Researching, building the routing algorithm in Zigbee networks

Zigbee wireless network built on IEEE 802.15.4 standard is becoming one of the most popular wireless networks in modern IoT devices. One of the disadvantages of Zigbee networks is the short transmission distance between devices. This paper focuses on researching and comparing routing algorithms in Zigbee networks, thereby building the optimal routing algorithm in the existing system. The paper’s objective is to form the basis for making Zigbee tree and mesh networks, which improves the transmission distance for Zigbee networks better than the star network.

Efficient collective search by BRT algorithm using fast rising threshold agent

The BRT algorithm is a method for the best-of-n problem that allows a group of distributed robots to find out the most appropriate collective option among many alternatives. Computer experiments show that the time required for finding out the best option is proportional to the number of options. In this paper, we aim to shorten this search time by introducing a few agents whose threshold increases faster than the normal one to achieve higher scalability of the BRT algorithm. The results show that the search time is reduced, and the variance is improved, especially under challenging problems where robots are required to make decisions out of a large number of options.

Constructing efficient and secure batch signature schemes

In ordinary signature schemes, such as RSA, DSA, ECDSA, the signing process is performed only for a single message. Due to performance issues, in some contexts, the above solutions will become unsuitable if a party needs to sign multiple messages simultaneously. For example, in the authenticated key exchange protocols based on signatures between client and server, the server is expected to handle multiple key exchange requests from different clients simultaneously. Batch signing is a solution that generates signatures for multi-messages simultaneously with a single (ordinary) signature generation. In this article, we will consider some of the existing batch signing solutions and point out a few of their weakness. To deal with these problems, the paper also proposes two secure types of batch signature schemes, but still ensures the same efficiency as the existing batch signing solution.

An assessment of attenuation correction of SPECT MPI images generated by deep learning model

This article evaluates the effectiveness of using a deep learning network model to generate reliable attenuation corrected the single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI). The authors collected myocardial perfusion imaging data of 88 patients from a SPECT/CT machine, with an average age of 62.47 years. Then, two datasets are created from the original data: set A includes the deep learning-based attenuation corrected images (Generated Attenuation Correction - GenAC), and the non-attenuation corrected images; set B contains only non-attenuation corrected images. These datasets were diagnosed by two doctors (in which, one has 7 years of experience and the other has 10 years of experience in reading SPECT MPI). The doctors diagnose based on the image data without knowing which dataset it belongs to. The sensitivity, specificity, diagnostic accuracy, and lesion rate were evaluated between the two data sets. Results: The average specificity, sensitivity, and accuracy of the set with the deep learning-based attenuation corrected images were 0.87, 0.86, 0.86, while the results with the non-attenuation corrected images are 0.69, 0.83, and 0.78.

The R-PSO algorithm solving multi-skill resource-constrained project scheduling problem

The Multi-Skill Resource-Constrained Project Scheduling Problem (MS-RCPSP) is a combinational optimization problem with many applications in science and practical areas. This problem aims to find out the feasible schedule for the completion of projects and workflows that is minimal duration or cost (or both of them - multi objectives). The researches show that MS-RCPSP is classified into NP-Hard classification, which could not get the optimal solution in polynomial time. Therefore, we usually use approximate methods to carry out the feasible schedule. There are many publication results for that problem based on evolutionary methods such as GA, Greedy, Ant, etc. However, the evolutionary algorithms usually have a limitation that is fallen into local extremes after a number of generations. This paper will study a new method to solve the MS-RCPSP problem based on the Particle Swarm Optimization (PSO) algorithm that is called R-PSO. The new improvement of R-PSO is re-assigning the resource to execute solution tasks. To evaluate the new algorithm's effectiveness, the paper conducts experiments on iMOPSE datasets. Experimental results on simulated data show that the proposed algorithm finds a better schedule than related works.

Computer control in human-machine interaction systems by hand movements

Computers have become popular in recent years. The forms of human-computer interaction are increasingly diverse. In many cases, controlling the computer is not only through the mouse and keyboard, but humans must control the computer through body language and representation. For some people with physical disabilities, controlling the computer through hand movements is essential to help them interact with the computer. The field of simulation also needs these interactive applications. This paper studies a solution to build a hand tracking and gesture recognition system that allows cursor movement and corresponding actions with mouse and keyboard. The research team confirms that the system works stably, accurately and can control the computer instead of a conventional mouse and keyboard through the implementation and evaluation.

An encryption - authentication algorithms developed from the elgamal cryptosystem

The paper proposes encryption - authentication algorithms developed from the Elgamal cryptosystem. There are algorithms included: system parameters, keys, encryption, and authenticated decryption. New proposed algorithms ensure a level of security against attacks: revealing secret keys - compared with RSA, GOST; security - compare with ElGamal; anti-forgery. Simultaneously verify the origin of e-doc and ensure the sender's authentication.

Researching and building a method for testing and adjusting the strapdown inertial navigation system

The strap-down inertial navigation system (SINS) is widely used and becoming very important in many areas, especially in the arms industry when the GPS signal is lost or not reliable. To ensure the precision of the system, in addition to optimizing the algorithm for the strap-down inertial navigation system, the testing, and adjusting of the SINS system when installed on vehicles also play a vital role. In the article, a method of displaying SINS data on a digital map is proposed. Furthermore, the article also proposes a method to assess the influence of misalignment angles on the navigation accuracy and how to estimate and correct them.

Research on the weapon target assignment problem in the combined air defense missile system for training simulation

In modern warfare, when the weapon system and the targets are constantly being improved and upgraded, ensuring the distribution of firepower to optimally destroy the target will help the commander to make quick and accurate decisions, thereby improving combat effectiveness. This paper proposes a method to build a command-control automatic system based on solving the weapon target assignment (WTA) problem in a combination of short and medium-range air defense missile systems so that the total damage of targets is maximum and the damage of protected area is minimum. Based on combinatorial optimization algorithms, the probability of kill, linear programming method using Hungarian algorithm, the paper presents a mathematical model of WTA and its optimal solution for short- and medium-range air defense missile systems serving the training simulation problem, thereby giving the results of evaluating the effectiveness of the algorithm.

Improving the robustness of binarized neural network using the EFAT method

In recent years with the explosion of research in artificial intelligence, deep learning models based on convolutional neural networks (CNNs) are one of the promising architectures for practical applications thanks to their reasonably good achievable accuracy. However, CNNs characterized by convolutional layers often have a large number of parameters and computational workload, leading to large energy consumption for training and network inference. The binarized neural network (BNN) model has been recently proposed to overcome that drawback. The BNNs use binary representation for the inputs and weights, which inherently reduces memory requirements and simplifies computations while still maintaining acceptable accuracy. BNN thereby is very suited for the practical realization of Edge-AI application on resource- and energy-constrained devices such as embedded or mobile devices. As CNN and BNN both compose linear transformations layers, they can be fooled by adversarial attack patterns. This topic has been actively studied recently but most of them are for CNN. In this work, we examine the impact of the adversarial attack on BNNs and propose a solution to improve the accuracy of BNN against this type of attack. Specifically, we use an Enhanced Fast Adversarial Training (EFAT) method to train the network that helps the BNN be more robust against major adversarial attack models with a very short training time. Experimental results with Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD) attack models on our trained BNN network with MNIST dataset increased accuracy from 31.34% and 0.18% to 96.96% and 85.08%, respectively.