Please wait a minute...
, No 8

■ Cover page(PDF 3.16 MB) ■  Table of Content, August 2021  (PDF 34 KB) 

  
  • Parallel Algorithms for Earth Observation Coverage of Satellites
    Hui Li, Cunlong Zhu, Xiaodan Wang, Wenfeng Liu, and Yutao Sun
    2021, 17(8): 657-665.  doi:10.23940/ijpe.21.08.p1.657665
    Abstract    PDF (607KB)   
    References | Related Articles
    Coverage-rate computation is crucial in satellite mission planning. However, with increasingly wider mission scope and stricter demands on planning efficacy, how to improve the computing performance of coverage rate has become an important topic. In this paper, traditional grid-point approach (GPA) was used as a primary algorithm to solve satellite's earth observation coverage. To find the causes of error and propose optimization strategies, this paper also summarized grid generation and ray casting method involved in the abovementioned traditional grid-point approach. In the analysis of optimization strategies, the parallel technique was used to improve the algorithm efficiency. Based on applicable scenarios of the grid-point approach, four different sorts of grid-point generating precisions were designed, and the performance of all the parallel grid-point algorithms was tested. The test findings showed that all the parallel algorithms have significantly better performance than their serial counterparts, and the parallel optimization effect of GPU is superior to that of CPU.
    Radar Detection Performability under Graceful Degradation
    Tyler D. Ridder and Ram M. Narayanan
    2021, 17(8): 666-675.  doi:10.23940/ijpe.21.08.p2.666675
    Abstract    PDF (1147KB)   
    References | Related Articles
    Operational reliability has been introduced recently as a useful metric to describe the performability of a radar's signal processing system. In this paper, the concept of operational reliability is investigated for a typical radar scenario in which a phased array radar on an airborne platform is tasked with detecting a vehicle on the ground. Various background clutter scenarios are modeled in order to study the effects of clutter type on operational reliability and the associated optimal threshold. The effects of radar component degradation on operational reliability are investigated for two different radar system components: the receiver chain low noise amplifier and the phased array antenna. The gain degradation of the low-noise amplifier and element failures in the phased array antenna are modeled and applied to the operational reliability calculations under the various clutter scenarios. By adjusting the detection threshold adaptively, it is seen that the operational reliability can be optimally tuned and maximized even under graceful degradation conditions.
    Reliability Evaluation for a Multistate Network with Time Attribute and Periodical Maintenance
    Yi-Chun Cheng, Yi-Kuei Lin, and Ping-Chen Chang
    2021, 17(8): 676-685.  doi:10.23940/ijpe.21.08.p3.676685
    Abstract    PDF (343KB)   
    References | Related Articles
    In a hybrid flow shop (HFS), each machine at a workstation may fail due to unexpected failure, human negligence, etc. As a result, the number of available machines at a workstation should be multistate. From a decision-making viewpoint, it is a practical issue whether the HFS with a stochastic capacity could complete a given demand or not. System reliability is one of the performance indexes to assess the system performance. It is defined as a probability that the HFS with multistate capacity can satisfy demand successfully. The HFS is firstly modeled as a multistate hybrid flow shop network (MHFSN). In real-world systems, the reliability of each machine may change as time elapses. Hence, machine reliability is considered in the MHFSN, and thus it is important to evaluate time-dependent system reliability. To consider a time attribute in the MHFSN, Weibull distribution is adopted to represent each machine reliability that will degrade as time passes. To prevent decreasing machine reliability and system reliability, managers need to maintain machines periodically to retain system capability. Hence, an algorithm is proposed to evaluate system reliability with periodical maintenance. Note that system reliability is provided in a time period, and thus managers are able to make a maintenance decision in time based on the changes of the system reliability.
    Experimental Investigation and Optimization on Microstructure & Mechanical Properties of AA5052 in Comparison with AA2024 and AA8090 using Friction Stir Welding
    Y Sai Ratnakar, P Srinivasa Reddy, M Gangadhar Rao, and D Appanna
    2021, 17(8): 686-694.  doi:10.23940/ijpe.21.08.p4.686694
    Abstract    PDF (665KB)   
    References | Related Articles
    Friction Stir Welding (FSW) is a contact welding process that uses the heat generated by friction to fuse two different materials. The heat generated makes the metals attain plasticity and thus, fusion occurs which results in a joint. In the present study, welding of dissimilar aluminium alloys AA5052 with AA2024 and AA5052 with AA8090 has been done using Friction stir welding and the microstructure along with the mechanical properties of the welded joint are determined. An investigation has been done by varying FSW process parameters which include axial load, rotational speed of the spindle, and traverse speed of the weld, while keeping the tilt angle constant. Microstructures of the weld joints have been studied through SEM analysis. The weld joint between AA2024 and AA5052 welded at 710rpm has shown the highest tensile strength of 80.068 N/mm². Studies on microstructures revealed that the weld zone has better microstructure with uniform material distribution and negligible flaws were identified. The weld joint between AA5052 and AA8090 welded at 710rpm has the highest Brinell hardness of 32.7271 kgf/mm² compared to other joints. In a surface roughness check by tallysurf, the sample of AA2024+AA5052 welded at 710rpm has a smoother weld surface. Optimal results state that a higher order transformation model can improve durability and reliability of the given specimen.
    A Review on the Literature of Fashion Recommender System using Deep Learning
    Angel Arul Jothi J and Razia Sulthana A
    2021, 17(8): 695-702.  doi:10.23940/ijpe.21.08.p5.695702
    Abstract    PDF (271KB)   
    References | Related Articles
    Over the years, much research has been conducted on fashion recommendation systems. Different techniques such as image processing, machine learning, or deep learning have been incorporated in the recommendation systems. Online e-stores like Amazon, eBay, etc. customize fashion recommendation systems to satisfy the daily requirements of their customers. A number of different approaches are proposed to study the purchase pattern of the customers. This article reviews various works in fashion recommenders using deep learning that are published from 2016 to 2020. Researchers have used deep learning models distinctly or by pairing with other machine learning models in building the recommendation system. The manuscript provides a brief description of the persuading deep learning models that owns a place in recommendation systems.
    A Survey on Challenges in Transforming No-SQL Data to SQL Data and Storing in Cloud Storage based on User Requirement
    S.P. Shantharajah and E. Maruthavani
    2021, 17(8): 703-710.  doi:10.23940/ijpe.21.08.p6.703710
    Abstract    PDF (170KB)   
    References | Related Articles
    The Internet and modern computing have produced explosive growth in data volumes, but also new ways to store data. When storing petabytes of data for analysis and to gain new insight, related decisions are made about the storage and processing of big data. Organizations physically store their data on servers in their own data centers. Storing data on servers in the own data center is called on-premises or on-prem storage. In the big data world, there is a growing need in providing useful insights from different kinds of data and using data to infer vital information towards better decision making. The main challenge is to store such data with velocity into a commercial data warehouse. The effective solution for storing such data is to represent the unstructured data into structured data and maintain it using distributed clusters and cloud storage based on the user requirements. The structured data maintained in the cloud are cost efficient and infinitely scalable. To query these huge datasets from cloud storage, there is a need for distributed query engines like Hive, Impala, Presto, and Drill. These open-source Structured Query Language (SQL) engines are capable of querying enormous datasets almost instantaneously. The present work has an emphasis on distributed SQL engines like Hive and Impala that can query extremely large datasets. The focus is more on Hive and Impala which are the most widely deployed of these query engines. The outcome of the research helps the readers to understand the challenges being faced in providing and maintaining structured data in the cloud.
    Secure ECG Signal Transmission for Smart Healthcare
    Harinee S and Anand Mahendran
    2021, 17(8): 711-721.  doi:10.23940/ijpe.21.08.p7.711721
    Abstract    PDF (667KB)   
    References | Related Articles
    The health care system is overwhelming with COVID-19 patients since the end of 2019. Many countries have postponed their public screening programs and are fighting to maintain balance between essential health care and COVID-19 pandemic. At this point of time, IoT along with artificial intelligence (AI) provides a solution for a remote IoT based healthcare system. IoT based wearable devices collect data and send them to the cloud via IoT Gateway. Further, the data are analyzed by AI to raise alerts at times of emergencies like a sudden drop-in pulse rate etc. AI needs a large database for training and processing. This threatens the patient's data security and privacy. To overcome this threat, an encryption algorithm can be used. When it comes to remote healthcare, monitoring of cardio vascular muscles is the most important of all. This can be done by an electrocardiogram (ECG). These ECG signals are transmitted in case of emergency and are stored in remote servers. The ECG signals are unique in nature and they can be used in biometric devices which attract hackers. To avoid the intruders, the data is encrypted and transmitted. In this paper, we provide a lightweight cellular automaton (CA) based encryption algorithm to suit limited power and low memory devices. For ECG based evaluation, the MIT-BIH arrhythmia database is considered.
    Identifying Video Tampering using Watermarked Blockchain
    Padma Priya R, Aditya Tiwari, Ayush Pandey, and Siddharth Krishna
    2021, 17(8): 722-732.  doi:10.23940/ijpe.21.08.p8.722732
    Abstract    PDF (877KB)   
    References | Related Articles
    Nowadays, it is clearly notable that crime percentages and burglaries are predominantly increasing. Therefore, the need for video clues and evidence are becoming more important than ever before. These clues not only have a significant impact in recognizing the burglary but also aid in conviction of the individuals engaged in the foul demonstration. When such video clues have more and more significance, the likelihood of evidence getting altered is alarmingly high. Towards this end, there is a serious need to ensure the validity of crime proof. We have proposed an enhanced encrypted watermarking technique which can recognize tampering in strong video proofs. A blockchain is where the previous chain of blocks is associated with the following blocks through encryption. Any block absent in the middle of the chain can confirm that the blockchain is altered. In our proposed work, the video's frame rate and duration are first segmented into frames. Among these frames, the first frame which is obtained is merged with the chosen watermark (genesis watermark) with the help of image processing techniques, namely Logistic mapping, Integer Wavelet Transform (IWT), and singular value decomposition (SVD), respectively in a sequential manner. Such an encoded frame is chosen and duly applied as a watermark for the second frame. This procedure is repeated for all the other consecutive frames, resulting in the blockchain-like structure in the to-be authenticated video. If in any circumstances the video proof is altered (morphed or removal of frame), our proposed algorithm is capable to aid the investigators in determining not only the frames but also the timestamp in the video where its integrity was breached. We were able to attain both functional and non-functional achievements. Referring to the non-functional requirement, our watermarked video was acquiring same memory usage before and after the encryption process. Hence, our work can be considered as a lightweight technique and we were able to identify tampers in video evidence with 97% confidence interval. To the best of our knowledge, our work will be the first work wherein a blockchain-like watermarking technique is proposed for identifying breaches in video evidence. Our algorithm is applicable to video surveillance-based applications in smart cities and any other CCTV recordings.
    K-means Under-Sampling for Hypertension Prediction using NHANES Dataset
    Kajal Dwivedi, Ramanathan Lakshmanan, and Rajeshkannan Regunathan
    2021, 17(8): 733-740.  doi:10.23940/ijpe.21.08.p9.733740
    Abstract    PDF (435KB)   
    References | Related Articles
    Supervised machine learning algorithms are extensively used in various sectors to extract useful patterns from a large dataset. In real-life applications such as in the medical dataset, it is common to have less instances of a positive class compared to negative classes. This leads to the biased performance of machine learning algorithms towards the negative class and misclassifies positive class instances. However, this problem becomes worse with class-overlap. In this paper, a novel framework K-means under-sampling (KUS) is proposed to solve the class imbalance and class overlap problem together to improve the classifier's performance for the diagnosis of hypertension using the NHANES dataset. KUS improves the visibility of minority class instances to the classifiers. The results show that the KUS improved the performance of the classifiers and the results also show that KUS performed better as compared to other resampling algorithms.
Online ISSN 2993-8341
Print ISSN 0973-1318