Utilization of Webots and Khepera II as a platform for Neural Q-Learning controllers

Ganapathy, Velappa; Soh, Chin Yun; Lui, Wen Lik Dennis

doi:10.1109/isiea.2009.5356361

Cited by 12 publications

(9 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Pandey et al [21], designed sensor integration based fuzzy logic controller to create a collision-free path, using webots simulator to check simulation result. In [22] [23], webots and Khepera II miniature mobile robot utilized as a platform for the investigation of Neurol Q-Learning controller. Almasri et al [24] present a collision-free mobile robot navigation using fuzzy logic fusion model.…”

Section: Related Workmentioning

confidence: 99%

Collision-Free Mobile Robot navigation using Fuzzy Logic Approach

Dirik¹

2018

IJCA

View full text Add to dashboard Cite

Autonomous mobile robots' navigation has become a very popular and interesting topic of computer science and robotics in the last decade. Many algorithms have been developed for robot motion control in an unknown (indoor/outdoor) and in various environments (static/dynamic). Fuzzy logic control techniques are an important algorithm developed for robot navigation problems. The aim of this research is to design and develop a fuzzy logic controller that enables the mobile robot to navigate to a target in an unknown environment, using WEBOTS commercial mobile robot simulation and MATLAB software. The algorithm is divided into two stages; In the first stage, the mobile robot was made to go to the goal, and in the second stage, obstacle avoidance was realized. Robot position information (x, y, Ø) was used to move the robot to the target and six sensors data were used during the obstacle avoidance phase. The used mobile robot (E_PUCK) is equipped with 12 IR sensors to measure the distance to the obstacles. The fuzzy control system is composed of six inputs grouped in doubles which are left, front and right distance sensors two outputs which are the mobile robot's left and right wheel speeds. To check the simulation result for proposed methodology, WEBOTS simulator and MATLAB software were used. To modeling the environment in different complexity and design, this simulator was used. The experimental results have shown that the proposed architecture provides an efficient and flexible solution for autonomous mobile robots and the objective of this research has been successfully achieved. This research also indicated that WEBOT and MATLAB are suitable tools that could be used to develop and simulate mobile robot navigation system.

show abstract

Section: Related Workmentioning

confidence: 99%

Collision-Free Mobile Robot navigation using Fuzzy Logic Approach

Dirik¹

2018

IJCA

View full text Add to dashboard Cite

show abstract

“…However, Neural Networks (NN), one of the most common form of function approximators, has not been extensively used in conjunction with RL. Although some successful applications have been reported [2] [3][4] [5], most of them choose sequential learning method for an obvious reason. Because of the on-line nature of RL, experience samples come in sequentially and thus incremental learning NN becomes an apparent choice.…”

Section: Introductionmentioning

confidence: 99%

“…It can be randomly initialized and then the weights of NN be tuned during interaction with the environment [3], but it may require a long time before NN can converge to an appropriate form. Another way is to first use tabular method to derive the training data and train a preliminary form of NN.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Receding Horizon Cache and Extreme Learning Machine based Reinforcement Learning

Shao

Huang

2012

2012 12th International Conference on Control Automation Robotics &Amp; Vision (ICARCV)

View full text Add to dashboard Cite

Function approximators have been extensively used in Reinforcement Learning (RL) to deal with large or continuous space problems. However, batch learning Neural Networks (NN), one of the most common approximators, has been rarely applied to RL. In this paper, possible reasons for this are laid out and a solution is proposed. Specifically, a Receding Horizon Cache (RHC) structure is designed to collect training data for NN by dynamically archiving state-action pairs and actively updating their Q-values, which makes batch learning NN much easier to implement. Together with Extreme Learning Machine (ELM), a new RL with function approximation algorithm termed as RHC and ELM based RL (RHC-ELM-RL) is proposed. A mountain car task was carried out to test RHC-ELM-RL and compare its performance with other algorithms.

show abstract

“…Robot path planning has been an active research area, and many methods have been developed to tackle this problem, such as A* algorithm [1]- [2], D* algorithm [3]- [4], reinforcement learning [5]- [6], Q-Learning [7]- [10], global Cspace methods [11], potential field methods [12], fuzzy logic [13]- [15] and neural networks [16]- [17]. Each method has its own strength over others in certain aspects.…”

Section: Introductionmentioning

confidence: 99%

Improved genetic algorithms based optimum path planning for mobile robot

Yun

Ganapathy

Chong

2010

2010 11th International Conference on Control Automation Robotics &Amp; Vision

View full text Add to dashboard Cite

Improved genetic algorithms incorporate other techniques, methods or algorithms to optimize the performance of genetic algorithm. In this paper, improved genetic algorithms of optimum path planning for mobile robot navigation are proposed. An Obstacle Avoidance Algorithm (OAA) and a Distinguish Algorithm (DA) are introduced to generate the initial population in order to improve the path planning efficiency to select only the feasible paths during the evolution of genetic algorithm.Domain heuristic knowledge based crossover, mutation, refinement and deletion operators are specifically designed to fit path planning for mobile robots. Proposed genetic algorithms feature unique, simple path representations, and simple but effective evaluation methods. Simulation studies and real time implementations are carried out to verify and validate the effectiveness of the proposed algorithms. Keywords-Genetic Algorithm (GA), Obstacle Avoidance Algorithm (OAA), Distinguish Algorithm (DA), optimum path planning, mobile robot, Team AmigoBot TM and MATLAB.I.

show abstract

Utilization of Webots and Khepera II as a platform for Neural Q-Learning controllers

Cited by 12 publications

References 6 publications

Collision-Free Mobile Robot navigation using Fuzzy Logic Approach

Collision-Free Mobile Robot navigation using Fuzzy Logic Approach

Receding Horizon Cache and Extreme Learning Machine based Reinforcement Learning

Improved genetic algorithms based optimum path planning for mobile robot

Contact Info

Product

Resources

About