Improving dependability in robotized bolting operations

Lorenzo Pagliara; Vincenzo Petrone; Violeta Redondo; Enrico Ferrentino; Manuel Ferre; Pasquale Chiacchio

cs.RONov 13, 2025

Improving dependability in robotized bolting operations

Authors: Lorenzo Pagliara, Vincenzo Petrone, Violeta Redondo, Enrico Ferrentino, Manuel Ferre, Pasquale Chiacchio

Organizations: Department of Information Engineering, Electrical Engineering and Applied Mathematics (DIEM), University of Salerno, Via Giovanni Paolo II, 132, Fisciano, 84084, Italy · GTD Science, Infrastructures & Robotics, Av. Leonardo Da Vinci, 2, Madrid, 28906, Spain · Universidad Politécnica de Madrid, Centre for Automation and Robotics (CAR) UPM-CSIC, José Gutiérrez Abascal, 2, Madrid, 28006, Spain

Abstract

Bolting operations are critical in industrial assembly and in the maintenance of scientific facilities, requiring high precision and robustness to faults. Although robotic solutions have the potential to improve operational safety and effectiveness, current systems still lack reliable autonomy and fault management capabilities. To address this gap, we propose a control framework for dependable robotized bolting tasks and instantiate it on a specific robotic system. The system features a control architecture ensuring accurate driving torque control and active compliance throughout the entire operation, enabling safe interaction even under fault conditions. By designing a multimodal human-robot interface (HRI) providing real-time visualization of relevant system information and supporting seamless transitions between automatic and manual control, we improve operator situation awareness and fault detection capabilities. A high-level supervisor (SV) coordinates the execution and manages transitions between control modes, ensuring consistency with the supervisory control (SVC) paradigm, while preserving the human operator's authority. The system is validated in a representative bolting operation involving pipe flange joining, under several fault conditions. The results demonstrate improved fault detection capabilities, enhanced operator situational awareness, and accurate and compliant execution of the bolting operation.

Figures & tables

Panel of Figure 1 — Figure 1: Example of coupling and tightening steps in a bolting operation.

Figure 2: Control system for dependable bolting operations.

Figure 3: Screenshot of the system’s GUI during operation.

Figure 4: Diagram of a generic telerobotic supervisory control operation.

Panel of Figure 5 — Figure 5: Experimental setup for the bolting operation: (a) overall remote site with the UR5e robot and the pipe assembly; (b) detailed view of the chain clamp and the captive nut arrangement.

Figure 6: Comparison between the mean of the interaction force in Scenario A and Scenario B.

Figure 7: Comparison between the planned coupling trajectory and the one actually followed by the FR after human intervention.

Figure 8: Forces rendered on the HD.

Figure 9: Tightening torque profile.

Condition	Nominal cost	Contingency cost	Operator occupancy	Source
Manual	$\approx$ $339623\text{\,}\mathrm{s}$ per 5-fastener sequence, including reading instructions/docs	—	Continuous, on-site	[ 26 ]
Fully teleoperated	$163240\text{\,}\mathrm{s}$ per bolt	$12635\text{\,}\mathrm{s}$ per event; $\approx$ 30% on overall task completion upon contingencies	Continuous, at controls	[ 10 ]
Autonomous	$\approx$ $5\text{\,}\mathrm{s}$ /operation, including $\approx$ $2.5\text{\,}\mathrm{s}$ perception	4/5 classes escalate w/o recovery path	0% nominal; undefined on failure, likely on-site	[ 43 , 25 ]
Proposed	Same as autonomous	Same as teleoperated, contingent fraction only	$\mathbb{P}(\text{contingency})$ $\times$ occupancy/event	ours

Table 1: Qualitative comparison of execution conditions.

Failure mode	Cause	How it becomes observable	Mitigation in current system	Covered?
Excessive interaction force	Operator commands a pose that would cause high contact force	Force/torque sensor reading; admittance controller limits force	Safety stop triggers when force exceeds threshold	Yes (Section 3.2 )
Self-collision	Operator unaware of robot’s joint configuration or external geometry	Visual feedback; GUI state display	Operator monitors via GUI; safety stop if collision occurs	Yes (one failure occurred in Section 3.4 )
Joint-limit violation	Operator commands a pose outside the robot’s kinematic envelope	GUI joint configuration display; robot stops	Operator monitors via GUI; controller stops	Yes
Inability to complete tightening in a single step	Joint limits and self-collision risk prevent full rotation	Operator observes partial completion	Procedure split into alternating coupling/tightening steps	Yes (Section 3.4 )
Communication delay	Network latency; visual feedback lag	Delayed response to commands	PREEMPT_RT patch [ 64 ] ; visual feedback lag is accepted as a known limitation	Partially (patched OS as in Section 3.1 )

Table 2: Teleoperation failure modes

Explore similar work

Learning-augmented robotic automation for real-world manufacturing
Apr 24, 2026Yunho Kim, Quan Nguyen, Taewhan Kim +2Robotic Control Deformable Object Manipulation
All authors are with Neuromeka Co., Ltd.
PanelShield: Verifiable Closed-Loop Safe Planning for Robotic Industrial Panel Operation
Aug 28, 2026Guipeng Xin, Jiahe Xu, Chenhui Wan +3Robot Task Planning Robot Safety
School of Mechanical Science and Engineering Huazhong University of Science and Technology Wuhan, China
Robust Operational Space Control with Conformal Disturbance Bounds for Safe Redundant Manipulation
Jul 1, 2026Wenhua Liu, Fan Zhang, Qin LinRobotic Control Control Barrier Functions
Department of Engineering Technology and the Department of Electrical and Computer Engineering, University of Houston, Houston, TX 77004, USA.

Apr 24, 2026J/K move · Enter open · S save

Learning-augmented robotic automation for real-world manufacturing

Yunho Kim, Quan Nguyen, Taewhan Kim, Youngjin Heo+1

All authors are with Neuromeka Co., Ltd.

Robotic Control Deformable Object Manipulation Robotics Robotic Manipulation Industrial Robotics

Open paper AskPDF arXiv

Industrial robots are widely used in manufacturing, yet most manipulation still depends on fixed waypoint scripts that are brittle to environmental changes. Learning-based control offers a more adaptive alternative, but it remains unclear whether such methods, still mostly confined to laboratory demonstrations, can sustain hours of reliable operation, deliver consistent quality, and behave safely around people on a live production line. Here we present Learning-Augmented Robotic Automation, a hybrid system that integrates learned task controllers and a neural 3D safety monitor into conventional industrial workflows. We deployed the system on an electric-motor production line to automate deformable cable insertion and soldering under real manufacturing constraints, a step previously performed manually by human workers. With less than 20 min of real-world data per task, the system operated continuously for 5 h 10 min, producing 108 motors without physical fencing and achieving a 99.4% pass rate on product-level quality-control tests. It maintained near-human takt time while reducing variability in solder-joint quality and cycle time. These results establish a practical pathway for extending industrial automation with learning-based methods.

Abstract

Explore similar work

Learning-augmented robotic automation for real-world manufacturing

PanelShield: Verifiable Closed-Loop Safe Planning for Robotic Industrial Panel Operation

Robust Operational Space Control with Conformal Disturbance Bounds for Safe Redundant Manipulation