Dynamic modelling and parameter estimation of a hydraulic robot manipulator using a multi-objective genetic algorithm

Abstract

This article concerns the problem of dynamic modeling and parameter estimation for a seven degree of freedom hydraulic manipulator. The laboratory example is a dual-manipulator mobile robotic platform used for research into nuclear decommissioning. In contrast to earlier control model orientated research using the same machine, the article develops a nonlinear, mechanistic simulation model that can subsequently be used to investigate physically meaningful disturbances. The second contribution is to optimize the parameters of the new model, i.e. to determine reliable estimates of the physical parameters of a complex robotic arm which are not known in advance. To address the nonlinear and non-convex nature of the problem, the research relies on the multi-objectivization of an output error single performance index. The developed algorithm utilises a multi-objective Genetic Algorithm (GA) in order to find a proper solution. The performance of the model and the GA is evaluated using both simulated (i.e. with a known set of 'true' parameters) and experimental data. Both simulation and experimental results show that multi-objectivization has improved convergence of the estimated parameters compared to the single objective output error problem formulation. This is achieved by integrating the validation phase inside the algorithm implicitly and exploiting the inherent structure of the multi-objective GA for this specific system identification problem.

Vision-Based Assisted Tele-Operation of a Dual-Arm Hydraulically Actuated Robot for Pipe Cutting and Grasping in Nuclear Environments

“…The present article concerns an investigation into some of the concepts raised above, using a previously developed dual-manipulator robotic platform in a non-active laboratory environment [1][2][3]. The aim is to develop semi-automatic control systems that reduce operator workload, speeding up task execution and reducing training time, whilst minimizing the introduction of additional sensors and other components.…”

“…However, in comparison to a typical machine driven by electric motors, hydraulic actuators generally have higher loop gains, wider bandwidths and lightly-damped, nonlinear dynamics. For the robotic platform in the present study, these dynamics have been previously represented using physically-based (mechanistic) equations [3,46]. However, the present article instead focuses on the identification from data of a relatively straightforward State-Dependent Parameter (SDP) model form, which is used here directly for nonlinear control design.…”

“…This article concerns assisted tele-operation systems for nuclear decommissioning and the characterisation of nuclear environments. To bring focus, the development and experimental work is based on an illustrative, hydraulically actuated, dual-manipulator device [1][2][3]. At present, direct tele-operation is standard in the nuclear industry and vision systems are rare, except for those used simply to allow for remote tele-operation [4].…”

Using an Improved Artificial Bee Colony Algorithm for Parameter Estimation of a Dynamic Grain Flow Model

“…The particle swarm optimization (PSO) to determine the unknown photovoltaic (PV) model parameters was proposed by Soon and Low [18]. A multiobjective genetic algorithm (MOGA) is utilized to determine reliable estimates of the physical parameters of a complex robotic arm which are not known in advance [19]. Artificial intelligence techniques are superior to the traditional methods and have the obvious advantages of solving the global optimization problem with high precision and short time.…”

An Electro-Hydraulic Servo with Intelligent Control Strategy

“…In the previous works, there are many type of control techniques have been reported, which can be utilized to control the tracking capability of a nonlinear electrohydraulic actuator system. Each of the control techniques required a proper tuning technique and some of the advanced tuning techniques have been reported recently such as Particle Swarm Optimization (PSO) [13][14][15][16][17][18], Genetic Algorithm (GA) [19][20][21], and Differential Evolution (DE) [22,23].…”