IEEE International Conference on Robotics and Automation

Display Omitted Chaotic maps have been embedded into Gravitational Search Algorithms (GSA) for the first time.The problem of trapping in local minima in GSA has been improved by the chaotic maps.The convergence rate of GSA has been improved.The statistical test allowed us to judge about the significance of the results.An adaptive normalization is proposed to smoothly transit from the exploration phase to the exploitation phase. In a population-based meta-heuristic, the search process is divided into two main phases: exploration versus exploitation. In the exploration phase, a random behavior is fruitful to explore the search space as extensive as possible. In contrast, a fast exploitation toward the promising regions is the main objective of the latter phase. It is really challenging to find a proper balance between these two phases because of the stochastic nature of population-based meta-heuristic algorithms. The literature shows that chaotic maps are able to improve both phases. This work embeds ten chaotic maps into the gravitational constant (G) of the recently proposed population-based meta-heuristic algorithm called Gravitational Search Algorithm (GSA). Also, an adaptive normalization method is proposed to transit from the exploration phase to the exploitation phase smoothly. As case studies, twelve shifted and biased benchmark functions evaluate the performance of the proposed chaos-based GSA algorithms in terms of exploration and exploitation. A statistical test called Wilcoxon rank-sum is done to judge about the significance of the results as well. The results demonstrate that sinusoidal map is the best map for improving the performance of GSA significantly.

Chaotic gravitational constants for the gravitational search algorithm

This paper proposed a new methodology to determine the optimal trajectory of the path for multi-robot in a clutter environment using hybridization of improved particle swarm optimization (IPSO) with an improved gravitational search algorithm (IGSA). The proposed approach embedded the social essence of IPSO with motion mechanism of IGSA. The proposed hybridization IPSO–IGSA maintain the efficient balance between exploration and exploitation because of adopting co-evolutionary techniques to update the IGSA acceleration and particle positions with IPSO velocity simultaneously. The objective of the algorithm is to minimize the maximum path length that corresponds to minimize the arrival time of all robots to their respective destination in the environment. The robot on the team make independent decisions, coordinate, and cooperate with each other to determine the next positions from their current position in the world map using proposed hybrid IPSO–IGSA. Finally the analytical and experimental results of the multi-robot path planning were compared to those obtained by IPSO–IGSA, IPSO, IGSA in a similar environment. The Simulation and the Khepera environment result show outperforms of IPSO–IGSA as compared with IPSO and IGSA with respect to optimize the path length from predefine initial position to designation position ,energy optimization in the terms of number of turn and arrival time.

A hybridization of an improved particle swarm optimization and gravitational search algorithm for multi-robot path planning

Gravitational Search Algorithm (GSA) is an optimization method inspired by the theory of Newtonian gravity in physics. Till now, many variants of GSA have been introduced, most of them are motivated by gravity-related theories such as relativity and astronomy. On the one hand, to solve different kinds of optimization problems, modified versions of GSA have been presented such as continuous (real), binary, discrete, multimodal, constraint, single-objective, and multi-objective GSA. On the other hand, to tackle the difficulties in real-world problems, the efficiency of GSA has been improved using specialized operators, hybridization, local search, and designing the self-adaptive algorithms. Researchers have utilized GSA to solve various engineering optimization problems in diverse fields of applications ranging from electrical engineering to bioinformatics. Here, we discussed a comprehensive investigation of GSA and a brief review of GSA developments in solving different engineering problems to build up a global picture and to open the mind to explore possible applications. We also made a number of suggestions that can be undertaken to help move the area forward.

A comprehensive survey on gravitational search algorithm

Human action recognition using genetic algorithms and convolutional neural networks

This paper proposes a new Gravitational Search Algorithm (GSA)-based approach for generating an optimal path for a robot travelling in partially unknown environments in the presence of multiple (static or dynamic) obstacles. The GSA-based approach is expressed as an algorithm which computes an optimal path for a robot that travels from an initial point to a target point while avoiding all the known obstacles in the environment but also any other static or dynamic object that could appear in the path of the robot to the target point. To validate the new approach for the path planning, the new algorithm is employed in the generation of obstacle-free paths for different robots that are participating in different missions in the framework of the nRobotic platform developed at the “Politehnica” University of Timisoara, Romania. A comparison focused on the resulted path length and performance with another well-known evolutionary algorithm represented by the Particle Swarm Optimization used for path planning is performed.

Optimal Robot Path Planning Using Gravitational Search Algorithm

This paper proposes the combination of tensor product (TP)-based model transformation and of fuzzy control as a cost effective cascade control system (CS) structure. A novel two-step design approach is suggested. A parallel distributed compensation technique is used in the first step in the TP-based design of the inner state feedback control loops which is next characterized by simple models using equivalent state feedback matrices. The modulus optimum method and the modal equivalent principle are used in the second step to design the Mamdani PI-fuzzy controller in the outer control loops. The cost effective feature of the original cascade CS combination concerns the simplicity of the CS structure, of the controllers and of the design approach. The presentation is focused on the level control in horizontal three tank systems where two cascade CS structures are separately designed to control two water levels. The efficiency of the cascade CS structure is proved by real-time experimental results for a laboratory three tank system. Comparisons with the Mamdani PI-fuzzy CS and with the cascade combination of TP and PI CS are included.

https://onlinelibrary.wiley.com/doi/pdf/10.1002/asjc.855

Cascade Control System‐Based Cost Effective Combination of Tensor Product Model Transformation and Fuzzy Control

This paper proposes an optimal path planning algorithm for mobile robots based on a hybridization between a Gravitational Search Algorithm (GSA) and a Particle Swarm Optimization (PSO) algorithm and referred to as hybrid PSO-GSA. The multi-objective optimization is considered as the PSO-GSA uses two objective functions to generate optimal trajectories for mobile robots in static environments while avoiding collisions with the obstacles and danger zones that might exist in the environment. The hybrid PSO-GSA solves the optimization problems by minimizing the objective functions, producing optimal collision-free trajectories in terms of minimizing the length of the path that needs to be followed by the robot and also assuring that the generated trajectories are at a safe distance from the danger zones. The proposed hybrid PSO-GSA path planning algorithm is validated by running several experiments with robots in different environments in the presence of multiple obstacles and multiple danger zones.

Hybrid PSO-GSA robot path planning algorithm in static environments with danger zones

This paper discusses new optimal path planning algorithms based on a Gravitational Search Algorithm (GSA) and on a Particle Swarm Optimization (PSO) algorithm applied to multiple mobile robots on holonomic wheeled platforms. Four path planning objectives are aggregated in the definition of a set of separate optimization problems for each robot. Each optimization problem involves a single scalar objective function (o.f.) expressed as the weighted sum of four o.f.s. The GSA and the PSO algorithm solve the optimization problems aiming the minimization of the o.f.s, and the path planning algorithms transform the solutions to the optimization problems into optimal solutions as optimal paths. The GSA-and PSO-based path planning algorithms are validated by several experiments with multiple robots that participate in missions.

Multi-robot GSA- and PSO-based optimal path planning in static environments

This paper presents aspects concerning embedding Gravitational Search Algorithms (GSAs) in Convolutional Neural Networks (CNNs) for Optical Character Recognition (OCR) systems. The GSAs are used in combination with the Back Propagation (BP) algorithm as optimization algorithms in the training process of a specific CNN architecture for OCR applications. The new algorithm consists of applying first the GSA and next the BP in order to ensure performance improvements by avoiding the algorithms' traps in local minima. A performance analysis for a given benchmark application shows the advantages of our algorithm over the classical BP algorithm for a six layer CNN dedicated to OCR applications.

Lucian-Ovidiu Fedorovici

Papers

Optimal Robot Path Planning Using Gravitational Search Algorithm

Cascade Control System‐Based Cost Effective Combination of Tensor Product Model Transformation and Fuzzy Control

Hybrid PSO-GSA robot path planning algorithm in static environments with danger zones

Multi-robot GSA- and PSO-based optimal path planning in static environments

Embedding Gravitational Search Algorithms in Convolutional Neural Networks for OCR applications