New issue arrives! Read a highlight review and some latest research.

  • Share:

A new issue is available now, covering the topics of deep learning, image processing, speech analysis, emotion recognition, etc.. Read full text at http://www.ijac.netor http://link.springer.com/journal/11633.

 

【Review】DeepLearning Based Single Image Super-resolution: A Survey

Viet Khanh Ha, Jin-ChangRen, Xin-Ying Xu, Sophia Zhao, Gang Xie, ValentinMasero, Amir Hussain

Abstract:

Single image super-resolution has attractedincreasing attention and has a wide range of applications in satellite imaging,medical imaging, computer vision, security surveillance imaging, remotesensing, objection detection, and recognition. Recently, deep learningtechniques have emerged and blossomed, producing "the state-of-the-art” inmany domains. Due to their capability in feature extraction and mapping, it isvery helpful to predict high-frequency details lost in low-resolution images.In this paper, we give an overview of recent advances in deep learning-basedmodels and methods that have been applied to single image super-resolutiontasks. We also summarize, compare and discuss various models from the past andpresent for comprehensive understanding and finally provide open problems andpossible directions for future research.

Keywords:

Image super-resolution, convolutionalneural network, high-resolution image, low-resolution image, deep learning.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-019-1183-x

http://link.springer.com/article/10.1007/s11633-019-1183-x

 

 



◇◇◇◇◇◇

Large-scale Data Collection and Analysisvia a Gamified Intelligent Crowdsourcing Platform

Simone Hantke, Tobias Olenyi, ChristophHausner, Tobias Appel, Björn Schuller

Abstract:

In this contribution, we present iHEARu-PLAY, an online,multi-player platform for crowdsourced database collection and labelling,including the voice analysis application (VoiLA), a free web-based speechclassification tool designed to educate iHEARu-PLAY users aboutstate-of-the-art speech analysis paradigms. Via this associated speech analysisweb interface, in addition, VoiLA encourages users to take an active role inimproving the service by providing labelled speech data. The platform allowsusers to record and upload voice samples directly from their browser, which arethen analysed in a state-of-the-art classification pipeline. A set ofpre-trained models targeting a range of speaker states and traits such asgender, valence, arousal, dominance, and 24 different discrete emotions isemployed. The analysis results are visualised in a way that they are easilyinterpretable by laymen, giving users unique insights into how their voicesounds. We assess the effectiveness of iHEARu-PLAY and its integrated VoiLAfeature via a series of user evaluations which indicate that it is fun and easyto use, and that it provides accurate and informative results.

Keywords:

Human computation, speech analysis,crowdsourcing, gamified data collection, survey.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-019-1180-0

http://link.springer.com/article/10.1007/s11633-019-1180-0

 

 



◇◇◇◇◇◇

Semi-supervised Ladder Networksfor Speech Emotion Recognition

Jian-Hua Tao, Jian Huang, Ya Li, Zheng Lian, Ming-Yue Niu

Abstract:

As a major component of speech signal processing, speech emotionrecognition has become increasingly essential to understanding humancommunication. Benefitting from deep learning, many researchers have proposedvarious unsupervised models to extract effective emotional features andsupervised models to train emotion recognition systems. In this paper, weutilize semi-supervised ladder networks for speech emotion recognition. Themodel is trained by minimizing the supervised loss and auxiliary unsupervisedcost function. The addition of the unsupervised auxiliary task providespowerful discriminative representations of the input features, and is alsoregarded as the regularization of the emotional supervised task. We alsocompare the ladder network with other classical autoencoder structures. Theexperiments were conducted on the interactive emotional dyadic motion capture(IEMOCAP) database, and the results reveal that the proposed methods achievesuperior performance with a small number of labelled data and achieves betterperformance than other methods.

Keywords:

Speech emotion recognition, the ladder network, semi-supervisedlearning, autoencoder, regularization.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-019-1175-x

http://link.springer.com/article/10.1007/s11633-019-1175-x

 

 



◇◇◇◇◇◇

Pedestrian Height Estimation and3D Reconstruction Using Pixel-resolution Mapping Method Without SpecialPatterns

Bing-Xing Wu, Suat Utku Ay, Ahmed Abdel-Rahim

Abstract:

Extracting the three-dimensional (3D) information including locationand height of a pedestrian is important for vision-based intelligent trafficmonitoring systems. This paper tackles the relationship between pixels′ actualsize and pixels′ spatial resolution through a new method named pixel-resolutionmapping (P-RM). The proposed P-RM method derives the equations for pixels′spatial resolutions (XY-direction) and object′s height (Z-direction) in thereal world, while introducing new tilt angle and mounting height calibrationmethods that do not require special calibration patterns placed in the real world.Both controlled laboratory and actual world experiments were performed andreported. The tests on 3D mensuration using proposed P-RM method showed overallbetter than 98.7% accuracy in laboratory environments and better than 96%accuracy in real world pedestrian height estimations. The 3D reconstructed imagesfor measured points were also determined with the proposed P-RM method whichshows that the proposed method provides a general algorithm for 3D informationextraction.

Keywords:

Traffic monitoring application, spatial resolution, pixel-resolutionmapping (P-RM) method, 3D information, pedestrian height estimation.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-019-1170-2

http://link.springer.com/article/10.1007/s11633-019-1170-2

 

 



◇◇◇◇◇◇

Synthesizing Robot Programs withInteractive Tutor Mode

Hao Li, Yu-Ping Wang, Tai-Jiang Mu

Abstract:

With the rapid development of the robotic industry, domestic robotshave become increasingly popular. As domestic robots are expected to bepersonal assistants, it is important to develop a natural language-basedhuman-robot interactive system for end-users who do not necessarily have muchprogramming knowledge. To build such a system, we developed an interactivetutoring framework, named “Holert”, which can translate task descriptions innatural language to machine-interpretable logical forms automatically. Comparedto previous works, Holert allows users to teach the robot by further explainingtheir intentions in an interactive tutor mode. Furthermore, Holert introduces asemantic dependency model to enable the robot to “understand” similar taskdescriptions. We have deployed Holert on an open-source robot platform,Turtlebot 2. Experimental results show that the system accuracy could besignificantly improved by 163.9% with the support of the tutor mode. Thissystem is also efficient. Even the longest task session with 10 sentences can behandled within 0.7 s.

Keywords:

Human-robot interaction, semantic parsing, program synthesis,intelligent robotic systems, natural language understanding.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-018-1154-7

http://link.springer.com/article/10.1007/s11633-018-1154-7

 

 



◇◇◇◇◇◇

Time-space Viewpoint Planning forGuard Robot with Chance Constraint

Igi Ardiyanto, Jun Miura

Abstract:

This paper presents a novel movement planning algorithm for a guardrobot in an indoor environment, imitating the job of human security. A movementplanner is employed by the guard robot to continuously observe a certainperson. This problem can be distinguished from the person following problemwhich continuously follows the object. Instead, the movement planner aims toreduce the movement and the energy while keeping the target person under itsvisibility. The proposed algorithm exploits the topological features of theenvironment to obtain a set of viewpoint candidates, and it is then optimizedby a cost-based set covering problem. Both the robot and the target person aremodeled using geodesic motion model which considers the environment shape.Subsequently, a particle model-based planner is employed, considering thechance constraints over the robot visibility, to choose an optimal action forthe robot. Simulation results using 3D simulator and experiments on a realenvironment are provided to show the feasibility and effectiveness of ouralgorithm.

Keywords:

Guard robot, viewpoint planning, state-time space, uncertainty,topology, chance-constraint.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-018-1146-7

http://link.springer.com/article/10.1007/s11633-018-1146-7

 

 



◇◇◇◇◇◇

Experimental Evaluation ofCertain Pursuit and Evasion Schemes for Wheeled Mobile Robots

Amit Kumar, Aparajita Ojha

Abstract:

Pursuit-evasion games involving mobile robots provide an excellentplatform to analyze the performance of pursuit and evasion strategies.Pursuit-evasion has received considerable attention from researchers in thepast few decades due to its application to a broad spectrum of problems thatarise in various domains such as defense research, robotics, computer games,drug delivery, cell biology, etc. Several methods have been introduced in theliterature to compute the winning chances of a single pursuer or single evaderin a two-player game. Over the past few decades, proportional navigationguidance (PNG) based methods have proved to be quite effective for the purposeof pursuit especially for missile navigation and target tracking. However, aperformance comparison of these pursuer-centric strategies against recentevader-centric schemes has not been found in the literature, for wheeled mobilerobot applications. With a view to understanding the performance of each of theevasion strategies against various pursuit strategies and vice versa, fourdifferent proportional navigation-based pursuit schemes have been evaluatedagainst five evader-centric schemes and vice-versa for non-holonomic wheeledmobile robots. The pursuer′s strategies include three well-known schemesnamely, augmented ideal proportional navigation guidance (AIPNG), modifiedAIPNG, angular acceleration guidance (AAG), and a recently introducedpursuer-centric scheme called anticipated trajectory-based proportionalnavigation guidance (ATPNG). Evader-centric schemes are classic evasion, randommotion, optical-flow based evasion, Apollonius circle-based evasion and anotherrecently introduced evasion strategy called anticipated velocity-based evasion.The performance of each of the pursuit methods was evaluated against fivedifferent evasion methods through hardware implementation. The performance wasanalyzed in terms of time of interception and the distance traveled by players.The working environment was obstacle-free and the maximum velocity of thepursuer was taken to be greater than that of the evader to conclude the game infinite time. It is concluded that ATPNG performs better than other PNG-basedschemes, and the anticipated velocity-based evasion scheme performs better thanthe other evasion schemes.

Keywords:

Pursuit-evasion, wheeled mobile robot, proportional navigation,trajectory planning, target interception.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-018-1151-x

http://link.springer.com/article/10.1007/s11633-018-1151-x

 

 



◇◇◇◇◇◇

Control of a Two-wheeled Machinewith Two-directions Handling Mechanism Using PID and PD-FLC Algorithms

Khaled M. Goher, Sulaiman O. Fadlallah

Abstract:

This paper presents a novel five degrees of freedom (DOF)two-wheeled robotic machine (TWRM) that delivers solutions for both industrialand service robotic applications by enlarging the vehicle′s workspace andincreasing its flexibility. Designing a two-wheeled robot with five degrees offreedom creates a high challenge for the control, therefore the modelling anddesign of such robot should be precise with a uniform distribution of mass overthe robot and the actuators. By employing the Lagrangian modelling approach,the TWRM′s mathematical model is derived and simulated in Matlab/Simulink®. Forstabilizing the system′s highly nonlinear model, two control approaches weredeveloped and implemented: proportional-integral-derivative (PID) and fuzzylogic control (FLC) strategies. Considering multiple scenarios with differentinitial conditions, the proposed control strategies′ performance has beenassessed.

Keywords:

Two-wheeled inverted pendulum (IP) with two direction handling,Lagrangian formulation, proportional-integral-derivative (PID), fuzzy logiccontrol (FLC), under-actuated systems.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-019-1172-0

http://link.springer.com/article/10.1007/s11633-019-1172-0

 

 



◇◇◇◇◇◇

Convergence Analysis of a NewMaxMin-SOMO Algorithm

Atlas Khan, Yan-Peng Qu, Zheng-Xue Li

Abstract:

The convergence analysis of MaxMin-SOMO algorithm is presented. TheSOM-based optimization (SOMO) is an optimization algorithm based on theself-organizing map (SOM) in order to find a winner in the network. Generally,through a competitive learning process, the SOMO algorithm searches for theminimum of an objective function. The MaxMin-SOMO algorithm is the generalizationof SOMO with two winners for simultaneously finding two winning neurons i.e.,first winner stands for minimum and second one for maximum of the objectivefunction. In this paper, the convergence analysis of the MaxMin-SOMO ispresented. More specifically, we prove that the distance between neuronsdecreases at each iteration and finally converge to zero. The work is verified withthe experimental results.

Keywords:

Optimization, self-organizing map (SOM), SOM-based optimization(SOMO) algorithm, particle swarm optimization (PSO), genetic algorithms (GAs).

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-016-0996-0

http://link.springer.com/article/10.1007/s11633-016-0996-0

 

 



◇◇◇◇◇◇

Stabilization for a Class ofDiscrete-time Switched Large-scale Systems with Parameter Uncertainties

Chang-Chun Sun

Abstract:

The problem of robust stabilization for a class of discrete-time switchedlarge-scale systems with parameter uncertainties and nonlinear interconnectedterms is considered. By using state feedback and Lyapunov function technique, adecentralized switching control approach is put forward to guarantee thesolutions of large-scale systems converge to the origin globally. A numericalexample and a corresponding simulation result are utilized to verify the effectivenessof the presented approach.

Keywords:

Switched large-scale systems, discrete-time, state feedback, decentralizedcontrol, switching law.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-016-0966-6

http://link.springer.com/article/10.1007/s11633-016-0966-6

 

 



◇◇◇◇◇◇

A Linear Quadratic ControllerDesign Incorporating a Parametric Sensitivity Constraint

Mohamed Yagoubi

Abstract:

The purpose of this paper is to propose a synthesis method ofparametric sensitivity constrained linear quadratic (SCLQ) controller for anuncertain linear time invariant (LTI) system. System sensitivity to parametervariation is handled through an additional quadratic trajectory parametricsensitivity term in the standard LQ criterion to be minimized. The main purposehere is to find a suboptimal linear quadratic control taking explicitly intoaccount the parametric uncertainties. The paper main contribution is threefold:1) A descriptor system approach is used to show that the underlying singularlinear-quadratic optimal control problem leads to a non-standard Riccatiequation. 2) A solution to the proposed control problem is then given based ona connection to the so-called Lur0e matrix equations. 3) A synthesis method ofmultiple parametric SCLQ controllers is proposed to cover the whole parametricuncertainty while degrading as less as possible the intrinsic robustnessproperties of each local linear quadratic controller. Some examples arepresented in order to illustrate the effectiveness of the approach.

Keywords:

Linear quadratic control, parametric uncertainties, trajectorysensitivity, non-standard Riccati equation, Lur’e matrix equations, linear timeinvariant (LMI), particle swarm optimization.

Full Text:

http://www.ijac.net/en/article/doi/10.1007/s11633-016-1048-5

http://link.springer.com/article/10.1007/s11633-016-1048-5

 

 

 



◆◆◆◆◆◆

For moreup-to-date information:

1) WeChat: IJAC

2) Twitter:IJAC_Journal

3) Facebook:International Journal of Automation and Computing

4) Linkedin: Int.J. of Automation and Computing

5) Sina Weibo:IJAC-国际自动化与计算杂志

Current Issue

2019 Vol.16 No.5

Table of Contents

ISSN 1476-8186

E-ISSN 1751-8520

CN 11-5350/TP

Editors-in-chief
Tieniu TAN, Chinese Academy of Sciences Guoping LIU, University of South Wales Huosheng HU, University of Essex
Global Visitors