Journal of Circuits, Systems and Computers

The development of mobile Internet, Internet of Things, and cloud computing has contributed to the unprecedented growth of information data. Big data plays a very important role in education. Currently, the literature review and in-depth research on big educational data are not very extensive, mainly involved in two fields: education mining and learning analysis. For a perfect research about education big data, this paper comprehensively reviewed three major aspects (Predictive Analytics, Learning Analytics, and Recommendation Systems) of educational data analytics for an intensive investigation and analysis: (1) Predictive Analytics: It predicts students’ learning performance by tracking students’ learning information and then analyzes students’ learning competence to build an academic early warning system; teachers can be allowed to intervene in students in time and adopts different teaching ways for different students. Therefore, both students’ learning and ability can be individualized and improved; (2) Learning Analytics: This part can identify the learners’ behavior patterns and obtain more implicit learner characteristics by studying the hidden meaning behind learning behaviors and strategies; (3) Recommendation Systems: It can match the needs of learners and recommend appropriate learning resources through different methods. All the above proved that the application of big data technology in education provides powerful data support for the development of education.

Several scheduling algorithms that have been proposed for Real-Time Operating System (RTOS) are supposed to be optimal. However, optimal scheduling is only theoretical due to the possibility of system overload where it cannot meet the deadlines of tasks. Besides, these algorithms are implemented in the RTOS, which generates additional overheads that can lead to the “nonscheduling” of certain independent tasks. In this paper, we propose an original solution for nonschedulable independent tasks in embedded systems. This solution, named Hybrid Fuzzy Earliest Deadline First Scheduling algorithm (HFEDFS), is based on the Earliest Deadline First algorithm (EDF) and Fuzzy Logic. It is characterized by a rejection policy and a rescheduling mechanism. The experimental results show that our proposed algorithm improves the system’s performance. To reduce extra overheads of RTOS, this algorithm is implemented on a Field-Programmable Gate Array (FPGA) circuit (Xilinx Virtex-5 LX50T-1156 board from DIGILENT).

This paper proposes a Finite State Machine (FSM) testing technique based on deep neural network (DNN). This technique verifies the correctness of an implementation FSM-B of a specification FSM-A. Using the back-propagation algorithm, a deep neural network is trained with the input–output patterns for a given set of transition functions that specify an FSM. Initially, for FSM-A, the input patterns and the corresponding output patterns (I/O pairs) are generated. Then most of the patterns are used to train the DNN. Once the training is over, the DNN is validated with the remaining I/O pairs (around 20%). The model can be used for verifying the correctness of FSM-B after training and validation of the DNN. Some inputs are applied to FSM-B and the generated output patterns are compared with the predicted values of the proposed DNN. The difference of accuracy percentages between FSM-A and FSM-B is recorded and zero difference between them indicates the fault-free condition of the implementation FSM-B. To check the effectiveness of the scheme, the output- and state-type faults are injected to derive mutant FSMs. Experimental results performed on the MCNC FSM benchmarks prove the efficacy of the proposed method. Only a few numbers of tests are needed to detect the presence of anomaly, if any. Hence, the test time reduces significantly — resulting in an average test time reduction of 85.67% compared to the conventional techniques. To the best of our knowledge, for the first time a DNN-driven testing scheme is being proposed.

In this research on range query technology for two-tiered wireless sensor networks, the existing researches still need improvement in privacy security and communication cost of the sensor node. In this paper, a security multi-dimensional range query protocol is proposed. The left 0-1 encoding mechanism (a new and optimized encoding mechanism), HMAC algorithm, and BASE64 encoding are employed to optimize the calculation of comparison factor. The encryption index chain is constructed by using AES algorithm to replace the long encryption constraint chain, so as to lower the energy consumption of computation and communication. The base station can verify the authenticity and integrity of query results on the basis of continuity of storage time and encryption index chain. The experiment has been conducted in the real systems. The protocol is compared and analyzed with the latest proposed Communication-Efficient Secure Range Queries (CSRQ). The results have shown that the proposed protocol can gain a better performance in terms of four aspects, i.e., the length of sensing data, the number of data collected in each cycle, the number of data dimensions and the number of sensor nodes in each unit. Meanwhile, compared with CSRQ, the communication cost of the proposed protocol is reduced by at least 10%.

Cyber-Physical System (CPS) is an integration of physical components like actuators, sensors and various types of equipment with the Internet possessing computational ability for efficient communication. A Heterogeneous Independent Network (HINT) is a realistic model that is used for the analysis of inter-dependability between the power grid and communications network. In the traditional Deep $Q$-Learning method, action needs to be stored in the $Q$ table for the prediction. In real case studies, many state and action values affect the performance of the model. Existing Deep $Q$-Network (DQN) model generates all possible actions for the $Q$-values and this involves the generation of excessive information that causes the model to overfit. In this research, the Neural Network is applied to estimate the state–action in the DQN and to store the particular state–action value instead of storing all the state–action values as followed in the traditional method. The HINT model provides realistic failure propagation in the network and its state–action value overfits the existing DQN method due to the presence of more information. The proposed DQN with reinforcement learning stores selected state–action values in the $Q$ tables and eliminates irrelevant information that helps to increase the accuracy and reduce the computational time. The DQN with reinforcement learning is applied to adaptively learn the system to select the optimal action in a continuous interaction with a stochastic environment. The proposed DQN model involves the application of reward function to store state–action value with higher probability based on prediction and eliminates other state–action values. Features such as intra-degree, inter-betweenness, substation-betweenness, relay-betweenness and feature vector are extracted and given as input to the DQN to characterize the critical nodes. The proposed DQN method is evaluated on the HINT network and synthetic network to analyze its efficiency in fault detection. The result shows that the HINT network has a lower prediction error compared to the existing Deep Neural Network (DNN) method. The proposed DQN and LSTM models have accuracies of 98% and 93% in fault prediction, respectively.

Self-supervised learning is a promising paradigm to address the problem of manual-annotation through effectively leveraging unlabeled videos. By solving self-supervised pretext tasks, powerful video representations can be discovered automatically. However, recent pretext tasks for videos rely on utilizing the temporal properties of videos, ignoring the crucial supervisory signals from the spatial subspace of videos. Therefore, we present a new self-supervised pretext task called Multi-Label Transformation Prediction (MLTP) to sufficiently utilize the spatiotemporal information in videos. In MLTP, all videos are jointly transformed by a set of geometric and color-space transformations, such as rotation, cropping, and color-channel split. We formulate the pretext as a multi-label prediction task. The 3D-CNN is trained to predict a composition of underlying transformations as multiple outputs. Thereby, transformation invariant video features can be learned in a self-supervised manner. Experimental results verify that 3D-CNNs pre-trained using MLTP yield video representations with improved generalization performance for action recognition downstream tasks on UCF101 ($+2.4\%$) and HMDB51 ($+7.8\%$) datasets.

With the development of the Internet, smart campus education and online social platforms have become the mainstream of establishing social relationships. Although many users communicate by social networks, the social networks are caught in the problem of relationship sparsity, which severely impedes users’ communication space. More fatally, in the course of building social relationships, the disclosure of sensitive information will cause users’ privacy vulnerable to be compromised by attackers. Therefore, this paper proposes a potential social relationships prediction approach based on locality-sensitive hashing (LSH) to address the above issues. Specifically, the LSH clusters similar users into the same bucket, and the fuzzy computing method is developed to predict the types of social relationships among these similar users. To further alleviate the relationship sparsity problem, the existing social network structure is utilized to predict users’ social relationships and relationship types. Furthermore, the rationality of prediction results is verified by using the social balance theory (SBT). Finally, massive experiments are executed on Epinions, and the experimental results further confirmed the efficiency and accuracy of our methodology in terms of link prediction while guaranteeing privacy-preservation.

The deterministic memory-to-cache mapping used by cache side channel attack causes the leakage of sensitive information such as secret keys, which seriously threatens user security and highlights the need to defend against this kind of attack. Recent table-based secure cache design requires more space to store the entry while a purely encryption-based design needs complex encryption units to ensure robustness. What is more, the newer attack algorithm enabling faster eviction set discovery may still break such defenses. Even though increasing the association of cache can be a potential solution, it introduces too much redundancy cache access and storage overhead. In this paper, we eliminate this problem. We present Encryption address-based Set Balance Cache (ESBC), a novel cache design to mitigate cache-based side channel attack. ESBC encrypts an address into two-level sets and displaces the data from a primary set into the secondary set when the primary set is full. The two-level mapping structure increases the complexity for attackers to build eviction sets, which is a vital step for the conflict-based attack. Furthermore, we adopt two different optimized variants from a temporal perspective, ESBC with remapping (ESBCR), which serves dynamic-remapping, and spatial perspective, ESBC with multi-mapping (ESBCM), which performs multi-mapping to improve the robustness. Our security analysis reveals that these designs can confuse the exploitation of conflicting addresses. Simulation-based evaluation on SPEC2017 shows 0.24% instruction per cycle (IPC) degradation for ESBCR with 1% remap rate. While the reduction of ESBCM that has five potential secondary sets is 0.69%. What is exciting is that both miss per kilo instructions (MPKI) and miss rates are even reduced because of the efficient cache usage. The storage overhead of these two variants is only 0.87%. By comparing the two schemes, we can observe that though ESBCR brings less performance overhead, the ESBCM has better scalability.

Sampling a signal at elevated sampling rates can be easily achieved by using time-interleaved analog-to-digital converters (TIADCs). TIADCs have more than one ADC in parallel. Each ADC samples the signal with a time gap of one sampling period and hence known as TIADC. The samples from all these ADCs are combined to reconstruct the signal. But the disadvantage of TIADCs is that they have mismatches like sampling time, gain and phase offset. The proposed work focuses on estimation and correction of these mismatches. For estimation of mismatches, teacher learner-based optimization (TLBO) algorithm was used and the estimated mismatches were used for correction by applying suitable operations. The proposed algorithm was applied for four-channel TIADCs. For estimation, a pilot signal is used which in this case is a monotonic sinusoidal signal. The estimation of mismatches was accurate and the correction was implemented for TIADCs with a sinusoidal input signal and the enhancement in signal quality was evaluated by finding signal-to-noise ratio (SNR) and signal-to-noise and distortion ratio (SNDR). There is a significant enhancement in SNR and SNDR. The average enhancements in SNR and SNDR are 50 and 46dB, respectively.

A gain and bandwidth tunable active-RC multiple-feedback (MFB) fourth-order low-pass filter is presented, which exhibits four different bandwidths of 10, 20, 30 and 40MHz and four different gain settings of 0, 4, 8 and 12dB to meet the requirements of the cellular vehicle-to-everything (C-V2X) standards. The filter uses the cascade of two biquad MFB cells. Gain and bandwidth programmability is achieved by using programmable capacitor and resistor arrays. A logic block is implemented in the filter to adjust the gain transfer function for every tuning option. Also, two-stage miller op-amp topology is chosen to implement biquad MFB cells for minimum complexity and maximum efficiency in low voltage operation. The filter is designed in 65-nm CMOS technology and occupies a 0.181mm² area and it totally consumes 13.41mW from the 1.2V supply voltage. To the best of the author’s knowledge, this work is the first CMOS baseband filter design that includes both gain and bandwidth programmability implemented for C-V2X applications.

In this paper, Pseudo-Boolean Satisfiability (PB-SAT)-based congestion reduction and overflow minimization techniques are introduced for constructing congestion-aware rectilinear Steiner trees for global routing in IC design. Since congestion in different routing channels is a big issue in routing nets, various congestion minimization approaches have been proposed by many researchers. However, the increasing complexity of integrated circuits and the reduction of their size make the congestion problem more complicated. The proposed method can produce a congestion-free routing solution by generating a suitable rectilinear Steiner tree for circuit nets. To reduce the complexity of the problem, the nets are solved individually using a Minimum Bounding Box (MBB)-based heuristic technique. The algorithm is validated by utilizing the well-known ISPD 98 global routing benchmarks. Experimental results prove that the methodology is capable of routing most of the nets within the given routing capacity satisfying all the routing parameters. The results also show that the algorithm can produce congestion-aware Steiner trees efficiently within a nominal time-bound.

Deep-learning-based object detection is widely used in unmanned aerial vehicle (UAV) systems. However, existing methods can only perform the object detection algorithm on visible images. In many scenarios, the infrared image performs better than the visible image because it can represent more invisible features. This paper proposes an object detection algorithm using image fusion to fully use the advantages of both visible light images and infrared images. Moreover, we optimize and re-design the standard object detection algorithm, YOLO V2, to improve its performance on embedded platforms. By evaluating the performance of the proposed method, we found that the recognition accuracy rate is 97.43%, while the recognition accuracy rate of the visible image is 91.09% and the recognition accuracy rate of the infrared image is 91.39%. The experimental results indicate that the proposed method can increase the recognition accuracy rate of the visible image by 6.34% and increase the recognition accuracy rate of the infrared image by 6.04%.

Applications relying on the Ethernet IEEE 802.3ba and IEEE802.3b standards require secure data encryption and authentication at extremely high speeds (at least 100 Gbit/s). The Galois/Counter Mode (GCM) is currently considered the de facto standard for hardware high-speed authenticated encryption, although other algorithms have been proposed in the literature such as the Offset Codebook Mode (OCB). The challenge in terms of providing security for high-speed applications is to achieve implementations that explore the parallelism of these algorithms; however, this translates into area cost. In this work, we propose an alternative to GCM and OCB. We show that a combination of the Offset Two-Round authenticated-encryption scheme with the AES block cipher (known as AES-OTR) is exceptionally well suited for exploiting fine-grained parallelism, and can therefore be used to achieve ultra-high-speed data encryption rates. The experiments reported in this paper show that our pipeline-parallel implementation of AES-OTR outperforms the GCM and OCB schemes in terms of throughput per area while using almost half of the logic resources. Our implementation used a Stratix 4 FPGA device as well as several devices from the Virtex family. Implementations of AES-OTR on Stratix 4 used 11kALMs and achieved a throughput of 143.65 Gbit/s. On the Virtex Ultrascale, our design used 31,859 LUTs with a throughput of 204.92 Gbit/s.

Deep State Sequence-based (DSS) equivalence checking and path-based equivalence checking have been successfully applied for verification of digital designs between System Level Model (SLM) and Register Transfer Level (RTL). The DSS-based equivalence checking method can validate designs without mapping information, but the query size for each DSS is large compared with path-based verification. The query size for path-based methods is small, but the number of comparisons is large. In this work, we combine the advantages of DSS-based methods and path-based methods. We use DSS-based methods to find the corresponding paths and use cut-points like in path-based methods to split the DSS to reduce the query size. Finite State Machine with Datapath (FSMD) is used to represent the SLM and RTL models. Experimental results demonstrate that our method can effectively validate the designs and reduce the query size for DSS-based equivalence checking method.

Application-Specific ICs (ASIC) are manufactured in bulk for a long time. In this paper, an approach to High-Speed Application-Independent IC (HS-AIIC) design is discussed. A resolution-selective (RS) and resolution-adaptive (RA) 8-bit Flash ADC are designed for use in various high-speed applications. With the choice of resolution, one can work with the trade-off between speed, power consumption, and resolution for a particular application. The proposed resolution selection algorithm can be implemented for any set of resolutions for a flash ADC design. Further, an adaptive block is added to make the ADC design adaptive in nature so that we do not have to select a particular resolution manually. The proposed design entrusts on saving manufacturing cost and increases the functionality of ADC on a single chip. Proposed resolution adaptive 8-bit flash ADC design dissipates 512mW of power with an ENOB of 7.56 bits and SNDR of 46.27dB for 1GHz sampling clock pulse.

A four-dimensional chaotic system with complex dynamical properties is constructed. The complexity of the system was evaluated by equilibrium point, Lyapunov exponential spectrum and bifurcation model. The coexistence of Lyapunov exponential spectrum and bifurcation model proves the coexistence of attractors. $C0$ and SE complexity algorithms are used to compare and analyze the corresponding complexity characteristics of the system, and the most complex integer-order system is obtained. In addition, the circuit to switch between different chaotic attractors is novel. In particular, more complex parameters are selected for the fractional-order chaotic system through the analysis of parameter complexity, and the rich dynamics of the system are analyzed. Experimental results based on Field-Programmable Gate Array (FPGA) platform verify the feasibility of the system. Finally, the most complex integer-order system is compared with its corresponding fractional-order system in image encryption, so that the fractional-order system has a better application prospect.

In this paper, four variants of Pehlivan–Uyarŏglu chaotic system (PUCS) have been proposed. The properties of the proposed PUCSs are examined through numerical simulations and the parameter values are obtained by observing the bifurcation diagrams for state variables. Further, the convergence/divergence of nearby orbits is investigated by noticing the evolution of Lyapunov exponents with time. It is found that the values of Lyapunov exponents are negative, zero and positive for all proposed variants thus confirming the chaoticity of the proposals. The strangeness of the proposed variants is also studied. The stability of PUCS and its proposed variants is examined using Jacobi stability analysis. A current feedback operational amplifier (CFOA) based circuit is put forward that can realize the existing PUCS and its proposed variants, by simply adjusting the component values. The proposed realization is compact (23% saving in overall component count) in comparison to its operational amplifier (OpAmp) based counterpart. The behavior of the proposed variants in time domain, frequency domain and phase space have been examined through simulations in LTspice design environment. Furthermore, the feasibility of the proposed variants is also discussed through presenting the electronic circuit implementation of two of the variants and the results obtained are in good agreement with the LTspice simulations. Monte Carlo (MC) simulations are also included to show the robustness of the proposed circuit against parameter variations.

Drought stress is one of the most significant abiotic stresses, adversely affecting the economy by tumbling or even eliminating agricultural productivity, development, and yield. Adverse effects of water scarcity can be reduced if certain precautionary actions could be taken in advance. Therefore, monitoring and early detection of the drought can be helpful for preparing a well-developed response plan. In response to the stresses, an intricate response system of the plants is involved which emits a range of Volatile Organic Compounds (VOCs) from different parts, such as flowers, leaves, roots and stems. These VOCs can be used as fingerprints for categorizing stressed and nonstressed plants. This paper addresses the optimization of an array of gas sensors used in an in-situ stress diagnosis system, and the data obtained after optimization have been used for the detection of induced stresses in plants by recording the VOCs emitted by the plants. The flow characteristics of the gas chamber were modeled using the Finite Element method before fabrication. The temperature modulation of the gas sensors used in the designed electronic nose system was accomplished. The optimization of the sensor array was performed using the radar plot and Wilks’ Lambda technique. The optimum operating temperatures for each gas sensor were selected using a radar plot. Furthermore, the number of the sensors in the sensor array was reduced by choosing the sensors having higher discriminant ability using the Wilks’ Lambda optimization technique. Twelve healthy Khasi Mandarin Orange saplings were considered for the investigation. Three different levels of water stresses are induced in the plants artificially for the experiments. The overall response and the optimized response of the developed electronic nose system are compared using Linear Discriminant Analysis (LDA) and bootstrap ensemble K-Nearest Neighbors (KNN) classifier. The Leaf Relative Water Content (LRWC) of the leaves is also measured concurrently to confirm the stress induction in the plants.