Neural Navigation Approach of an Autonomous Mobile Robot in a Partially Structured Environment

Chohra, Amine; Sif, F.; Talaoubrid, S.

doi:10.1016/s1474-6670(17)46976-x

Cited by 9 publications

(15 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Neural network based recognition and classification schemes have been widely investigated by researchers in a diverse range of research fields, such as pattern recognition [23,24], image processing [22], and mobile robotics [19][20][21]. The key idea, adopting methods that are inspired by biological systems, is unchanged even though the motivation of research comes from different research areas.…”

Section: Neural Network As a Binary Classifiermentioning

confidence: 99%

See 1 more Smart Citation

A neural network based retrainable framework for robust object recognition with application to mobile robotics

Kang

Choi

et al. 2010

Appl Intell

View full text Add to dashboard Cite

In this paper, we address object recognition for a mobile robot which is deployed in a multistory building. To move to another floor, a mobile robot should recognize various objects related to an elevator, e.g., elevator control, call buttons, and LED displays. To this end, we propose a neural network based retrainable framework for object recognition, which consists of four components-preprocessing, binary classification, object identification, and outlier rejection. The binary classifier, a key component of our system, is a neural network that can be retrained, the motivation of which is to adapt to varying environments, especially with illuminations. Without incurring any extra process to prepare new training samples for retraining, they are freely obtained as a result of the outlier rejection component, being extracted on-line. To realize a practical system, we adopt a parallel architecture integrating both recognition and retraining processes for seamless object recognition, and furthermore detect and cope with the deterioration of a retrained neural network to ensure high reliability. We demonstrate the positive effect of retraining on the object recognition performance by conducting experiments over hundreds of images obtained in daytime and nighttime.

show abstract

Section: Neural Network As a Binary Classifiermentioning

confidence: 99%

“…However, if all initial training samples except the representative ones are removed from each image queue, then (19) and (20) are slightly modified as follows:…”

Section: How To Organize a New Training Setmentioning

confidence: 99%

A neural network based retrainable framework for robust object recognition with application to mobile robotics

Kang

Choi

et al. 2010

Appl Intell

View full text Add to dashboard Cite

show abstract

“…In navigation, the NN learns the relationship between the current sensor measurements and the robot's local environment, to plan a collisionfree path toward the goal [28][29][30]. In localization, the NN learns the location of the feature positions observed from various robot poses [31].…”

Section: Neural Network As a Model For Estimating The Robot Trajectormentioning

confidence: 99%

A new approach to simultaneous localization and map building with implicit model learning using neuro evolutionary optimization

Kang

Kim

et al. 2010

Appl Intell

View full text Add to dashboard Cite

This paper presents Neuro-Evolutionary Optimization SLAM (NeoSLAM) a novel approach to SLAM that uses a neural network (NN) to autonomously learn both a nonlinear motion model and the noise statistics of measurement data. The NN is trained using evolutionary optimization to learn the residual error of the motion model, which is then added to the odometry data to obtain the full motion model estimate. Stochastic optimization is used, to accommodate any kind of cost function. Prediction and correction are performed simultaneously within our neural framework, which implicitly integrates the motion and sensor models. An evolutionary programming (EP) algorithm is used to progressively refine the neural model until it generates a trajectory that is most consistent with the actual sensor measurements. During this learning process, NeoSLAM does not require any prior knowledge of motion or sensor models and shows consistently good performance regardless of the robot and the sensor noise type. Furthermore, NeoSLAM does not require the data association step at loop closing which is crucial in most other SLAM algorithms, but can still generate an accurate map. Experiments in various complex environments with widely-varying types of noise show that the learning capability of NeoSLAM ensures performance that is consistently less sensitive to noise and more accurate than that of other SLAM methods.

show abstract

“…In this paper, an interesting navigation approach suggested in Azouaoui et al (2008), Azouaoui and Chohra (2002) and Chohra et al (1998) is applied with some modifications K 42,2 to a car-like robot Robucar with addition of an action behavior to deal with the generation of smooth motions. Indeed, this approach is based on four basic behaviors; three of them are fused under a neural paradigm using gradient back-propagation (GBP) and reinforcement learning (RL) algorithms and the last behavior uses a FL controller.…”

Section: Introductionmentioning

confidence: 99%

Soft‐computing based navigation approach for a bi‐steerable mobile robot

et al. 2013

View full text Add to dashboard Cite

Purpose -The purpose of this paper is to present an implementation of a soft-computing (SC) based navigation approach on a bi-steerable mobile robot, Robucar. This approach must provide Robucar with capability to acquire the obstacle avoidance, target localization, decision-making and action behaviors after learning and adaptation. This approach uses three neural networks (NN) and fuzzy logic (FL) controller to achieve the desired task. The NNs corresponding to the obstacle avoidance and target localization are trained using the back-propagation algorithm and the last one is based on the reinforcement learning paradigm while the FL controller uses the Mamdani search and match algorithm. Simulation and experimental results are presented, showing the effectiveness of the overall navigation control system. Design/methodology/approach -In this paper, an interesting navigation approach is applied to a car-like robot, Robucar, with addition of an action behavior to deal with the generation of smooth motions. Indeed, this approach is based on four basic behaviors; three of them are fused under a neural paradigm using Gradient Back-Propagation (GBP) and reinforcement learning (RL) algorithms and the last behavior uses a FL controller. It uses a set of suggested rules to describe the control policy to achieve the action behavior. Findings -In the implemented SC-based navigation, the intelligent behaviors necessary to the navigation are acquired by learning using GBP algorithm and adaptation using FL. The proposed approach provides Robucar with more autonomy, intelligence and real-time processing capabilities. Indeed, the proposed NNs and FLC are able to remedy problems of analytical approaches, missing or incorrect environment knowledge and uncertainties which can lead to undesirable effects as the rough velocity changes. The simulation and experimental results display the ability of the proposed SC-based navigation approach to provide Robucar with capability to intelligently navigate in a priori unknown environment, illustrating the robustness and adaptation capabilities of the approach. Research limitations/implications -This work can be extended to consider mobile obstacles with a velocity higher than the velocity of the robot. Originality/value -This paper presents a learning approach to navigating a bi-steerable mobile robot in an unknown environment using GBP and RL paradigms.

show abstract

Neural Navigation Approach of an Autonomous Mobile Robot in a Partially Structured Environment

Cited by 9 publications

References 19 publications

A neural network based retrainable framework for robust object recognition with application to mobile robotics

A neural network based retrainable framework for robust object recognition with application to mobile robotics

A new approach to simultaneous localization and map building with implicit model learning using neuro evolutionary optimization

Soft‐computing based navigation approach for a bi‐steerable mobile robot

Contact Info

Product

Resources

About