Soft Computing Optimization of Industrial Robot Inverse Kinematics for Intelligent Manufacturing: A Statistical Benchmark of Metaheuristic Algorithms
5 affiliations
- 1Departamento de Estudios en Ingeniería para la Innovación, Universidad Iberoamericana Ciudad de México, Ciudad de México 01219, Mexico
- 2Área de Industria, Materiales y Energía, Universidad EAFIT, Medellín 050022, Colombia
- 3Facultad de Ingenierías, Institución Universitaria ITM, Medellín 050036, Colombia
- 4Facultad de Ingeniería Eléctrica y Electrónica, Universidad Veracruzana, Boca del Río 94294, Mexico
- 5ADC—IT Innovation Architecture, Bancolombia, Medellín 050010, Colombia
Abstract
Accurate and computationally efficient inverse kinematics is a fundamental requirement for industrial manipulators in intelligent manufacturing, where positioning, assembly, and handling demand reliable pose computation under nonlinear constraints. However, the inverse mapping of six-degree-of-freedom manipulators may admit multiple solutions, become ill-conditioned near singularities, or violate joint limits, while metaheuristic comparisons rely on heterogeneous protocols emphasizing best-run results. Herein, a statistical benchmark of five metaheuristics is presented for solving the inverse kinematics of a UR3 manipulator: Covariance Matrix Adaptation Evolution Strategy (CMA-ES), Multi-Verse Optimizer (MVO), Non-Dominated Sorting Genetic Algorithm II (NSGA-II), Salp Swarm Algorithm (SSA), and Whale Optimization Algorithm (WOA). A master–slave architecture separates the stochastic search from a deterministic evaluator that computes an unweighted full-pose residual subject to joint bounds and a Jacobian-determinant screening penalty. Three poses represent reoriented, nominal-height, and near-singular upper-workspace conditions, each assessed over 1000 independent runs with fixed algorithm settings. Nonparametric tests are interpreted with Cliff’s Δ effect sizes. The results reveal distinct accuracy–time trade-offs: CMA-ES obtains the smallest best-run objective at modest runtimes, MVO yields the smallest mean objective but the longest runtimes, and WOA is the fastest with higher residuals. The findings support algorithm selection for offline inverse-kinematics computation and supervisory robotic planning.
Related
- Energy-Efficient UAV Covert Communications via Metaheuristic Optimization
- Optimization Strategy of PID-Inspired Metaheuristic Algorithm-Tuned Fractional-Order Controllers Applied to Nonlinear AVR System
- Discrete Phase-Shift Optimization of Reconfigurable Intelligent Surfaces using Uniform Reference Directions
- Optimization method for flexible scheduling of digital twin industrial loads based on load side potential elasticity
- Intelligent Reflecting Surface Deployment for Low-Altitude Coverage: Illumination Geometry, Directional Characteristics, and Optimization
BibTeX
@article{caballeromora2026soft,
title = {{Soft Computing Optimization of Industrial Robot Inverse Kinematics for Intelligent Manufacturing: A Statistical Benchmark of Metaheuristic Algorithms}},
author = {Julio Antonio Caballero-Mora and Daniel Sanin-Villa and Victoria Bello-Hernández and Vanessa Botero-Gómez and Rogelio de Jesús Portillo-Vélez and Juan C. Tejada},
journal = {Mathematics},
year = {2026},
month = oct,
doi = {10.3390/math14203719},
url = {https://www.mdpi.com/2227-7390/14/20/3719},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Soft Computing Optimization of Industrial Robot Inverse Kinematics for Intelligent Manufacturing: A Statistical Benchmark of Metaheuristic Algorithms
AU - Julio Antonio Caballero-Mora
AU - Daniel Sanin-Villa
AU - Victoria Bello-Hernández
AU - Vanessa Botero-Gómez
AU - Rogelio de Jesús Portillo-Vélez
AU - Juan C. Tejada
PY - 2026
DA - 2026/10/13
JO - Mathematics
DO - 10.3390/math14203719
UR - https://www.mdpi.com/2227-7390/14/20/3719
AB - Accurate and computationally efficient inverse kinematics is a fundamental requirement for industrial manipulators in intelligent manufacturing, where positioning, assembly, and handling demand reliable pose computation under nonlinear constraints. However, the inverse mapping of six-degree-of-freedom manipulators may admit multiple solutions, become ill-conditioned near singularities, or violate joint limits, while metaheuristic comparisons rely on heterogeneous protocols emphasizing best-run results. Herein, a statistical benchmark of five metaheuristics is presented for solving the inverse kinematics of a UR3 manipulator: Covariance Matrix Adaptation Evolution Strategy (CMA-ES), Multi-Verse Optimizer (MVO), Non-Dominated Sorting Genetic Algorithm II (NSGA-II), Salp Swarm Algorithm (SSA), and Whale Optimization Algorithm (WOA). A master–slave architecture separates the stochastic search from a deterministic evaluator that computes an unweighted full-pose residual subject to joint bounds and a Jacobian-determinant screening penalty. Three poses represent reoriented, nominal-height, and near-singular upper-workspace conditions, each assessed over 1000 independent runs with fixed algorithm settings. Nonparametric tests are interpreted with Cliff’s Δ effect sizes. The results reveal distinct accuracy–time trade-offs: CMA-ES obtains the smallest best-run objective at modest runtimes, MVO yields the smallest mean objective but the longest runtimes, and WOA is the fastest with higher residuals. The findings support algorithm selection for offline inverse-kinematics computation and supervisory robotic planning.
SN - 2227-7390
ER -