A multi-strategy enhanced secretary bird optimization algorithm for high-precision inverse kinematics in robotic arms

Xuhao Chen, Nan Lin, Fan Zhang, Xuhai ZhaoPLoS 2025 · DOI

TL;DR

This paper introduces an Improved Secretary Bird Optimization Algorithm (ISBOA) to solve the high-precision inverse kinematics (IK) required for the dangerous task of pyrotechnic grain assembly. By enhancing a bio-inspired algorithm with three distinct mathematical strategies, the researchers achieved significantly higher accuracy and better singularity avoidance than traditional analytical and numerical methods.

What Problem Does This Paper Address?

Assembling pyrotechnic products (explosives) is a high-risk manufacturing process that demands extreme precision due to tiny shell apertures. The "brain" of this process is Inverse Kinematics (IK)—the mathematical process of calculating exactly how every joint in a robotic arm must move to reach a specific target.

Traditional IK methods fall into two camps: Analytical (fast but only work for specific robot designs) and Numerical (flexible but prone to getting stuck in "local optima" or failing at "singularities" where the math breaks down). While newer bio-inspired optimization algorithms show promise, they often converge too slowly or lack the precision needed for sub-millimeter assembly.

Methods

The authors took the recently proposed Secretary Bird Optimization Algorithm (SBOA)—which mimics the hunting and escape behaviors of the African raptor—and "supercharged" it with three strategies:

  1. Oppositional Variational Perturbation: In the early search phase, this expands the exploration area, ensuring the algorithm doesn't miss the best possible solution by looking too narrowly.
  2. Golden Sine Development Strategy: In the middle phase, it uses the "golden section" method to rapidly narrow down the search, drastically speeding up convergence.
  3. Evolutionary Strategy: In the final phase, it uses biological cross-evolution techniques to help the algorithm "jump out" of local traps (sub-optimal solutions).

Crucially, they also redesigned the Objective Function (the "goal" the algorithm tries to reach). It doesn't just look at position; it includes a "soft and smooth" criterion that penalizes jerky joint movements, which naturally helps the robot avoid dangerous singular positions.

Datasets

The researchers validated their approach using three distinct tiers of testing:

  • Numerical Benchmarks: The CEC2017 test set, a standard suite of 30 complex mathematical functions used to stress-test optimization algorithms.
  • Simulated Robots: Digital models of 4-DOF (SCARA), 6-DOF (UR16e), and 7-DOF robotic arms to test how the algorithm handles increasing mechanical complexity.
  • Physical Validation: A real-world experimental platform using a UR16e robotic arm and 3D-printed substitutes for pyrotechnic grains to test assembly success rates.

Key Insights and Ideas

  • Diversity is a Strength: By using Principal Component Analysis (PCA) and K-means clustering, the authors proved that ISBOA finds a much wider variety of valid joint configurations than other algorithms. This "solution diversity" allows engineers to pick the smoothest path for the robot.
  • The Accuracy-Stability Trade-off: Traditional Newton-based methods are incredibly fast but "brittle"—if they start with the wrong guess, they fail. ISBOA is "robust"; it takes longer to calculate but almost always finds a high-precision path without crashing.
  • Smoothness as a Constraint: By mathematically rewarding small joint changes, the algorithm inherently avoids "jitter" and singularities, which are the primary causes of mechanical failure in high-precision robotics.

Key Results

  • Algorithm Ranking: In the CEC2017 tests, ISBOA ranked 1st in 18 out of 30 functions, outperforming seven other modern algorithms like Particle Swarm Optimization (PSO) and the Slime Mould Algorithm (SMA).
  • Extreme Precision: For a complex 7-DOF arm, ISBOA achieved a pose error of $1.28 \times 10^{-13}$, which is eight orders of magnitude more accurate than standard PSO ($7.03 \times 10^{-5}$).
  • Physical Success: In 30 real-world assembly trials, ISBOA achieved a 96.7% success rate, compared to 86.7% for the traditional analytical method.
  • Reliability: ISBOA encountered only 2 singular positions during testing, whereas the traditional analytical method hit 4 and the Newton iterative method hit 3.

Limitations and Future Work

The primary drawback of ISBOA is computational speed. While it is far more accurate, its running time for a full path was approximately 8.08 seconds, compared to just 0.02 seconds for the Newton iterative method. This makes it difficult to use for real-time, high-speed reactive movements.

The authors suggest that future work should involve using Neural Networks to "learn" the solutions provided by ISBOA. This could combine the high-precision "wisdom" of the ISBOA algorithm with the millisecond-level execution speeds of a trained AI model.

Summary updated

Abstract

The assembly of pyrotechnic grain demands high precision and stability in robotic arm motion control due to the small shell apertures and stringent assembly accuracy requirements. Inverse kinematics is a core technology in robotic arm motion control. This paper constructs a robotic arm inverse kinematics model by reformulating the inverse kinematics problem as a constrained optimization problem and employs a multi-strategy improved Secretary Bird Optimization Algorithm (ISBOA) to achieve high-precision solutions. Aiming at the problems of restricted solution set exploration, easy to fall into local optimization and slow convergence when solving the inverse kinematics of multi-DOF robotic arm by SBOA, this paper introduces the oppositional variational perturbation, golden sine development and evolutionary strategy to optimize the formation of ISBOA, and verifies its effectiveness through numerical experiments. Simulation experiments using 4, 6, and 7-DOF robotic arms are conducted, with inverse solution results analyzed via PCA dimensionality reduction and K-means clustering, demonstrating the superiority of ISBOA in inverse solution diversity. Finally, a MATLAB-CoppeliaSim-UR16e experimental platform is developed to compare ISBOA with traditional analytical and Newton iterative method. Results are analyzed in terms of assembly accuracy, singular position handling, grasping success rate, and assembly success rate, confirming ISBOA's advantages in pyrotechnic grain assembly and its potential for engineering applications.
BibTeX
@inproceedings{chen2025multi,
  author = {Xuhao Chen and Nan Lin and Fan Zhang and Xuhai Zhao},
  title = {{A multi-strategy enhanced secretary bird optimization algorithm for high-precision inverse kinematics in robotic arms}},
  booktitle = {PLoS},
  year = {2025},
  doi = {10.1371/journal.pone.0331041},
}
Figure