Work

Model-Based Posture Estimation With Structured Light System

Year: 2025

Type: article

Abstract: Mechanical inaccuracies and assembly tolerances are inevitable in a wide range of engineered systems, often leading to errors in relative position and orientation, referred to as “posture.” These posture errors can reduce the control precision and overall system performance. To mitigate this problem, a model-based method was developed for estimating the six-degree-of-freedom posture of a structured light module (SLM) that integrates projection and sensing within a compact structure. In this approach, three laser devices emit rays,, and the resulting points are captured by an imaging device. The spatial configuration of these points in a global coordinate system is utilized to determine the posture of the SLM. In the first stage, a Jacobian-based iterative estimation algorithm updates the posture by minimizing discrepancies between the laser points predicted from the initial estimate and those obtained through measurements. In the second stage, a misalignment parameter characterizing the geometric deviations among the laser axes is introduced and refined using nonlinear optimization. This two-stage process makes robust convergence possible and sustains estimation accuracy, even when laser axis misalignments are present. The method was validated through simulations and physical experiments on a custom-designed testbed under nominal and perturbed conditions. Simulation results demonstrated stable estimation performance despite laser axis deviations and testbed imperfections. In contrast, physical experiments showed increased estimation errors, primarily owing to subpixel inaccuracies in laser point localization by the imaging system. The method can be applied to manipulators, industrial automation systems, and other engineered platforms that support posture estimation in dynamic or sensor-constrained environments.

Source: IEEE Access

DOI: 10.1109/access.2025.3593270

Authors: Jeongyeop Lee, Bumjoo Lee

Institution: Myongji University

Language: English

Topic: Optical measurement and interference techniques

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $2,075

Funder: National Research Foundation of Korea