Catálogo de publicaciones - libros

This paper deals with the design of the capacitor insertion method for the three degree-of-freedom (DOF) flexible and deformation mechanisms aimed to eliminate hysteresis effects in piezoelectric actuators. By inserting a capacitor in series with the piezoelectric actuator is applied a 3-DOF nano-precision platform and laser-measuring systems.The Z-tilts(z, pitch, and roll motion) error compensation stage of the nano-measuring machine is accomplished. In addition, a high resolution laser interferometer is used to measure position accurately. Therefore, above the method is effectively applied to a piezoelectric actuator are presented that compensate substantial improvements in positioning control precision and control performance. With the aid of positioning control, this system provides +/-60nm positioning resolution over the total range of 1000nm and +/-0.1 arcsec angle resolution over the total range of 3 arcsec for the stage along the z-direction.

The ATIS (Advanced Traveler Information System) provides travelers with real-time and precise information about the shortest path to the destination, the traffic condition, travel time estimation, and so on. To offer these services, we have to collect the speed data which are necessary to ATIS. However many data are lost due to communication or sensor errors during collecting the data. In order to provide accurate service, the lost data have to be compensated. Thus, a lot of prediction methods have been proposed to compensate the lost speed data. In this paper, we propose new prediction method adopting the back propagation neural network under neighboring links’ speed as well as weekday and time. Experimental results show that our method reduces prediction error up to 41.8 % compared to the previous method.

Based on DCT coefficients, we propose a steganographic technique. Our method embeds a message into the DCT coefficients of an image according to the relative size of a selected DCT coefficient value and the average value of its adjacent coefficients in the block to embed and extract the hiding message bit. The coefficients are chosen middle-to-low frequency coefficients in order to defend against JPEG compression. The method decreases the likelihood of being detected, and the resulted stego-image can be stored in JPEG format. The hidden message can be securely transformed. Our experiments show that our method can extract the data efficiently and blindness.

With the growth of critical worm threats, many researchers have studied worm-related topics and internet anomalies. The researches are mainly concentrated on worm propagation and detection more than the fundamental characteristics of worms. It is very important to know worms’ characteristics because the characteristics provide basic resource for worm prevention. Unfortunately, this kind of research cases are very few until now. Moreover the existing researches only focus on understanding the function structure of the worm propagation or the taxonomy of the worm according to a sequence of worm attacks. Thus, in this paper, we try to confirm the formalized pattern of the worm action from the existing researches and analyze the report of the anti-virus companies. Finally, we define the formalized actions based on temporal behaviors and worm packet flows, and we apply our proposed method for the new worm classification.

System call sequences are useful criteria to judge the behaviors of processes. How to generate an efficient matching algorithm and how to build up an implementable system are two of the most difficult problems. In this paper, we explore the possibility of extending consecutive system call to incorporate temporal signature to the Host-based Intrusion Detection System. In this model, we use the real-time detected system call sequences and their consecutive time interval as the data source, and use temporal signature to filter the real model. During the monitoring procedure, we use data mining methods to analyze the source dynamically and implement incremental learning mechanism. Through studying small size samples and incremental learning, the detecting ability of the system can be still good when the sample’s size is small. This paper also introduces the key technologies to build such a system, and verifies this intrusion detection method in real time environment. Finally, this paper gives the experiments results to verify the availability and efficiency of our system.

Data hiding strategy based on equivalence class is proposed. We transform information hiding problem into finding the representative element in specific equivalence class. Then minimizing the distortion in the equivalence class (MDEC) is proposed, and this is used in the LSB hiding scheme. The theoretic performance of LSB hiding based on MDEC is analyzed in detail. Then a variable LSB method based on MDEC is also proposed. It can solve efficiently the problem of selecting different LSB methods to fit message with different length. Similarly, the performance is also proposed. In fact, there exists a tradeoff between distortion and length of information. However, most spatial hiding scheme based on LSB will reach larger distortion in hiding less information. The proposed hiding strategy can resolve this issue efficiently and can meet such applications where the size of message is very unstable. In addition, proposed strategy not only improves the quality of steg image but also does not sacrifice its security.

In this paper, two new properties of singular value decomposition (SVD) on images are proved. The first property demonstrates the quantitative relationship between singular values and power spectrum. The second one proves that under the condition of losing equal power spectrum, the square-error of the reconstructed image is much smaller when we reduce all singular values proportionally instead of neglect the smaller ones. Based on the two properties, a new data-hiding scheme is proposed. It performs well as for robustness, for it satisfies power-spectrum condition (PSC), and PSC-compliant watermarks are proven to be most robust. Besides, the proposed scheme has a good performance as for capacity and adaptability.

Recent advances in networking technology have increased the potential collaborations for virtual enterprises (VEs) on a global scale. Due to the dynamic nature of collaborations, building and evolving trust is essential to support the formation of VEs. There is a critical need for the new approach suited to this environment. This paper presents a trust-based approach for partner selection problem in the VEs. The proposed model explores the dynamic properties of trustworthy index, which provides a multi-perspective and interactive overview of potential partners to the decision-makers. The model uses the hybrid learning algorithm of fuzzy cognitive map (FCM). As FCM has the excellent ability in modeling complex systems, the proposed model can support historical data mining automatically, and revise the existing model according to the new requirements. Results of three experiments show that this model provides reasonable performance and high adaptability for diverse partner selection problems.

Involvement of decision support systems in the process of selecting optimal decisions in transport logistics requires preliminary estimation of cargo delivery quality. The number of delivery criteria, their ranking according to degree of importance, and correlations among these criteria play an important role in the process of decision making. For the elaboration of effective approaches to the above mentioned tasks in the transport logistics we propose use of an intelligent system that applies methods from formal concept analysis and paraconsistent annotated logic.

The reading comprehension (RC) task- accepting arbitrary text input (a story) and answering questions about it. The RC system needs to draw upon external knowledge sources to achieve deep analysis of passage sentences for answer sentence extraction. This paper proposes an approach towards RC that attempts to utilize semantic information to improve performance beyond the baseline set by the bag-of-words (BOW) approach. Our approach emphasizes matching of linguistic features (i.e. verbs, named entities and base noun phrases) and semantic extending for answer sentence extraction. The approach gave improved RC performance in the Remedia corpus, attaining accuracies of 41.3%. In particular, performance analysis shows that a relative performance of 19.7% is due to the application of linguistic feature matching and a further 10.3% is due to the semantic extending.