研究目的
To develop a camera-based path planning framework that reduces the time and cost for programming industrial robots in small and medium-sized enterprises (SMEs) focusing on low volume and high variant manufacturing, by enabling fast preparation and execution of robot tasks in dynamic environments.
研究成果
The framework effectively reduces the time and cost for programming industrial robots in SMEs by enabling fast, collision-free path planning independent of workpiece position. The case study demonstrates its feasibility and performance, with total execution time of 183.88 seconds. Future work should focus on extending the solution space to use both position and orientation in path planning, automated task generation, and ensuring process stability.
研究不足
The path planning algorithm currently uses a fixed orientation for path planning, which may limit flexibility. Automated task generation based on differences between CAD model and scanned data is not fully implemented, and process stability requirements need further assurance. Computing times, though reduced, could be optimized further for real-time applications.
1：Experimental Design and Method Selection:
The framework uses a pipeline split into offline and online parts. Offline part involves path planning in CAD models and feature extraction using Clustered Viewpoint Feature Histogram (CVFH). Online part uses a laser scanner for 3D scanning, preprocessing with filters (e.g., statistical outlier removal, Moving Least Squares), feature extraction, matching with K nearest neighbor and Interactive Closest Point (ICP) algorithms, and path planning with adaptive subspace division and collision detection using a three-layer approach (Octomap, bounding box, GJK algorithm). Trajectory optimization is done with Bernstein-Bezier splines, and execution via TCP-IP interface.
2：Sample Selection and Data Sources:
A case study with an IRB1600 ID welding robot and a box workpiece on a euro pallet. Data sources include CAD models and laser scan data from the sensor.
3：List of Experimental Equipment and Materials:
Industrial robot (IRB1600 ID), line laser scanner, CAD software, computer for processing.
4：Experimental Procedures and Operational Workflow:
Place workpiece, scan environment with laser scanner, preprocess data, match with CAD model, plan collision-free path using inverse kinematics and path planning algorithms (e.g., A*), optimize trajectory, and execute via robot control.
5：Data Analysis Methods:
Execution times are benchmarked for various functions (e.g., Octomap generation, segmentation, matching, exploration, inverse kinematics, collision detection).
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
