# DecompGrind: A Decomposition Framework for Robotic Grinding via Cutting-Surface Planning and Contact-Force Adaptation

Source: [DecompGrind: A Decomposition Framework for Robotic Grinding via Cutting-Surface Planning and Contact-Force Adaptation](<https://www.sota2.com/research/paper/69c36c3c0636e5bf4afc6637>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-01T05:00:12.613Z

Authors: Shunsuke Araki, Takumi Hachimine, Yuki Saito, Kouhei Ohnishi, Jun Morimoto, Takamitsu Matsubara

Published: 2026-03-24T06:52:09.000Z

## Abstract

Robotic grinding is widely used for shaping workpieces in manufacturing, but it remains difficult to automate this process efficiently. In particular, efficiently grinding workpieces of different shapes and material hardness is challenging because removal resistance varies with local contact conditions. Moreover, it is difficult to achieve accurate estimation of removal resistance and analytical modeling of shape transition, and learning-based approaches often require large amounts of training data to cover diverse processing conditions. To address these challenges, we decompose robotic grinding into two components: removal-shape planning and contact-force adaptation. Based on this formulation, we propose DecompGrind, a framework that combines Global Cutting-Surface Planning (GCSP) and Local Contact-Force Adaptation (LCFA). GCSP determines removal shapes through geometric analysis of the current and target shapes without learning, while LCFA learns a contact-force adaptation policy using bilateral control-based imitation learning during the grinding of each removal shape. This decomposition restricts learning to local contact-force adaptation, allowing the policy to be learned from a small number of demonstrations, while handling global shape transition geometrically. Experiments using a robotic grinding system and 3D-printed workpieces demonstrate efficient robotic grinding of workpieces having different shapes and material hardness while maintaining safe levels of contact force.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Robotic Grinding on WP-E1](<https://www.sota2.com/research/sota/robotic-grinding-on-wp-e1.md>): Robotic Grinding; WP-E1; Execution Time (s): 182.6.

- [Grinding on WP-E2](<https://www.sota2.com/research/sota/grinding-on-wp-e2.md>): Robotic Grinding; WP-E2; Execution Time (s): 192.3.

- [Robotic Grinding on WP-E1 1.0 (test)](<https://www.sota2.com/research/sota/robotic-grinding-on-wp-e1-1-0-test.md>): Robotic Grinding; WP-E1 1.0 (test); Execution Time (min): 8.

- [Robotic Grinding on WP-E2 1.0 (test)](<https://www.sota2.com/research/sota/robotic-grinding-on-wp-e2-1-0-test.md>): Robotic Grinding; WP-E2 1.0 (test); Execution Time (min): 8.

- [Robotic Grinding on WP-E3 1.0 (test)](<https://www.sota2.com/research/sota/robotic-grinding-on-wp-e3-1-0-test.md>): Robotic Grinding; WP-E3 1.0 (test); Execution Time (min): 8.

- [Grinding on WP-S1](<https://www.sota2.com/research/sota/grinding-on-wp-s1.md>): Grinding; WP-S1; Execution Time (s): 5.3.

- [Grinding on WP S2](<https://www.sota2.com/research/sota/grinding-on-wp-s2.md>): Grinding; WP S2; In-limit Force Ratio: 100.

- [Grinding on WP-S3](<https://www.sota2.com/research/sota/grinding-on-wp-s3.md>): Grinding; WP-S3; In-limit Force Ratio: 100.

- [Grinding on WP-S4](<https://www.sota2.com/research/sota/grinding-on-wp-s4.md>): Grinding; WP-S4; In-limit Force Ratio (%): 100.

- [Grinding on WP S5](<https://www.sota2.com/research/sota/grinding-on-wp-s5.md>): Grinding; WP S5; In-limit Force Ratio (%): 100.

- [Robotic Grinding on WP-E3](<https://www.sota2.com/research/sota/robotic-grinding-on-wp-e3.md>): Robotic Grinding; WP-E3; Execution Time (s): 180.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.22859>)

[PDF](<https://arxiv.org/pdf/2603.22859.pdf>)
