Coverage Planning for Robotic Tooth Preparation in Densely Constrained Environments

Yunwen Li; Chen Chen; Xiangjie Yan; Chang Shu; Jianxia Hou; Shiji Song; Xiang Li

cs.ROAug 25, 2026

Coverage Planning for Robotic Tooth Preparation in Densely Constrained Environments

Authors: Yunwen Li, Chen Chen, Xiangjie Yan, Chang Shu, Jianxia Hou, Shiji Song, Xiang Li

Organizations: Department of Automation, Tsinghua University, China · Department of Mechanical and Process Engineering, ETH Zürich, Switzerland · Tsingscribe Medical Ltd., China · Department of Priority Oral Care, Peking University School and Hospital of Stomatology

Abstract

Tooth preparation refers to the controlled removal of tooth structure to create an optimal substrate for fixed restorations and is a core procedure in restorative dentistry. Automating this task is particularly challenging for robots because the dental bur must operate within a densely constrained intraoral workspace, where even sub-millimeter deviations can compromise outcomes or damage adjacent structures. This paper presents a novel robotic system for autonomous full-crown tooth preparation. The proposed framework includes: 1) an anatomy-aware toolpath planning algorithm that conforms precisely to a technician-designed preparation model while protecting adjacent teeth, and 2) a clearance-oriented end-effector yaw assignment strategy that allows intraoral access while reducing the risk of soft-tissue interference. Together, these features enable the robot to accurately mill the irregular tooth surface with an average geometric deviation of 0.117 mm (RMSE), achieving both restoration quality and clinical safety. A series of simulations and phantom-head experiments validate the system's feasibility and effectiveness.

Figures & tables

Fig. 1: Task and key challenges in tooth preparation. The procedure is performed in a confined intraoral workspace and requires complex surface shaping, posing challenges to geometric accuracy and collision avoidance.

Fig. 2: Overview of the proposed robotic tooth preparation framework. The system integrates (1) a technician-designed digital model as the preparation target, (2) anatomy-aware toolpath planning that respects bur geometry and adjacent-tooth safety, and (3) six-axis robotic execution with optimized handpiece insertion to ensure feasible intraoral access and precise material removal.

Fig. 3: The prototype of the tooth preparation robot.

Fig. 4: Occluso-axial separation. (a) Candidate points for each radial direction in the dynamic programming algorithm. (b) Finishing line and occluso-axial separation.

Fig. 5: Local tooth frames for toolpath generation. (a) Criteria for $z$ -axis optimization. (b) Sample local frames in oral scan frame.

Fig. 6: Toolpath generation for incisors, canines, and the axial surfaces of (pre)molar teeth. (a) Geometric model for calculating facet contact in each pushing direction. The candidate facets are offset by $r$ in their normal direction projected onto the $XY$ plane. (b) Cutter location result for all layers and pushing directions. The finishing line is inwardly offset by the cutter radius $r$ and added as the first layer, providing both the initial shoulder shape and a safety boundary for adjacent-teeth protection.

Fig. 7: (a) Toolpath for the occlusal surfaces of (pre)molar teeth. (b) Pre-finishing toolpath for (pre)molar teeth.

Fig. 8: Definition of the handpiece orientation for implicit obstacle avoidance. By aligning the handpiece $y$ -axis (perpendicular to the bur axis and maximally aligned with the physical longitudinal direction of the handpiece body) with the outward-pointing $y$ -axis of the tooth frame, the five-DOF toolpath is extended to a six-DOF trajectory that maintains a safe orientation relative to the buccal mucosa.

Fig. 9: Complete toolpaths generated by the proposed planning algorithm for (a, b) incisor, (c, d) canine, and (e, f) molar preparation. The upper row shows the global toolpath distribution, while the lower row shows the feedrate profiles and the corresponding preparation time. The relatively long preparation time results from the conservative, safety-oriented feedrate selection.

Fig. 10: Simulated tooth preparation process on a molar (a–d), a canine (e–g), and an incisor (h–j).

Fig. 11: Original and prepared tooth on a phantom head.

Fig. 12: Surface deviation of the prepared tooth relative to the designed target. (a,b) Alignment based on other unprepared teeth. (c,d) Best-fit rigid alignment between the prepared and designed tooth.

Metric	In-situ frame ( $\mathrm{mm}$ )	Best-fit frame ( $\mathrm{mm}$ )
Minimum signed error	-0.511	-0.423
Maximum signed error	0.757	0.245
Root mean square error	0.321	0.117
90% absolute deviation	0.561	0.210

TABLE I: Surface deviation metrics

Fig. 13: Anatomy-aware toolpath and clearance-oriented handpiece orientation. (a) Adjacent teeth were protected by the enforced toolpath boundary. (b) The proposed yaw selection strategy enables safe insertion. (c,d) Without the yaw assignment, the handpiece may interfere with the soft tissues and optical markers.

Robotics

Robotics studies the design and control of machines that perceive, move, and act in physical environments.

See all 751 papers

RobOralScan: Learning Active Intraoral Scanning for Robotic Dental Reconstruction
Jun 25, 2026Jinhyung Lee, Haeun Yun, Siwon Kim +4RL for Robotics Robot Policy Learning
DGIST
Haptic bilateral teleoperation system for free-hand dental procedures
Mar 27, 2025Lorenzo Pagliara, Vincenzo Petrone, Enrico Ferrentino +2Healthcare Robotics Robot Teleoperation
Department of Information Engineering, Electrical Engineering and Applied Mathematics (DIEM), University of Salerno, 84084 Fisciano, Italy · Specialization School in Oral Surgery, University of Naples Federico II, 80138 Napoli, Italy
3D Geometric Tooth Alignment Planning via Deep Reinforcement Learning
Jul 16, 2026Yong Li, Jianwen Lou, Jiayue Ma +3Reinforcement Learning Healthcare
School of Software Technology, Zhejiang University, Hangzhou, China · State Key Laboratory of CAD&CG, Zhejiang University, Hangzhou, China · Stomatology Hospital, Zhejiang University School of Medicine, Hangzhou, China

Jun 25, 2026J/K move · Enter open · S save

RobOralScan: Learning Active Intraoral Scanning for Robotic Dental Reconstruction

Jinhyung Lee, Haeun Yun, Siwon Kim, Gihyun Baek+3

DGIST

RL for Robotics Robot Policy Learning 3D Reconstruction Dental Imaging

Open paper AskPDF arXiv

Intraoral scanning is widely used for digital optical impressions in prosthodontic, implant, and orthodontic treatment, but full-arch and long-span scanning remain labor-intensive tasks with limited automation. In the confined oral cavity, operators must continuously adjust scanner motion while accumulating narrow field-of-view observations, making reconstruction quality sensitive to missing tooth surfaces and operator workload. We propose RobOralScan, which, to the best of our knowledge, is the first reinforcement learning (RL)-based pipeline for robotic automatic intraoral scanning. RobOralScan introduces a geometric memory-based observation space that accumulates partial scan observations into a tri-state geometric representation, allowing the policy to reason over scan history and insufficiently observed regions. It further introduces tooth-wise coverage learning, combining coverage-aware reward signals and a progressive training scheme to improve global reconstruction coverage while reducing uneven coverage across individual teeth. The learned policy selects relative scanner motions from accumulated geometric memory and robot proprioception for closed-loop scan control within the oral workspace. RobOralScan achieves a Chamfer Distance of 0.00838, an average coverage of 92.58%, a lower-tail per-tooth coverage of 88.45%, and a normalized AUC of 0.6674, completing the scan criterion in 8 of 10 evaluation episodes. Furthermore, zero-shot sim-to-real experiments demonstrate its practical feasibility on a physical robot-scanner setup.

Abstract

Robotics

Explore similar work

RobOralScan: Learning Active Intraoral Scanning for Robotic Dental Reconstruction

Haptic bilateral teleoperation system for free-hand dental procedures

3D Geometric Tooth Alignment Planning via Deep Reinforcement Learning