Large Language Models (LLMs) have recently emerged as a leading approach for multivariate time series forecasting. However, their effectiveness is hampered by a fundamental architectural mismatch: the permutation-invariant self-attention of Transformers lacks inductive biases for the strict temporal order and complex cross-variable dependencies inherent in time series. Existing methods often sidestep this issue with input-level alignment techniques rather than endowing the model itself with structural awareness. To address this gap, we introduce GraFT (Graph-infused Forecasting Transformer), a framework that systematically embeds relational priors into a pre-trained backbone by constructing a heterogeneous patch relation graph, which represents both universal temporal principles with static edges and instance-specific patterns with dynamic adaptive edges. To process this multi-relational structure, a relational graph convolutional network generates structure-aware representations, which are infused into the patch embeddings to provide explicit structural guidance to the Transformer's attention mechanism. Extensive experiments show that GraFT achieves state-of-the-art performance on long-term forecasting and zero-shot learning, outperforming leading LLM-based methods on eight standard benchmarks with an average Mean Squared Error (MSE) reduction of 14.4%.

Active learning effectively reduces annotation costs while enhancing model performance in medical image segmentation tasks. One-shot active learning presents a highly practical scenario where valuable samples for annotation are selected in a single round. However, current one-shot active learning methods predominantly rely on sampling selection methods based on global information. In contrast, focusing on the selection of features specific to local lesion regions would be more targeted and effective. In this work, we introduce a novel deep active learning framework specifically designed for pathological lesion segmentation tasks. To enable the model to effectively capture lesion-related regions of interest, we propose a strong augmentation strategy for image samples in self-supervised contrastive training. These strong augmentation samples are generated through cluster-based background subtraction using cluster projector, thereby emphasizing the features of the target lesion and improving the model's sensitivity to these areas. We utilize the Segment Anything Model as the base model to facilitate training and sample selection in a one-shot manner. Experimental results on two lesion segmentation datasets demonstrate that the proposed framework outperforms several existing active learning methods.

The Segment Anything Model (SAM), serving as a foundational vision model, has demonstrated an extraordinary capability in segmenting natural images. However, its efficacy in the domain of medical image analysis leaves much to be desired, primarily due to the irregular shapes and indistinct edges characteristic of lesions. There is a pressing need to augment SAM’s proficiency in recognizing lesion boundaries. Achieving precise segmentation of such lesions requires a blend of high-level global semantic information and low-level local boundary details. In response to this challenge, we introduce an auxiliary boundary-aware Convolutional Neural Network (CNN) module, equipped with a boundary generator, to enhance the model’s focus on boundary feature extraction. Furthermore, to leverage both the intricate low-level features in the lower layers and the high-level textural features in the deeper layers, we employ feature adapters to fuse the multi-scale features derived from the SAM encoder, thereby aggregating a wealth of enriched information. The performance superiority of our model is demonstrated through comprehensive evaluation on three different medical image segmentation tasks, and experimental results highlight the effectiveness of our proposed model.

The precise detection of mild cognitive impairment (MCI) is of significant importance in preventing the deterioration of patients in a timely manner. Although hypergraphs have enhanced performance by learning and analyzing brain networks, they often only depend on vector distances between features at a single scale to infer interactions. In this paper, we deal with a more arduous challenge, hypergraph modelling with synchronization between brain regions, and design a novel framework, i.e., A Multi-scale Hypergraph Network for MCI Detection via Synchronous and Attentive Fusion (MHSA), to tackle this challenge. Specifically, our approach employs the Phase-Locking Value (PLV) to calculate the phase synchronization relationship in the spectrum domain of regions of interest (ROIs) and designs a multi-scale feature fusion mechanism to integrate dynamic connectivity features of functional magnetic resonance imaging (fMRI) from both the temporal and spectrum domains. To evaluate and op-timize the direct contribution of each ROI to phase synchronization in the temporal domain, we structure the PLV coefficients dynamically adjust strategy, and the dynamic hypergraph is modelled based on a comprehensive temporal-spectrum fusion matrix. Experiments on the real-world dataset indicate the effectiveness of our strategy. The code is available at https://github.com/Jia-Weiming/MHSA.

Unmanned Aerial Vehicles (UAVs) offer a solution for remote inspection of wind turbines. However, in stormy weather conditions, the visual quality of UAV-taken images is affected by contaminated suspended atmospheric particles. To address this problem, a double-patch lightweight convolutional dehazing neural network (DPLDN) is proposed to reconstruct hazy images and enhance the image quality. Unlike other learning-based methods that measure transmission map and atmospheric light separately, the proposed DPLDN uses a transformed atmospheric scattering model to jointly transmission map and atmospheric light, employs depth-separable convolution instead of conventional convolution, and splits the image into double patches. In addition, a super-resolution reconstruction method is proposed to transform the processed low-resolution images into higher-quality images. Extensive experiments shows that our proposed method has better dehazing performance compared to other state-of-the-art image dehazing techniques. Meanwhile, the applicability of the method in wind turbine blade image segmentation is experimentally verified.

Malware proliferation is a major threat to computer systems, and malware classification techniques are effective for analyzing and identifying malware. Recent intelligent malware classifiers intend to integrate natural language processing techniques for better identification performance, however, representation learning presents a fundamental challenge in this new paradigm. Currently, representation models either rely on predetermined structures or ignore structure. Thus, we propose a malware vector representation model utilizing an improved context-separated bi-directional long short-term memory (CS-Bi-LSTM) network with reinforcement learning (RL) to discover optimized structures for learning sentence representations. Our model filters out irrelevant representations and captures long-term dependencies. To generate word vectors, we use the CS-Bi-LSTM network, which employs two unique unidirectional long short-term memory (LSTM) cells for each context. Following that, we develop a proximal policy optimization (PPO)-based RL structure to capture relevant representations with feedback from a quality estimator. Through intrinsic and extrinsic evaluations on two malware datasets, our proposed model demonstrates better performance by filtering redundant information and achieving an acceptable vector representation. Then, our proposed method achieves the highest predictive performance with a classification accuracy of 98.54%.

Named entity recognition (NER) as a crucial technology is widely used in many application scenarios, including information extraction, information retrieval, text summarization, and machine translation assisted in AI-based smart communication and networking systems. As people pay more and more attention to NER, it has gradually become an independent and important research field. Currently, most of the NER models need to manually adjust their hyper-parameters, which is not only time-consuming and laborious, but also easy to fall into a local optimal situation. To deal with such problem, this paper proposes a machine learning-guided model to achieve NER, where the hyper-parameters of model are automatically adjusted to improve the computational performance. Specifically, the proposed model is implemented by using bi-directional encoder representation from transformers (BERT) and conditional random field (CRF). Meanwhile, the collaborative computing paradigm is also fused in the model, while utilizing the particle swarm optimization (PSO) to automatically search for the best value of hyper-parameters in a collaborative way. The experimental results demonstrate the satisfactory performance of our proposed model.

Abstract Recent developments in the field of Internet of things (IoT) have aroused growing attention to the security of smart devices. Specifically, there is an increasing number of malicious software (Malware) on IoT systems. Nowadays, researchers have made many efforts concerning supervised machine learning methods to identify malicious attacks. High‐quality labels are of great importance for supervised machine learning, but noises widely exist due to the non‐deterministic production environment. Therefore, learning from noisy labels is significant for machine learning‐enabled Malware identification. In this study, motivated by the symmetric cross entropy with satisfactory noise robustness, the authors propose a robust Malware identification method using temporal convolutional network (TCN). Moreover, word embedding techniques are generally utilised to understand the contextual relationship between the input operation code (opcode) and application programming interface function names. Here, considering the numerous unlabelled samples in real‐world intelligent environments, the authors pre‐train the TCN model on an unlabelled set using a word embedding method, that is, Word2Vec. In the experiments, the proposed method is compared with several traditional statistical methods and more recent neural networks on a synthetic Malware dataset and a real‐world dataset. The performance comparisons demonstrate the better performance and noise robustness of their proposed method, especially that the proposed method can yield the best identification accuracy of 98.75% in real‐world scenarios.

To address foreseeable challenges during the penetration of wind energy into the power grid, including accurate wind power forecasting and smart power generation scheduling, this study proposes a novel short-term wind speed forecasting model, named EMD-KM-SXL, which is based on empirical mode decomposition (EMD),$K $-means clustering and machine learning, and a new two-stage short-term wind power forecasting model based on wind speed forecasting and wind power curve (WPC) modeling. The former wind speed forecasting model regards historical wind speed observations as model input and the latter power forecasting model utilizes knowledge augmentation, introducing wind power conversion relationship, environmental factors as well as wind power system status parameters. In the proposed wind speed forecasting model, three machine learning models, including support vector regressor, XGBoost regressor, and Lasso regressor, are employed to forecast three types of frequency components that generated via EMD and$K $-means clustering. Then, the WPC model is utilized to compute potential outpower based on wind speed prediction value speed, which is regarded as the first stage of the proposed wind power forecasting model. In the second stage, environmental factors and wind power system status parameters are introduced and an artificial neural network model, considering preliminary predicted power, environmental factors, and wind power system status parameters as model input is built to make final power prediction. Computational results show that the proposed models achieve the best performance in terms of wind speed and power forecasting on different forecasting horizons ranging from 10 to 40 min, compared with benchmarking methods.

Coronavirus disease-2019 (COVID-19) as a new pneumonia which is extremely infectious, the classification of this coronavirus is essential to effectively control the development of the epidemic. Pathological changes in the chest computed tomography (CT) scans are often used as one of the diagnostic criteria of COVID-19. Meanwhile, deep learning-based transfer learning is currently an effective strategy for computer-aided diagnosis (CAD). To further improve the performance of deep transfer learning model used for COVID-19 classification with CT images, in this article, we propose a hybrid model combined with a semi-supervised domain adaption model and extreme learning machine (ELM) classifier, and the application of a novel multikernel correntropy induced loss function in transfer learning is also presented. The proposed model is evaluated on open-source datasets. The experimental results are compared to some baseline models to verify the effectiveness, while adopting accuracy, precision, recall,$F_{1}$score and area under curve (AUC) as the evaluation metrics. Experimental results show that the proposed method improves the performance of original model and is more suitable for CT images analysis.

In recent years, joint detection and embedding (JDE) has become the research focus in multi-object tracking (MOT) due to its fast inference speed. JDE models are designed and widely utilized to train the detection task and the re-identification (Re-ID) task jointly. However, there exists a severe issue overlooked by previous JDE models, i.e., the detection task requires category-level features but the Re-ID task requires instance-level features. This could lead to feature conflict, which would hurt the performance of JDE models. Furthermore, inaccurate detection results can degrade the final tracking accuracy even when discriminative Re-ID features are provided. In this article, we propose a new balancing method for training JDE models, which monitors the training process of the detection task and adjusts the weights of the detection task and Re-ID task in the training phase. Our proposed balancing method ensures a well-trained detection model and a good trade-off between the detection task and Re-ID task. Comprehensive experiments on two public MOT benchmarks demonstrate the effectiveness and superiority of our proposed balancing method. In particular, our proposed balancing method could achieve new state-of-the-art results on MOT challenges without additional training data.

With the rapid development of internet-of-things (IoT) and communication technologies, the quality of our daily life has improved with the applications of smart communications and networking, such as intelligent transportation, mobile computing, and edge computing. How to enable such a smart life has become a popular research topic. 5G-powered communication supports massive data transmission ensuring mobile users to have high-quality experiences, which bridges the gap between IoT and cloud computing. For example, connected 4K cameras can handle object tracking using functions in the cloud computing platform via the support of smart communications. Moreover, in smart communications and networking, we can collect, measure and analyse vast volumes of data using the technologies of artificial intelligence and big data. Such an advantage will bring tremendous opportunities for smart cities, such as unmanned vehicles and smart transportation. However, there are also many issues to resolve as we need to study architecture, applications, and future challenges within smart communications and networking. We accepted 19 papers for publication in this special issue after peer review. These selected papers are categorised into three topics: Topic A (Optimisation in smart communications), Topic B (Networks security and optimisation) and Topic C (Smart technology in IoT). The summary of each topic is given below. Wu et al., in their paper ‘Completion time minimisation for UAV enabled data collection with communication link constrained’, study the completion time minimisation problem under the communication link contained for data collection via designing the safe flight trajectory of the UAV in a complex environment. The authors first transform the original problem to a TSP-like problem based on the hover point, which can satisfy the link constraints of data collection. Then A* algorithm and SCA algorithm are used to construct the adjacency matrix and, respectively, the classical DP is used to solve the TSP-like problem. Besides, the slack variables are introduced and the successive convex approximation is leveraged to reformulate the communication link constraint, obstacle avoidance constraints, and discrete region threat avoidance constraints. Compared with the TSP-like problem with hovering, a continuously flying UAV usually has less time to perform a mission. The simulation results are presented to verify the proposed two-path planning algorithms under various parameter configurations. Zhou et al., in their paper ‘User-centric data communication service strategy for 5G vehicular networks’, develop a UCDCS strategy and the ARSUGs are updated in real-time according to predictions of vehicle mobility. Then, after the vehicle sends out a data communication request, the network comprehensively considers the RSU load cost, throughput cost, and vehicle income to flexibly allocate vehicle service resources via ARSUGs. Finally, in the allocated ARSUG, the network sorts and scores the ARSU in the ARSUG according to the communication service preferences of different vehicles and assists the vehicles to select the best ARSU for data transmission. The simulation results show that, compared with the traditional IMM, NCNS, and THOM strategies, the downlink transmission rate, link reliability, and network delay of the UCDCS strategy are 47.74%, 0.21%, and 5.96% higher, respectively. The experimental results verify that the strategy proposed in this paper can achieve better network load balancing than previous strategies. Hu et al., in their paper ‘Orthogonal frequency division multiplexing with cascade index modulation’, propose a novel orthogonal frequency division multiplexing with cascade index modulation (OFDM-CIM), which combines the conventional IM with the multiple-mode IM together, to increase the proportion of the index bits in the transmission. Subcarrier-wise and sub-block-wise cascade IM schemes are proposed to achieve different spectral efficiency and diversity order for diverse scenarios in the next-generation wireless communications. The optimal maximum likelihood (ML) detector is proposed for OFDM-CIM. To reduce the demodulation complexity, a novel tree search-based detector and a log-likelihood ratio (LLR) based low complexity detector, which can avoid the illegal indices patterns in the searching process, are proposed for OFDM-CIM. Monte Carlo simulations show that the proposed scheme achieves better BER performance than OFDM-IM. Tong et al., in their paper ‘Low pilot overhead channel estimation for CP-OFDM-based massive MIMO OTFS system’, first analyse the CP-OFDM-based massive MIMO OTFS system channel with antenna directivity pattern and transform the burst sparsity in the angle domain into block sparsity by using non-uniform Fourier Transform (NUFT). Furthermore, according to the general sparsity in the delay domain, the block sparsity in the Doppler domain, and the angle domain, a three-dimensional dynamic support search (DSD) algorithm is proposed. Compared with the traditional OMP algorithm and the 3DSOMP algorithm, simulation results demonstrate the proposed DSD algorithm has higher channel estimation accuracy and lower pilot overhead. Gao et al., in their paper ‘Low drift visual inertial odometry with UWB aided for indoor localisation’, propose a low drift visual inertial odometry with ultra-wideband (UWB) aided for indoor localisation. First, a single UWB anchor was dropped in an unknown position, and a cost function was formed by the position information and the UWB ranging information to obtain the position of the anchor. Then, the single anchor position and the UWB ranging constraints were added to the tightly coupled visual inertial fusion algorithm framework, thereby improving the robustness of motion tracking and reducing the drift of the odometry. Finally, the effectiveness of the proposed method was verified in the actual indoor environment, and the experiment results demonstrated that, compared with state-of-the-art localisation methods, the positioning accuracy and robustness were improved significantly. Wang et al., in their paper ‘An approach to adaptive filtering with variable step size based on geometric algebra’, propose the novel approach to adaptive filtering with variable step size based on Sigmoid function and geometric algebra (GA). First, the proposed approach to adaptive filtering with variable step size based on geometric algebra represents the multi-dimensional signal as a GA multi-vector for the vectorisation process. Second, the proposed approach to adaptive filtering with variable step size based on geometric algebra solves the contradiction between the steady-state error and the convergence rate by establishing a nonlinear function relationship between the step size and the error signal. Finally, the experimental results demonstrate that the proposed approach to adaptive filtering with variable step size based on geometric algebra achieves better performance than that of the existing adaptive filtering algorithms. Zheng, et al., in their paper ‘Unequal Error Protection Transmission for Federated Learning’, design an unequal error protection (UEP) scheme based on multi-rate channel coding and multi-layer modulation. The numerical simulation verifies that the proposed UEP transmission schemes have significant benefits in accuracy, robustness and efficiency, especially when the channel condition is poor. Zhang et al., in their paper ‘Secrecy outage probability analysis of energy-aware relay selection for energy-harvesting cooperative systems’, employ relay selection to improve the physical-layer security for a CCR-EH system consisting of a CS, multiple CRs and a CD in the face of an E. To prevent confidential information leaking to E, an optimal relay selection (ORS) scheme and a suboptimal relay selection (SRS) scheme are proposed. In the ORS scheme, the whole channels state information (CSI) of wireless links is available to CRs while SRS only needs to know the CSI of main channels from CRs to CD. Moreover, the closed-form expressions of secrecy outage probabilities for both ORS and SRS schemes are derived. The classical round-robin relay selection (RRRS) is also analysed in terms of secrecy outage probability. Finally, the numerical results show that ORS achieves the best performance and RRRS performs the worst in terms of secrecy outage probability. Wang et al., in their paper ‘Applying an auction optimisation algorithm to mobile edge computing for security’, study the mobile blockchain network based on edge computing and propose a new assumption regarding the mobile communication blockchain based on the traditional blockchain. By analysing attacks on the mobile blockchain, a security model based on edge computing is designed, and the smart contract in the blockchain is combined with a court trial. In the algorithm optimisation process, a price utility function is constructed based on maximising social welfare, and both models are used as joint optimisation indexes. The profit of the provider is guaranteed, which is conducive to the development of the blockchain. Simulation results verify that system security increases with the blocked funds and duration, and the forking attack success rate approaches zero as the number of validators increases. Zhang et al., in their paper ‘A PUF-based lightweight authentication and key agreement protocol for smart UAV networks’, propose a two-stage lightweight identity authentication and key agreement protocol for UAV. The entire process only uses hash and XOR operations, which significantly improves the authentication efficiency. Simultaneously, the physical unclonable function (PUF) is introduced and embedded into the UAV hardware to ensure UAV network communication security when a UAV suffers a physically capture attack. Moreover, the security of the proposed protocol is proved with Burrows–Abadi–Needham (BAN) logic, real-or-random (ROR) model, and AVISPA simulation tools. An informal security analysis is also provided to illustrate that the protocol satisfies the security requirements of UAV networks. Finally, the protocol is compared with other existing protocols regarding function properties, computation cost, and communication cost. The results show that the protocol has effectiveness and practicality. Akhunzada et al., in their paper ‘MalDroid: Secure DL-enabled intelligent malware detection framework’, present a secure by design efficient and intelligent Android detection framework against prevalent, sophisticated and persistent malware threats and attacks. A novel and highly proficient CUDA-enabled multi-class malware threat detection and identification deep learning (DL)-driven mechanism that leverages ConvLSTM2D and CNN is proposed. The devised approach is extensively evaluated on publicly available state-of-the-art datasets of Android applications (i.e., Android Malware Dataset (AMD), Androzoo). Standard and extended assessment metrics are employed to thoroughly evaluate the proposed technique. Moreover, the performance of the proposed algorithm is verified both with the constructed hybrid DL-driven algorithms and current benchmarks. Additionally, to explicitly show unbiased results, the proposed scheme is validated. Shang et al., in their paper ‘An efficient MAC protocol design for adaptive compressed sensing based underwater WSNs’, design an adaptive compressive sensing-based MAC protocol to optimise energy efficiency and bandwidth utilisation. In the feedback UWSN structure, based on the adaptive compressive sensing method, a TDMA mechanism is designed to collect data from both the compressive sensor nodes and non-compressive sensing nodes in the UWSN. Super-frame-based MAC protocol is designed to minimise the energy consumption per bit according to the designed UWSN. An optimisation problem is to solve the parameters of the super-frame to satisfy both data latency and recovery quality requests. Considering the compression sensing method and packets loss, the slot allocation algorithm is designed to maximise bandwidth utilisation. Simulations show that the proposed method performs better than most of the state-of-art protocols and also a testbed is built up to show that the battery life can be prolonged by 11%. Liu, et al., in their paper ‘Reliability Modelling and Optimization for Microservicebased Cloud Application Using Multi-agent System’, proposes a scheduling scheme of Multi-agent system to optimize the reliability of cloud applications through flexible resource combination. The reliability optimization (PCPRO) algorithm based on partial critical path is introduced. Experiments on scientific workflow verify the effectiveness of the proposed algorithm. Ai et al., in their paper ‘Anti-collision algorithm based on slotted random regressive-style binary search tree in RFID technology’, propose an anti-collision algorithm based on slotted random regressive-style binary search tree (SR-RBST). Based on slotted ALOHA (SA), the method proposed in this paper uses the regressive-style binary search tree (RBST) to process the RFID labels in the collision time slot. With the same size of tags, the SR-RBST algorithm needs less total time slot and has higher efficiency and shorter identification time, while with the increase of the number of tags, the SR-RBST anti-collision algorithm has more obvious advantages. The SR-RBST algorithm effectively improves the time slot utilisation efficiency of the system. Siddiqi et al., in their paper ‘FANET: Smart city mobility off to a flying start with self-organised drone-based networks’, propose a reliable RTA monitoring scheme using enhanced ant colony optimisation (eACO) technique based on self-organised drone FANETs. The proposed scheme addressed several challenges including coverage of larger geographical areas and data communication links between FANETs nodes. The experiment results are presented to compare the proposed technique against different network lifetimes and the number of received packets. The presented results show that the proposed techniques perform better compared to other state-of-the-art techniques. Guo et al., in their paper ‘Payoff-maximisation-based adaptive hierarchical wireless charging algorithm for the mobile charger in IoT’, propose a payoff-maximisation-based adaptive hierarchical wireless charging algorithm for the mobile charger. According to energy allocation, anchor point deployment, and time allocation, decomposing it into three layers by the hierarchical decomposition method to obtain optimal solution quickly. The process of energy allocation and anchor point deployment in each mesh is optimised in the first two layers based on Karush–Kuhn–Trucker (KKT) condition and greedy strategy respectively. Based on the feedback of the first two layers, the most complex problem of time allocation in the last layer is solved by the innovative gain recall mechanism. The trade-off between the number of recharged devices and recharging time in each cycle can be achieved by only charging the devices in the meshes which are without recall gains. The simulation results prove our algorithm can adaptively adjust the ratio of moving time to recharge time in a fixed cycle, and mobile chargers can always work in efficient recharging positions, whose effect is exploited at the utmost. Wang et al., in their paper ‘Improving the performance of tasks offloading for internet of vehicles via deep reinforcement learning methods’, propose an offloading scheme combining mobile edge computing (MEC) and deep reinforcement learning (DRL). First, a realistic map is simulated, while initialising the tasks queue, and building a task offloading environment with the base station (BS), roadside units (RSUs), and idle vehicles. Then, an algorithm that combines deep learning with reinforcement learning, i.e., the deep Q-learning network (DQN) algorithm, is developed to optimise the offloading scheme, to further reduce the offload latency. Finally, given that the complete information cannot be observed effectively in the environment, the long short term memory (LSTM) model is applied to train neural networks within DQN to improve its learning efficiency, in consideration of the satisfactory performance of LSTM in processing time-series data. The simulation results show that the MEC-based vehicle task offloading can effectively reduce the latency of vehicle offloading. Hou et al., in their paper ‘A data-driven method to predict service level for call centers’, investigate how to use the data-driven method to solve the service level prediction problem. To solve this problem, the relationship between service level and other factors, such as number of calls, number of agents, and time is explored. To model the relationship between service level and input features, some features based on empirical analyses are extracted and proposed to use decision tree-based ensemble methods, like random forest and GBDT. The experiment results show that the proposed method outperforms other baselines significantly. Wang, et al., in their paper ‘Short-term Passenger Flow Forecasting Using CEEMDAN meshed CNN-LSTM-Attention Model Under Wireless Sensor Network’, propose a complete ensemble empirical mode decomposition with adaptive noise algorithm (CEEMDAN) and attention-based CNN-LSTM network to extract both temporal and spatial characteristics of passenger flow data. By adding the attention mechanism, the problem of insufficient peak value prediction can be solved effectively. The experiment result shows that the CEEMDAN-ConvLSTM-Attention model has a significant performance improvement than the existing network models. All of the papers selected for this special issue show the development of different emerging technologies and creative strategies in various fields. However, there are still many challenges in all of those fields that require future research attention. The authors have no conflict of interest to disclose. Honghao Gao is currently with the School of Computer Engineering and Science, Shanghai University, China. He is also a Professor at Gachon University, South Korea. Prior to that, he was a research fellow with the Software Engineering Information Technology Institute at Central Michigan University, USA, and was an adjunct professor at Hangzhou Dianzi University, China. His research interests include software formal verification, industrial IoT networks, vehicle communication, and intelligent medical image processing. He has publications in IEEE TII, IEEE T-ITS, IEEE TNNLS, IEEE TSC, IEEE TNSE, IEEE TNSM, IEEE TCCN, IEEETGCN, IEEE TCSS, IEEE TETCI, IEEE/ACM TCBB, IEEE IoT-J, IEEE JBHI, IEEE Network, ACM TOIT, ACM TOMM, ACM TOSN, ACM TMIS. He is the recipient of the Best Paper Award at IEEE TII 2020 and EAI CollaborateCom 2020. Xiong Luo currently works at the Department of Computer Science and Technology, University of Science and Technology Beijing. His main research interests include data mining and machine learning, complex system modelling and computational intelligence, cognitive neural networks, intelligent optimal control, Internet of Things applications. He is a senior member of IEEE, a senior member of The Chinese Computer Society, a member of the Intelligent Automation Committee of the Chinese Association of Automation, a deputy secretary general of the Intelligent Medical Committee of the Chinese Association for Artificial Intelligence, and a member of the Cognitive System and Information Processing Committee of the Chinese Association for Artificial Intelligence. Ramón J. Durán Barroso received ‘a telecommunication’ engineer degree in 2002 and obtained his PhD in 2008, both from Universidad de Valladolid (UVa), Spain. From 2002 to 2010, he was an Assistant Professor at Universidad de Valladolid, Spain. From 2010 to the present, he was an associate professor at Universidad de Valladolid, Spain. From 2004, he focused on communication networks, in particular in the following topics: design and optimisation of wavelength routed optical networks, hybrid optical networks (proposing polymorphic networks), cognitive heterogeneous optical networks (proposing CHRON networks) and access optical networks. He has also actively researched the use of ICT technologies in education. Moreover, he has actively collaborated with the other line of his research group devoted to research about wireless communications and location techniques using some methodologies previously used in his research in network optimisation. Walayat Hussain received a PhD from the University of Technology Sydney, Australia. Currently, he is serving as a lecturer (assistant professor) at Victoria University, Melbourne, Australia. Before joining Victoria University, he worked for six years as a lecturer and research ‘fellow’ at the FEIT, University of Technology Sydney, Australia. He has served as an assistant professor at BUITEMS University for many years. He has published in various top-ranked ERA-A*, JCR/SJR Q1 journals such as The Computer Journal, Info. Systems, Info. Sciences, IJIS, FGCS, IEEE Access, Comput & Ind Eng, MONET, Journal of AIHC, IEEE TETCI, IEEE TSC, IEEE TGCN, IJCS and WCMC. He has served as a guest editor in various Q1 journals. He has won multiple national and international research awards and recognitions. He is the recipient of the Best Paper Award at 3PGCIC 2015, 2016 Poland, South Korea, Ministry of Higher Education Govt. of Oman and FEIT HDR Publication Award by the UTS Australia. Guest Editorial.

Abstract With the rapid development of communication technologies, the quality of our daily life has been improved with the applications of smart communications and networking, such as intelligent transportation and mobile service computing. However, high user demands for quality of service (QoS) are forcing intelligent transportation to continuously improve immediacy and reduce the tasks offloading delay for the internet of vehicles (IoV). To meet the low latency of vehicle tasks offloading, an offloading scheme combining mobile edge computing (MEC) and deep reinforcement learning (DRL), is proposed in this article. Firstly, a realistic map is simulated, while initializing the tasks queue and building a tasks offloading environment with multiple service nodes. Then, an algorithm that combines deep learning with reinforcement learning, that is, the deep Q‐learning network (DQN) algorithm, is developed to optimize the offloading scheme by reducing the offload latency. Finally, given that the complete information cannot be observed effectively in the environment, a long short‐term memory (LSTM) model is applied within the DQN to train its neural network to improve offloading efficiency. The simulation results show that the MEC‐based vehicle tasks offloading can effectively reduce the latency of vehicle offloading.

This paper investigates the multi-objective optimal operation problem of an integrated energy system (IES) which integrates grid-connected photovoltaic (PV) generator, gas boiler, battery energy storage system, and thermal storage to satisfy energy demand in forms of electricity and heat. To handle the changes from the system uncertainty (e.g., PV generation, electrical loads, thermal loads, etc.) and unknown thermal dynamic model for temperature control, deep reinforcement learning-based model-free optimization method is proposed to solve the multi-objective optimization problem in which the multi-objective optimization problem is firstly formulated as a multi-objective Markov decision process (MDP) problem. The multi-objective MDP problem is converted to many single-objective MDP problems by the sum technique which are solved by multi-agent deep deterministic policy gradient (DDPG) algorithm. To improve the performance of multi-agent DDPG algorithm, evolutionary computing-based parameter-tuning method is further proposed to fine-tune the policy parameters in DDPG algorithm. The proposed methods are verified on real data. Experiments results illustrate that the multi-agent DDPG algorithm can efficiently solve the multi-objective optimal operation problem of the IES while the evolutionary computing-based policy parameter-tuning method can further improve the approximation of Pareto frontier.

In this paper, we present a study on person identification using static features extracted from Kinect skeleton data. On the contrary to previous reports that the dynamic features such as gait parameters are more discriminative than static features, we find that by using a combination of a set of easy to obtain static features, we can achieve nearly perfect accuracy in identifying persons with only a few frames. In our study, we experimented with several classifiers, including k-nearest neighbor (KNN), decision tree, Gaussian Naive Bayesian, neural network with multiplayer perception (MLP), and support vector machine (SVM), and several combinations of static skeleton features. In all scenarios, KNN outperforms other classifiers consistently. MLP and SVM require a huge amount of parameter tuning and training time and they do not perform well compared with KNN except for small gallery sizes when all static features available.

This article proposes a hybrid symbolic aggregate approximation and vector space model (SAX-VSM) method for automatically classifying soil water content cycles. In the proposed method, a novel similarity measure, the distance weighted cosine (DWC) similarity measure, is introduced to improve the classification performance of the SAX-VSM. The DWC similarity measure incorporates both direction and distance information of feature vectors. Meanwhile, a mixed-integer optimization problem is formulated to determine hyperparameters. An extended Rao-1 algorithm, I-Rao-1 algorithm, is developed to solve such optimization problems. To verify the feasibility and effectiveness of the proposed method, three soil moisture data sets collected from the Florida research trials are employed. Compared with state-of-the-art methods, the proposed method has achieved the best performance based on all data sets in terms of the highest accuracy, precision, and recall values. Therefore, it is promising to apply the proposed method into real applications in the smart irrigation system.

In this article, we provide a concise but systematic review on blockchain-enabled cyber-physical systems (CPS). We dissect various blockchain-enabled CPS as reported in the literature in terms of their operations and the features of blockchain that have been used. We identify key common CPS operations that can be enabled by blockchain, and classify them in terms of their time sensitivity and throughput requirements. We also elaborate and classify features of blockchain in terms of different levels of benefits to CPS, including security, privacy, immutability, fault tolerance, interoperability, data provenance, atomicity, automation, data/service sharing, and trust. Finally, we point out two primary open research issues for developing blockchain-enabled CPS, namely, excessive delay in reaching consensus and limited throughput, and outline future research directions.

With the popularization of computer-aided diagnosis (CAD) technologies, more and more deep learning methods are developed to facilitate the detection of ophthalmic diseases. In this article, the deep learning-based detections for some common eye diseases, including cataract, glaucoma, and age-related macular degeneration (AMD), are analyzed. Generally speaking, morphological change in retina reveals the presence of eye disease. Then, while using some existing deep learning methods to achieve this analysis task, the satisfactory performance may not be given, since fundus images usually suffer from the impact of data imbalance and outliers. It is, therefore, expected that with the exploration of effective and robust deep learning algorithms, the detection performance could be further improved. Here, we propose a deep learning model combined with a novel mixture loss function to automatically detect eye diseases, through the analysis of retinal fundus color images. Specifically, given the good generalization and robustness of focal loss and correntropy-induced loss functions in addressing complex dataset with class imbalance and outliers, we present a mixture of those two losses in deep neural network model to improve the recognition performance of classifier for biomedical data. The proposed model is evaluated on a real-life ophthalmic dataset. Meanwhile, the performance of deep learning model with our proposed loss function is compared with the baseline models, while adopting accuracy, sensitivity, specificity, Kappa, and area under the receiver operating characteristic curve (AUC) as the evaluation metrics. The experimental results verify the effectiveness and robustness of the proposed algorithm.

This article solves the exponential synchronization issue of memristor-based complex-valued neural networks (MCVNNs) with time-varying uncertainties via feedback control. Compared with the traditional control methods, a more practical and general control scheme with the available uncertain information of the parameters is newly developed for MCVNNs. Our approach considers the proposed neural networks as two dynamic real-valued systems. Then, the less conservative exponential synchronization criteria are proposed by incorporating the framework of the Lyapunov method and inequality techniques. Under the proposed algorithm, not only can the stability of MCVNNs be guaranteed but also the behavior of such a system is appropriate for image protection. Meanwhile, the sensitive measure of the encryption and decryption can be converted into synchronization error. When monitoring the secure mechanism as a whole, the influence of error feasible domain on image decryption is analyzed. Simulation examples are provided to verify the efficacy of the proposed synchronization criterion and the results of practical application on image protection.

Current 3D object detection methods adopt an analogous box prediction structure with the 2D methods, which predict center and size of the object simultaneously in a box regression procedure, leading to the poor performance of 3D detector to a great extent. In this work, we propose S-VoteNet, which converts the prediction of 3D bounding box into two parts: center prediction and size prediction. By introducing a novel spherical proposal, S-VoteNet uses vote groups to predict the center and radius of object rather than all parameters of 3D bounding box. The prediction of radius is used to constrain the object size, and the radius-based spherical center loss is applied to measure the geometric distance between the proposal and ground-truth. To make better use of the geometric information provided by point cloud, S-VoteNet aggregates seeds by the votes indices to generate seed groups. The seed groups are then used for box size regression and orientation estimation. By decoupling the localization and size estimation, our method effectively reduces the regression burden of the 3D detector. Experimental results on SUN RGB-D 3D detection benchmark demonstrate that our S-VoteNet achieves state-of-the-art performance by using only point cloud as input.

In this paper, we present an exploratory work towards the recognition of activities and performing real-time objective assessment in human patient simulation (HPS). Although HPS has been pervasively used in medical and nursing programs in developed countries, there is a huge need in providing consistent and objective assessment on student performance during HPS. Current methods all depend on instructor subjective observation, which not only could lead to inconsistency in evaluation across different students and different instructors, but also are very time and resource intensive. Recognizing complex human activities in the context of HPS is very challenging because it involves the recognition of human actions, gestures, as well as human-object and human-mannequin interactions. Hence, we study the feasibility of developing such a system for a particular simulation where a student is required to first identify the patient and then place a neck brace on the patient's neck. The system we that we have developed identifies the actions and activities in the simulation and provides qualitative assessment on the student performance using computer vision, OpenPose, and TensorFlow. The system also consists of a debriefing mobile app that the student and instructor could use to view an automatically generated report with supporting key frames captured and annotated by our system.

Image semantic segmentation tasks and methods based on weakly supervised conditions have been proposed and achieve better and better performance in recent years. However, the purpose of these tasks is mainly to simplify the labeling work. In this paper, we establish a new and more challenging task condition: weaklier supervision with one image level annotation per category, which only provides prior knowledge that humans need to recognize new objects, and aims to achieve pixel-level object semantic understanding. In order to solve this problem, a three-stage semantic segmentation framework is put forward, which realizes image level, pixel level, and object common features learning from coarse to fine grade, and finally obtains semantic segmentation results with accurate and complete object regions. Researches on PASCAL VOC 2012 dataset demonstrates the effectiveness of the proposed method, which makes an obvious improvement compared to baselines. Based on fewer supervised information, the method also provides satisfactory performance compared to weakly supervised learning-based methods with complete image-level annotations.

In this paper, a novel method is developed for day-ahead daily river-flow forecasting based on convolutional neural network (CNN). The proposed method incorporates both spatial and temporal information to improve the forecasting performance. A CNN model is usually trained by minimizing the mean squared error which is, however, sensitive to few particularly large errors. This character of squared error loss function will result in a poor estimator. To tackle the problem, a robust loss function is proposed to train the CNN. To facilitate the training of CNNs for multiple sites forecasting, transfer learning is also applied in this study. With transfer learning, a new CNN inherits the structure and partial learnable parameters from a well-trained CNN to reduce the training complexity. The forecasting performance of the proposed method is validated with real data of four rivers by comparing with widely used benchmarking models including the autoregressive model, multilayer perception network, kernel ridge regression, radial basis function neural network, and generic CNN. Numerical results show that the proposed method performs best in terms of the root mean squared error, mean absolute error, and mean absolute percentage error. The two-sample Kolmogorov-Smirnov test is further applied to assess the confidence on the conclusion.

In recent years, more and more wearable sensors have been employed in smart health applications. Wearable sensors not only can be used to collect valuable health-related data of their users, they can be also used in conjunction with other infrastructure-bound sensors, such as Microsoft Kinect sensor, to facilitate privacy-aware fine-grained activity tracking. This fusion of multimodal data promises a new type of smart health applications that coach a user to live a healthier life style by monitoring the user in realtime and reminding him or her when he or she engages in an unhealthy activity. In this paper, we investigate how to achieve fine-grained activity recognition in the context of such an application. In our scheme, the identification accuracy is improved by incorporating a nonlinear and local similarity measure, namely kernel risk-sensitive loss, into a novel multilayer neural network learning algorithm, called as stacked extreme learning machine. Furthermore, to achieve a good generalization performance with minimal human intervention, Jaya as a popular optimization algorithm, is also used to adjust key parameters in our proposed approach. The experiments are conducted to verify the effectiveness of the proposed scheme.

Active compliance control is one of the necessary prerequisites for the fine manipulation of hydraulic heavy duty manipulators (HHDMs). The establishment of a rigid-flexible coupling machine-hydraulic multibody dynamics model and an active compliance control algorithm for HHDM are the key problems to be solved urgently in the compliance control of heavy duty manipulators. In this paper, a multibody dynamics model of a machine-hydraulic system with seven degrees of freedom is developed with consideration of HHDM characteristics, such as multi-input multioutput, nonlinearity, and rigid-flexible coupling. Meanwhile, a position/force same loop control algorithm for the compliance control of HHDM is proposed based on genetic neural network. Specifically, the force control system is decomposed into subsystems, while the feedback position and force are output through the processing of the joint position controller, torque controller, and multibody dynamics model. Cosimulation results present the feasibility and effectiveness of the proposed dynamics model and control algorithm. Moreover, the experimental environment and the HHDM control system are also developed, while the operation experiment of the control system for HHDM is accordingly conducted. Experiment results show that the control system can realize precise control of position and force, and can successfully complete the function of active compliance control operation.

In this paper, we introduce a novel mechanism to facilitate automatic user authentication for privacy-aware vision based human activity tracking. The mechanism relies on the Bluetooth beacon technology, which localizes a user who is present in the field of view of a camera. Unlike our previous mechanism designed for the same purpose, which requires a user to push a button on the smartwatch he or she is wearing and make a predefined gesture to register with the activity tracking system, this new mechanism does not require the user to alter his or her work routine when entering the view of the tracking system. Hence, the proposed mechanism significantly increases the usability of the tracking system. While the importance of automatic user authentication might not be obvious to researchers, it is essential for the acceptance of the activity tracking technology in its intended venues where workers will be monitored on their jobs as demonstrated by our previous field study and by our interviews with business owners. We report the experimental result based on two types of Bluetooth beacon devices, one from Estimote, and the other from Gimbal. Through the experiments, we identify several challenges in using these commercial-off-the shelf beacon devices for automatic user authentication, including the lack of synchronized beacon signal transmission by different beacon devices, nonuniform beacon transmission with occasional large gaps, and delay in beacon signal detection or reporting, and propose solutions to these issues.

Recently, wind speed forecasting as an effective computing technique plays an important role in advancing industry informatics, while dealing with these issues of control and operation for renewable power systems. However, it is facing some increasing difficulties to handle the large-scale dataset generated in these forecasting applications, with the purpose of ensuring stable computing performance. In response to such limitation, this paper proposes a more practical approach through the combination of extreme-learning machine (ELM) method and deep-learning model. ELM is a novel computing paradigm that enables the neural network (NN) based learning to be achieved with fast training speed and good generalization performance. The stacked ELM (SELM) is an advanced ELM algorithm under deep-learning framework, which works efficiently on memory consumption decrease. In this paper, an enhanced SELM is accordingly developed via replacing the Euclidean norm of the mean square error (MSE) criterion in ELM with the generalized correntropy criterion to further improve the forecasting performance. The advantage of the enhanced SELM with generalized correntropy to achieve better forecasting performance mainly relies on the following aspect. Generalized correntropy is a stable and robust nonlinear similarity measure while employing machine learning method to forecast wind speed, where the outliers may exist in some industrially measured values. Specifically, the experimental results of short-term and ultra-short-term forecasting on real wind speed data show that the proposed approach can achieve better computing performance compared with other traditional and more recent methods.

In this paper, we report our experiences in designing, deploying, and making continuous improvements of a Privacy-Aware Compliance Tracking System (PACTS) at a skilled nursing facility. The purpose of PACTS is to help state tested nursing assistants (STNAs) get into the habit of using proper body mechanics when performing bedside cares. The system has been deployed in six resident rooms and seven STNAs have been participating our study for over ten weeks. This study makes the following contributions: (1) A registration mechanism that enables an STNA to register with any of the rooms that have PACTS installed, which is essential to protect the privacy of patients and non-participating persons; (2) A wrong activity detection mechanism that is robust against occlusions due to furniture and against temporary inability of floor determination; (3) A lease-based mechanism to improve the usability of PACTS, which allows an STNA to be continuously monitored without having to register with PACTS repeatedly when she/he goes in and out of the view of the Kinect sensor for the duration of the lease.

Poor form of body mechanics has been attributed to as a major risk factor for lower back injuries, which costs billions of dollars a year in the US alone. In this paper, we report a case study on using a Kinect-based system during an annual competency training at a local nursing home to promote safe resident handling. Each participant of the study was asked to perform three specific bedside care tasks while being monitored by our system, and he/she was provided with realtime feedback in the form of a vibration via smart watch that he/she worn on detection of a wrong activity by our system. At the end of the session, the participant was asked to complete a short survey regarding the performance of our system and his/her opinion about the system usability. There are two major findings in this case study: (1) the majority of the nursing assistants engaged in poor body mechanics frequently when performing the designated tasks, which indicated that traditional training is not rigorous and may fail to accomplish its purpose, and (2) most participants expressed positive attitude towards using our system for competency training as well as during their jobs to reduce the risk of injuries.

Lost productivity from lower back injuries in workplaces costs billions of U.S. dollars per year. A significant fraction of such workplace injuries are the result of workers not following best practices. In this paper, we present the design, implementation, and evaluation of a novel computer-vision-based system that aims to increase the workers' compliance to best practices. The system consists of inexpensive programmable depth sensors, wearable devices, and smart phones. The system is designed to track the activities of consented workers using the depth sensors, alert them discreetly on detection of noncompliant activities, and produce cumulative reports on their performance. Essentially, the system provides a valuable set of services for both workers and administrators toward a healthier and, therefore, more productive workplace. This study advances the state of the art in the following ways: 1) a set of mechanisms that enable nonintrusive privacy-aware selective tracking of consented workers in the presence of people that should not be tracked; 2) a single sign-on worker identification mechanism; 3) a method that provides realtime detection of noncompliant activities; and 4) a usability study that provides invaluable feedback regarding system design and deployment, as well as future areas of improvements.

The identification and resolution technology are the prerequisite for realizing identity consistency of physical-cyber space mapping in the Internet of Things (IoT). Face, as a distinctive noncoded and unstructured identifier, has especial advantages in identification applications. With the increase of face identification based applications, the requirements for computation, communication, and storage capability are becoming higher and higher. To solve this problem, we propose a fog computing based face identification and resolution scheme. Face identifier is first generated by the identification system model to identify an individual. Then, a fog computing based resolution framework is proposed to efficiently resolve the individual's identity. Some computing overhead is offloaded from a cloud to network edge devices in order to improve processing efficiency and reduce network transmission. Finally, a prototype system based on local binary patterns (LBP) identifier is implemented to evaluate the scheme. Experimental results show that this scheme can effectively save bandwidth and improve efficiency of face identification and resolution.

Kernel least-mean mixed-norm (KLMMN) algorithm as a special kernel adaptive filter method achieves good performance when the measured noises are distributed with a linear combination of long-tails and short-tails. In order to reduce the computational efforts and improve the accuracy, this paper proposes a novel entropy optimized kernel learning algorithm, called E-KLMMN, on the basis of information entropy and KLMMN. The first step of E-KLMMN algorithm is to calculate the entropy weights of input vectors in the training set which contains the linear combination of long-tailed and short-tailed distribution noises. Then we remove the input vectors and their corresponding outputs whose entropy weights are less than the average value. Finally, using the modified training set to train KLMMN model, the following data points thus could be predicted. Through the use of information entropy, the proposed algorithm E-KLMMN has the advantages of high precision and low cost, while employing it to noise environment. We use the actual data to conduct the experiment, and the comparisons among E-KLMMN, KLMS, and KLMMN demonstrate the effectiveness and superiority of our algorithm.

Extreme learning machine (ELM) as a learning algorithm for neural networks (NN) could provide the best generalization performance at extremely fast leaning speed. Through the use of ELM, it is thus possible to improve the existing schemes especially the ones whose learning speed is not fast enough while addressing control problems. As a popular NN-based approach for control applications, direct heuristic dynamic programming (DHDP) with a good capability of adaptive learning has been successfully applied to solve control problems. But limited by slow learning algorithms in NN, it imposes very challenging obstacles to the real-time controller design of DHDP, which keeps it from widely applied. In this paper, driven by the interest of improving learning speed of DHDP while maintaining its good approximation performance, we employ ELM as a learning algorithm in DHDP. The proposed ELM-based DHDP learning scheme is tested on a cart-pole balancing control problem. The simulation results show the proposed scheme has better learning performance than traditional DHDP. Furthermore, this paper provides a novel idea of applying ELM in control problems.

It is obvious that the highly nonlinear nature of dynamic behavior in hypersonic vehicle system impose very challenging obstacles to the controller design. In this paper, we discuss the design of a novel type-2 fuzzy dynamic characteristic modeling method to attitude tracking control problem for hypersonic vehicles in gilding phase. The type-2 (T2) fuzzy logic is introduced into the characteristic modeling (CM) method. Unlike the traditional fuzzy dynamic modeling method normally with a fixed local linear model in every fuzzy subspace, our approach performs CM in subspace, which actually can handle the nonlinearity well while reducing the number of fuzzy rules. After dividing the whole restriction region into several subspaces, the whole nonlinear system can be regarded as a T2 fuzzy “blending” of each individual characteristic model. Then this novel T2 fuzzy logic system modelled by decomposition of a complex nonlinear system into a collection of local CMs, can overcome the deficiencies of traditional fuzzy dynamic modeling and CM approaches. Therefore, it can achieve a better trade-off between tracking accuracy and convergence efficiency in the controller design for hypersonic vehicles. Simulation results under the conditions of certainty and uncertainty are given to show the effectiveness of the novel method for the attitude tracking control of hypersonic vehicles in gilding phase.

Since the launch of the scramjet, recent years have witnessed a growing interest in the study of airbreathing hypersonic vehicles. Due to its strong coupling characteristics, high nonlinearity, and uncertain parameters, the control of hypersonic vehicle becomes a great challenge. To deal with those design issues, we propose an adaptive learning control method based on direct heuristic dynamic programming (direct HDP), which is used to track the angle of attack despite the presence of bounded uncertain parameters. Inspired by the adaptive critic designs, direct HDP is one of the adaptive dynamic programming (ADP) methods, which is a model-free reinforcement learning algorithm using the online learning scheme to solve dynamic control problems in realistic complex environment. In this paper, this direct HDP method is improved by embedding the fuzzy neural network (FNN) in the controller design to enhance its self-learning ability and robustness. Simulation results are provided to demonstrate the effectiveness of our proposed method.

There has been a growing interest in the study of adaptive/approximate dynamic programming (ADP) in recent years. The ADP technique provides a powerful tool to understand and improve the principled technologies of machine intelligence system. As one of the ADP algorithms based on adaptive critic neural networks (NNs), the direct heuristic dynamic programming (direct HDP) has demonstrated some successful applications in solving realistic engineering control problems. In this study, based on a three-network architecture in which the reinforcement signal is approximated by an additional NN, a novel integrated design method for intensified direct HDP is developed. The new design approach is implemented by using multiple PID neural networks (PIDNNs), which effectively takes into account structural knowledge of system states and control that are usually present in a physical system. By using a Lyapunov stability approach, a uniformly ultimately boundedness (UUB) result is proved for our PIDNNs-based intensified direct HDP learning controller. Furthermore, the learning and control performances of the proposed design is tested using the popular cart-pole example to illustrate the key ideas of this paper.

The issue of designing a high performance controller to track a desired system trajectory is one of most important problems in control theory and practice. More recently, there has been a growing interest in the study of tracking control problem. In this paper, we discuss the design and stability properties of a special approximate/adaptive dynamic programming (ADP) method for a general multiple-input-multiple-output (MIMO) discrete-time nonlinear optimal tracking control problem. The direct heuristic dynamic programming (HDP) design algorithm is firstly derived by incorporating the PID control rule into neural networks (NNs). This design approach considers using not only the typical state variables but also their derivatives and cumulative sums as inputs to the controller output. It is therefore expected to retain PID controller properties with additional learning capability. Moreover, our nonlinear control problem is formulated under a general condition that system nonlinearity is unknown and therefore it introduces modelling errors for the controller design. By using a Lyapunov stability construct, we provide new results of uniformly ultimately boundedness (UUB) for the proposed PIDNN-based direct HDP controller in discrete-time nonlinear tracking setting with desired tracking performance.