Search results

Estimating the pose – position and orientation – of a moving object such as a robot is a necessary task for many applications, e.g., robot navigation control, environment mapping, and medical applications such as robotic surgery. The purpose of this paper is to introduce a novel method to fuse the information from several available sensors in order to improve the estimated pose from any individual sensor and calculate a more accurate pose for the moving platform.

Pose estimation is usually done by collecting the data obtained from several sensors mounted on the object/platform and fusing the acquired information. Assuming that the robot is moving in a three-dimensional (3D) world, its location is completely defined by six degrees of freedom (6DOF): three angles and three position coordinates. Some 3D sensors, such as IMUs and cameras, have been widely used for 3D localization. Yet, there are other sensors, like 2D Light Detection And Ranging (LiDAR), which can give a very precise estimation in a 2D plane but they are not employed for 3D estimation since the sensor is unable to obtain the full 6DOF. However, in some applications there is a considerable amount of time in which the robot is almost moving on a plane during the time interval between two sensor readings; e.g., a ground vehicle moving on a flat surface or a drone flying at an almost constant altitude to collect visual data. In this paper a novel method using a “fuzzy inference system” is proposed that employs a 2D LiDAR in a 3D localization algorithm in order to improve the pose estimation accuracy.

The method determines the trajectory of the robot and the sensor reliability between two readings and based on this information defines the weight of the 2D sensor in the final fused pose by adjusting “extended Kalman filter” parameters. Simulation and real world experiments show that the pose estimation error can be significantly decreased using the proposed method.

To the best of the authors’ knowledge this is the first time that a 2D LiDAR has been employed to improve the 3D pose estimation in an unknown environment without any previous knowledge. Simulation and real world experiments show that the pose estimation error can be significantly decreased using the proposed method.

Neuromarketing, which is an interdisciplinary area, concentrates on evaluating consumers’ cognitive and emotional reactions to different marketing stimuli. In spite of advantages, neuromarketing still requires development and lacks a strong theoretical framework. Techniques that are used in neuromarketing studies have different superiorities and limitations, and thus, there is a need for the evaluation of the relevance of these techniques. The purpose of this study is to introduce a novel integrated approach for the neuromarketing research area.

The proposed approach combines 2-tuple linguistic representation model and data envelopment analysis to obtain the most efficient neuromarketing technique. It is apt to handle information provided by using both linguistic and numerical scales with multiple information sources. Furthermore, it allows managers to deal with heterogeneous information, without loss of information.

The proposed approach indicates that functional magnetic resonance imaging (fMRI) is the best performing neuromarketing technology. Recently, fMRI has been widely used in neuromarketing research. In spite of its high cost, its main superiorities are improved spatial and temporal resolutions. On the other hand, transcranial magnetic stimulation (TMS) and positron emission tomography (PET) are ranked at the bottom because of their poor resolutions and lower willingness of participants.

This paper proposes a common weight data envelopment analysis (DEA)-based decision model to cope with heterogeneous information collected by the experts to determine the best performing neuromarketing technology. The decision procedure enables the decision-makers to handle the problems of loss of information and multi-granularity by using the fusion of 2-tuple linguistic representation model and fuzzy information. Moreover, a DEA-based common weight model does not require subjective experts’ opinions to weight the evaluation criteria.

The purpose of this paper is to develop a novel recommendation method using online reviews with emotional preferences for facilitating online purchase decisions. This leads to better use of information-rich online reviews for providing users with personalized recommendations.

A novel method is developed for producing personalized recommendations in online purchase decision-making. Such a method fuses the belief structure and the Shapley function together to effectively deal with the emotional preferences in online reviews and adequately tackle the interaction existent between product criteria with the use of a modified combination rule for making better online recommendations for making online purchase decisions.

An example is presented for demonstrating the applicability of the method for facilitating online purchase. The results show that the recommendation using the proposed method can effectively improve customer satisfaction with better purchase decisions.

The proposed method can better utilize online reviews for satisfying personalized needs of consumers. The use of such a method can optimize interface design, refine customer needs, reduce recommendation errors and provide personalized recommendations.

The proposed method adequately considers the characteristics of online reviews and the personalized needs of customers for providing customers with appropriate recommendations. It can help businesses better manage online reviews for improving customer satisfaction and create greater value for both businesses and customers.

With the advancement of social economy, science and technology, nowadays, people face increasingly complex decision-making problems and ever-growing decision-associated information contents. Owing to the unique advantages of the evidence theory, evidence decision flow fits in well with the cognitive process of human beings, which provides us with an effective decision method. However, traditional evidence theories are built upon the one-dimensional evidence recognition framework, which merely reflects the reliability of information determined by evidence source and as such may fail to present the characteristic information of evidence source itself or the information required for the process of determining reliability. This greatly influences processes in decision-making, such as evidence processing and combination. This essay aims to propose a two-dimensional evidence-reasoning theory to address specific decision problems.

This article uses a two-dimensional evidence framework to reflect the process and characteristic information of evidence source based on the traditional evidence framework. As a result, evidence is processed and combined by two-dimensional modified information.

This research is of theoretical and practical significance by extending theoretical connotation, fully utilizing precise evidence information and therefore meeting the requirements of efficient and accurate decision-making performances.

The theory adds a two-dimension to modify and capture evidence on the basis of the traditional evidence framework. This proposal has significant theoretical and practical value to expand evidence theory and provides more accurate use of evidence information, higher efficiency and quality requirements and more precise decision-making.

Content‐based image retrieval (CBIR) is an important research area for automatically retrieving images of user interest from a large database. Due to many potential applications, facial image retrieval has received much attention in recent years. Similar to face recognition, finding appropriate image representation is a vital step for a successful facial image retrieval system. Recently, many efficient image feature descriptors have been proposed and some of them have been applied to face recognition. It is valuable to have comparative studies of different feature descriptors in facial image retrieval. And more importantly, how to fuse multiple features is a significant task which can have a substantial impact on the overall performance of the CBIR system. The purpose of this paper is to propose an efficient face image retrieval strategy.

In this paper, three different feature description methods have been investigated for facial image retrieval, including local binary pattern, curvelet transform and pyramid histogram of oriented gradient. The problem of large dimensionalities of the extracted features is addressed by employing a manifold learning method called spectral regression. A decision level fusion scheme fuzzy aggregation is applied by combining the distance metrics from the respective dimension reduced feature spaces.

Empirical evaluations on several face databases illustrate that dimension reduced features are more efficient for facial retrieval and the fuzzy aggregation fusion scheme can offer much enhanced performance. A 98 per cent rank 1 retrieval accuracy was obtained for the AR faces and 91 per cent for the FERET faces, showing that the method is robust against different variations like pose and occlusion.

The proposed method for facial image retrieval has a promising potential of designing a real‐world system for many applications, particularly in forensics and biometrics.

Internet has endorsed a tremendous change with the advancement of the new technologies. The change has made the users of the internet to make comments regarding the service or product. The Sentiment classification is the process of analyzing the reviews for helping the user to decide whether to purchase the product or not.

A rider feedback artificial tree optimization-enabled deep recurrent neural networks (RFATO-enabled deep RNN) is developed for the effective classification of sentiments into various grades. The proposed RFATO algorithm is modeled by integrating the feedback artificial tree (FAT) algorithm in the rider optimization algorithm (ROA), which is used for training the deep RNN classifier for the classification of sentiments in the review data. The pre-processing is performed by the stemming and the stop word removal process for removing the redundancy for smoother processing of the data. The features including the sentiwordnet-based features, a variant of term frequency-inverse document frequency (TF-IDF) features and spam words-based features are extracted from the review data to form the feature vector. Feature fusion is performed based on the entropy of the features that are extracted. The metrics employed for the evaluation in the proposed RFATO algorithm are accuracy, sensitivity, and specificity.

By using the proposed RFATO algorithm, the evaluation metrics such as accuracy, sensitivity and specificity are maximized when compared to the existing algorithms.

The proposed RFATO algorithm is modeled by integrating the FAT algorithm in the ROA, which is used for training the deep RNN classifier for the classification of sentiments in the review data. The pre-processing is performed by the stemming and the stop word removal process for removing the redundancy for smoother processing of the data. The features including the sentiwordnet-based features, a variant of TF-IDF features and spam words-based features are extracted from the review data to form the feature vector. Feature fusion is performed based on the entropy of the features that are extracted.

To survey the approaches to design optimization based on possibility theory and evidence theory comparatively, as well as their prominent characteristics mainly for epistemic uncertainty.

Owing to uncertainties encountered in engineering design problems and limitations of the conventional probabilistic approach in handling the impreciseness of data or knowledge, the possibility‐based design optimization (PBDO), evidence‐based design optimization (EBDO) and their integrated approaches are investigated from the viewpoints of computational development and performance improvement. After that, this paper discusses the fusion technologies and an example of integrated approach in reliability to reveal the qualitative value and efficiency.

It is recognized that more conservative results are obtained with both PBDO and EBDO, which may be appropriate for design against catastrophic failure compared with the probability‐based design. Furthermore, the EBDO design may be less conservative compared with the PBDO design.

How to perfect already‐existing integration approaches in a more generally analytical framework is still an active domain of research.

The paper is a holistic reference for design engineers and industry managers.

The paper is focused on decomposition strategies and fusion technologies, especially addressing epistemic uncertainty for large‐scale and complex systems when statistical data are scarce or incomplete.

This paper aims to present a system that is fully capable of addressing the issue of detection, tracking and following pedestrians, which is a very challenging task, especially when it is considered for using in large outdoors infrastructures. Three modules, detection, tracking and following, are integrated and tested over long distances in semi-structured scenarios, where static or dynamic obstacles, including other pedestrians, can be found.

The detection is based on the probabilistic fusion of a laser scanner and a camera. The tracking module pairs observations with previously detected targets by using Kalman Filters and a Mahalanobis-distance. The following module allows to safely pursue the target by using a well-defined navigation scheme.

The system can track pedestrians from static position to 3.46 m/s (running). It handles occlusions, crossings or miss-detections, keeping track of the position even if the pedestrian is only detected in 55/per cent of the observations. Moreover, it autonomously selects and follows a target at a maximum speed of 1.46 m/s.

The main novelty of this study is the integration of the three algorithms in a fully operational system, tested in real outdoor scenarios. Furthermore, the addition of labelling to the detection algorithm allows using the full range of a single sensor while preserving the high performance of a combined detection. False-positives’ rate is reduced by handling the uncertainty level when pairing observations. The inclusion of pedestrian speed in the model speeds up and simplifies tracking process. Finally, the most suitable target is automatically selected by a scoring system.

At airport security checkpoints, baggage screening is aimed to prevent transportation of prohibited and potentially dangerous items. Observing the projection images generated by X-rays scanner is a critical method. However, when multiple objects are stacked on top of each other, distinguishing objects only by a two-dimensional picture is difficult, which prompts the demand for more precise imaging technology to be investigated for use. Reconstructing from 2D X-ray images to 3D-computed tomography (CT) volumes is a reliable solution.

To more accurately distinguish the specific contour shape of items when stacked, multi-information fusion network (MFCT-GAN) based on generative adversarial network (GAN) and U-like network (U-NET) is proposed to reconstruct from two biplanar orthogonal X-ray projections into 3D CT volumes. The authors use three modules to enhance the reconstruction qualitative and quantitative effects, compared with the original network. The skip connection modification (SCM) and multi-channels residual dense block (MRDB) enable the network to extract more feature information and learn deeper with high efficiency; the introduction of subjective loss enables the network to focus on the structural similarity (SSIM) of images during training.

On account of the fusion of multiple information, MFCT-GAN can significantly improve the value of quantitative indexes and distinguish contour explicitly between different targets. In particular, SCM enables features more reasonable and accurate when expanded into three dimensions. The appliance of MRDB can alleviate problem of slow optimization during the late training period, as well as reduce the computational cost. The introduction of subjective loss guides network to retain more high-frequency information, which makes the rendered CT volumes clearer in details.

The authors' proposed MFCT-GAN is able to restore the 3D shapes of different objects greatly based on biplanar projections. This is helpful in security check places, where X-ray images of stacked objects need to be distinguished from the presence of prohibited objects. The authors adopt three new modules, SCM, MRDB and subjective loss, as well as analyze the role the modules play in 3D reconstruction. Results show a significant improvement on the reconstruction both in objective and subjective effects.

Perception has been identified as the main cause underlying most autonomous vehicle related accidents. As the key technology in perception, deep learning (DL) based computer vision models are generally considered to be black boxes due to poor interpretability. These have exacerbated user distrust and further forestalled their widespread deployment in practical usage. This paper aims to develop explainable DL models for autonomous driving by jointly predicting potential driving actions with corresponding explanations. The explainable DL models can not only boost user trust in autonomy but also serve as a diagnostic approach to identify any model deficiencies or limitations during the system development phase.

This paper proposes an explainable end-to-end autonomous driving system based on “Transformer,” a state-of-the-art self-attention (SA) based model. The model maps visual features from images collected by onboard cameras to guide potential driving actions with corresponding explanations, and aims to achieve soft attention over the image’s global features.

The results demonstrate the efficacy of the proposed model as it exhibits superior performance (in terms of correct prediction of actions and explanations) compared to the benchmark model by a significant margin with much lower computational cost on a public data set (BDD-OIA). From the ablation studies, the proposed SA module also outperforms other attention mechanisms in feature fusion and can generate meaningful representations for downstream prediction.

In the contexts of situational awareness and driver assistance, the proposed model can perform as a driving alarm system for both human-driven vehicles and autonomous vehicles because it is capable of quickly understanding/characterizing the environment and identifying any infeasible driving actions. In addition, the extra explanation head of the proposed model provides an extra channel for sanity checks to guarantee that the model learns the ideal causal relationships. This provision is critical in the development of autonomous systems.