axis, expanding the robot's diameter to press wheels against the pipe wall. By connecting three units, the system functions as a differential mechanism. In pipes with varying diameters, the wire tension causes the linkage at the larger diameter to deform less than the one at the smaller diameter, allowing the robot to adapt its posture to the pipe geometry using a single wire.
• Obstacle Negotiation Strategy: To address the issue of getting stuck at pipe joint steps, the authors devised a control method based on motor current monitoring rather than additional sensors (like encoders or IMUs) to minimize size and complexity. The algorithm detects a step when the drive wheel motor current exceeds a threshold (2.5 A), indicating the wheel has stalled against a step. Upon detection, the system reverses the winding reel by 150 degrees to slacken the wire, allows the X-shaped linkage to shrink via contact force from the step, moves backward briefly, and then proceeds forward to cross the obstacle.

Key Contributions

1. Mechanical Improvement: The transition from a single propulsion unit (Xbot-1) to a connected multi-unit system (Xbot-2) significantly enhances traction force.
2. Control Algorithm: The development of a sensor-less step-detection and negotiation algorithm that utilizes drive wheel motor current waveforms to autonomously manage wire tension.
3. Experimental Verification: Validation of the proposed control method through simulated pipe experiments and actual robot trials involving pipe joints.

Results

• Traction Force: Comparative measurements showed that Xbot-2 with two propulsion units achieved approximately 1.5 times the traction force of Xbot-1 (75.0 N vs. 50.6 N). With three propulsion units, the traction force increased to approximately 2.0 times that of the original (103.1 N).
• Step Negotiation: In preliminary tests with two units, the robot stalled at pipe joints, with motor current rising from a baseline of 1.2–2.0 A to a sustained 2.8–2.9 A.
• Autonomous Traversal: In trials with three propulsion units and the proposed control algorithm, the robot successfully detected the step (current rising to ~2.9 A at 3.8 seconds), executed the wire-slackening sequence (reel reversal of 150 degrees), and smoothly traversed the pipe joint without getting stuck.

Significance and Claims
The paper claims that Xbot-2 successfully improves upon the limitations of its predecessor by significantly increasing traction force through the connection of propulsion units and enabling the traversal of pipe joints through an autonomous, current-based control method. The authors state that the proposed control method "significantly improved the step-over capability" of the articulated robots. While the experimental threshold for step detection was verified to work in principle, the authors modestly note that a more general step detection method is required for practical application. Future work is planned to include long-distance inspections in actual piping and further verification of emergency evacuation performance.