cnc machining

This study addresses the challenge of coupling discrete operation planning with continuous toolpath generation in CNC machining of B-rep models by proposing the CNCGEN framework. This framework introduces a novel "persistent manufacturing object" modeling mechanism that, combined with a learned agent verifier providing material removal feedback, dynamically correlates local predictions with geometric evolution to enable stepwise generation and state updating of operations and toolpaths. The approach integrates deep learning for three-axis machining, B-rep representations, and parametric toolpath algorithms, supported by a synthetically generated dataset incorporating geometric verification. Experimental results demonstrate that, compared to baseline methods, the proposed framework significantly improves workpiece geometric accuracy while effectively mitigating residual material and overcutting phenomena.

Conventional CNC woodworking for middle school students faces high entry barriers due to reliance on complex CAD software and abstract digital modeling. Method: This paper introduces a direct hand-drawing programming paradigm grounded in physical materials—users sketch lines, symbols, and color annotations directly onto wood surfaces; the system employs real-time computer vision for markerless pose estimation, spatial registration between physical and virtual spaces, and semantic parsing to automatically generate toolpaths. Contribution/Results: We propose the first “Physical Material Annotation Language” (PMAL), integrated with real-time interactive preview, an open-source customizable platform, and a lightweight toolpath mapping engine. A user study demonstrates that novices achieve high-precision custom wood part fabrication within an average of five minutes, with design task success rates increasing 3.2×. This approach substantially lowers the barrier to creative expression and advances the democratization of CNC manufacturing education.

Economical desktop GRBL-based CNC machines commonly lack rotational axes, hindering the machining of rotationally symmetric or polyhedral parts. Conventional solutions—such as hardware retrofits, controller replacements, or commercial CAM software—entail high cost and technical barriers. Method: This paper proposes a purely software-based indexing rotary post-processing framework that requires no hardware modification or firmware update. Leveraging a browser-based interactive interface, it automatically maps 2D toolpaths to discrete rotational indexing commands, enabling 4-axis indexed machining. Contribution/Results: The approach is fully compatible with standard GRBL controllers, substantially lowering both technical and economic barriers for multi-axis manufacturing in educational and maker environments. Experimental validation on off-the-shelf desktop CNC systems demonstrates efficient fabrication of rotationally symmetric and polyhedral components. This work establishes a practical, low-cost pathway toward accessible multi-axis CNC manufacturing.

Conventional robot trajectory planning methods—centered on tool motion—are ill-suited for machining scenarios involving moving workpieces and a fixed tool center point (TCP). Method: This paper proposes a workspace-coordinate-system-based trajectory planning method that takes the workpiece’s motion path as input. It employs B-spline parameterization to model the path while simultaneously incorporating inverse kinematics resolution and velocity-constrained optimization, ensuring adherence to arbitrary orientation constraints and TCP-end velocity requirements. The result is a continuous, smooth, and high-precision joint-space trajectory. Contribution/Results: Compared to traditional tool-centric paradigms, the method significantly enhances trajectory flexibility and geometric fidelity in complex freeform surface machining. Experimental validation was conducted on a real industrial robot platform, supporting diverse inputs—including mathematical functions and CAD-exported point clouds—demonstrating strong engineering applicability and robustness.

Manufacturing SMEs face critical bottlenecks in visual assembly quality control—including scarce real-image acquisition, high annotation costs, and insufficient training data. To address these challenges, this paper proposes a CAD model–driven fully synthetic data framework. It establishes an end-to-end virtual generation pipeline integrating parametric CAD modeling, physics-based rendering, and YOLO-family object detection, enabling efficient simulation-to-reality transfer learning. This work represents the first systematic deployment of a purely synthetic data approach to industrial inspection of planetary gear assemblies. Experiments demonstrate a 99.5% mAP@0.5:0.95 on synthetic data; after domain adaptation, detection accuracy remains at 93% on real-world images. The framework significantly reduces dependence on manual annotation and physical image collection, validating its feasibility for lightweight, reusable, and low-cost industrial deployment.

Traditional CNC machining safety verification relies heavily on repetitive simulation, which struggles to accommodate changing requirements and lacks a formal, scalable approach to collision detection. This work proposes a formal verification framework based on separation logic, modeling the physical workspace as a Spatial Heap where occupancy relations are treated as logical resources. By introducing a Parser-Prover Handshake mechanism, the framework decouples kinematic computation from logical reasoning and, for the first time, defines physical collisions as Spatial Data Races in a logical sense. Leveraging concurrent separation logic, it supports ownership-transfer-based verification in collaborative scenarios. The method eliminates the need for geometric simulation and provides a mathematically rigorous, scalable guarantee of collision freedom, establishing a formal safety foundation for autonomous manufacturing.

Existing CAD generation methods prioritize visual similarity over geometric precision and suffer from quantization-induced errors in parametric representations, rendering them inadequate for industrial applications demanding exact dimensional accuracy. To address this, this work proposes a Plan-Then-Construct paradigm: first generating a structured design plan with explicit continuous dimensional parameters, then employing a pointer mechanism during the construction phase to directly reference these parameters, thereby enabling high-fidelity modeling. This approach achieves the first end-to-end prediction of continuous parameters, eliminating quantization errors, and ensures dimensional consistency by decoupling parameter inference from geometric construction. We introduce the first large-scale CAD dataset annotated with design plans and propose a three-level geometric evaluation metric—vertices, edges, and faces. Experiments demonstrate that our method significantly outperforms existing approaches across all geometric levels, making it suitable for precision-sensitive engineering tasks.

This work addresses the limitation of existing CAD model evaluation methods, which predominantly emphasize visual fidelity while neglecting engineering functionality. To bridge this gap, the authors propose CADEngBench, a dual-track benchmark that systematically incorporates engineering behavior validation—including finite element analysis (FEA) alignment, design-for-manufacturing (DFM) checks, and kinematic joint dynamics—into the assessment framework, covering both parametric parts and assemblies. The benchmark employs techniques such as B-Rep validity verification, parameter perturbation tests, functional editing tasks, and linear static simulations using CalculiX to comprehensively evaluate the engineering-grade capabilities of generated and edited models. Experimental results reveal that while current multimodal models outperform in CAD editing over generation, they still struggle with complex edits, FEA consistency, and accurately reconstructing real-world assembly mating relationships.

This work addresses the challenge of achieving both generality and high fidelity in modeling instantaneous cutter-workpiece engagement (CWE) geometry for multi-axis milling under complex geometries and arbitrary toolpaths. To this end, the authors present the first open-source, reproducible high-precision CWE computation framework built upon the boundary representation (B-Rep) solid modeling kernel of Autodesk Fusion 360. The proposed method supports arbitrary tool types, workpiece geometries, and toolpaths without reliance on proprietary software, overcoming limitations of conventional discretization-based approaches in general scenarios. Through publicly accessible APIs, a cloud-ready architecture, and a shared experimental dataset, the study demonstrates the feasibility and accuracy of B-Rep–based CWE modeling, establishing an open benchmark for high-fidelity virtual machining and fostering reproducible research and collaboration in digital manufacturing.

为应对多样化生产带来的挑战，本文提出一种基于CAD模型的编程和执行系统，通过动态参数化的行为树控制结构实现人-机器人-起重机协同任务。