Large language models guided evolutionary frameworks for program synthesis have demonstrated strong results across combinatorial benchmarks; however, such frameworks have not been widely explored for robotic control systems. This article presents an automated framework for synthesizing control logic and optimizing numerical coefficients, enabling the deployment of generated controllers onto robotic platforms. To produce deployable controllers, the framework separates control logic synthesis from numerical parameter optimization, with controller coefficients optimized via particle swarm optimization using a Markov decision process reward signal. The resulting controllers are evaluated in a custom uncrewed aerial vehicles (UAV) simulation environment, validated using PX4 software-in-the-loop, and subsequently deployed on a physical UAV. Controller performance is evaluated on lemniscate and lissajous trajectory-tracking tasks and compared against proportional–integral–derivative with disturbance observer and linear quadratic regulator baselines. Across both trajectories, the framework-generated control law achieves improved tracking accuracy relative to the baseline controllers, with a minimum reduction in mean squared error of approximately 38% in real-world experiments. These results demonstrate the feasibility of deploying automatically synthesized control logic on a physical UAV, bridging automated program synthesis and real-world control deployment.

Boundary tracking of sea continuous objects (e.g., oil spills and radioactive waste) is a challenging task that can be tackled viaunderwater acoustic sensor networks. Existing methods operate by selecting sensor nodes in the proximity of the boundary, and tend to over- or underestimate the actual boundary of the continuous object. In this article, a boundary tracking algorithm termedfeasible region search for continuous objects(FRSCO) is proposed. To determine a feasible region where the actual boundary lies, the proposed method first bounds the continuous object inside a minimum elliptical boundary. Within the minimum boundary ellipse, cell partition is performed through a binary tree structure, from which a set of backbone cells is selected. These roughly localize the feasible region. By constructing and deconstructing convex hulls of nodes in each backbone cell, the feasible region location uncertainty is further narrowed down. Virtual nodes are introduced in the final feasible region to determine the boundary nodes – by applying the principle of maximum entropy. Similar to virtual nodes, the selected boundary nodes do not need to correspond to the actual sensor nodes in the proximity of the boundary. Results from realistic testbed experiments and simulations show that the FRSCO exhibits effective tracking accuracy.

The massive growth in interconnected devices from a multiplicity of networks goes hand-in-hand with the emergence of the Internet of Things (IoT) paradigm. As a critical component of the IoT, sensor networks have become ubiquitous and widely used in various application domains. However, open-ended wireless communication brings severe threats to user privacy and security. Attackers from outside the network can trace back along the data stream to capture the source node, which poses a significant threat to the privacy of the data source. A feasible defense method is to interfere with the attacker’s tracking process through forged data streams. However, the related traditional solutions generally have shortcomings in terms of balancing security and efficiency. Therefore, this article proposes a pseudopacket scheduling algorithm (PPSA), which aims at reasonably regulating the process of pseudopacket generation to interfere with the adversary’s tracking to the data source. The algorithm comprises three phases. First, the sink node performs geographic information acquisition and neighbor node discovery with a flood-based method. Then, the sink node uses a self-adapting proxy selection method to construct backbone routes with both randomness and low latency to receive actual packets. Finally, the nodes on both sides of the backbone routes follow a pseudopacket scheduling strategy to interfere with the adversary’s tracking of the source locations. The experimental results showcase that our proposed scheme effectively controls the additional energy consumption and transmission delays within acceptable ranges while ensuring adequate location privacy.

Intelligent equipment and flexible production lines are at the cores of smart manufacturing. Meanwhile, Internet of Things and Artificial Intelligence have provided new solutions for the intelligent equipment management and maintenance in mixed model assembly (MMA). This article focuses on knowledge-driven techniques, and it proposes a knowledge sharing-enabled multirobot collaboration (KS-enabled MRC) strategy for preventive maintenance of robots in MMA. First, a formal semantic environment for MMA is constructed by way of ontology-enabled semantic modeling. Then,task-related action primitives and ontology-based robot skill bases are established according to robot capability and task environment. Finally, the Wu-Palmer similarity metric and first-order logic are leveraged to match and reason new tasks according to the semantic rules, and a knowledge sharing and update mechanism are developed for this application. Experimental results demonstrate that the proposed KS-enabled MRC can reduce unscheduled downtime and assist in achieving a load balance for robots in MMA. The studied MRC can potentially avoid severe equipment degradation, thus acting as a preventive maintenance paradigm of complex equipment. Furthermore, it is applicable across different platforms and exhibits high deployment efficiency without intense programming requirements.

The development of theindustrial Internet of Thingsparadigm brings forth the possibility of a significant transformation within the manufacturing industry. This paradigm is based on sensing large amounts of data, so that it can be employed by intelligent control systems (i.e.,artificial intelligencealgorithms) eliciting optimal decisions in real time. Ensuring the accuracy and reliability of the intelligent wireless sensing and control system pipeline is crucial toward achieving this goal. Nevertheless, the presence of noise in actual wireless transmission processes considerably affects the quality of the sensed data. Typically, noise and anomalies present in the data are very difficult to distinguish from each other. Conventional anomaly-detection techniques generate many error reports, which cause the control systems to issue incorrect responses that hinder the industrial production. In this article, a novel solution is proposed to denoise data while simultaneously preserving the actual anomalies. The proposed approach operates by measuring both the neighbor and background contrasts in computing a noise score. The trust level of each data point is then calculated through a correlation measure to purge spurious data. Extensive experiments on real datasets demonstrate that the proposed approach yields effective performance, as compared to existing methods, and it meets the requirements of low latency—facilitating the normal operation of the monitored control systems.

Underwater acoustic sensor networks (UASNs) have been widely promoted for developing various categories of marine applications, where the sensor nodes cooperate to complete specific tasks. Given the fact that the sensor nodes are unattended while continuously exposed to harsh environments, an associatedtrust modelplays a significant role in node trustworthiness evaluation and defective node detection, such as the case of adverse attacks on the network. However, the existing trust models only evaluate the communication behavior and the energy of the sensor nodes, ignoring the effects of underwater environmental noise on trust reliability. Further, most trust models are designed with arbitraty weighted trust metrics, causing inevitable evaluation errors. To achieve the accurate calculation of node trust, we propose a new anomaly and attack resilient trust model, based on the isolation forest. We refer to this model asITrust. The proposed ITrust model consists of two phases: trust metrics specifics and defective node detection. In the first phase, the trust dataset is integrated from four types of trust metrics: communication trust, data trust, energy trust, and environment trust. In the second stage, trust is evaluated with the obtained trust dataset using the isolation forest algorithm. Simulation results demonstrate that the proposed ITrust can detect defective nodes effectively, and achieves higher detection accuracy than that of the existing trust models in a noisy environment.

Underwater acoustic sensor networks (UASNs) have been widely applied in marine scenarios, such as offshore exploration, auxiliary navigation and marine military. Due to the limitations in communication, computation, and storage of underwater sensor nodes, traditional security mechanisms are not applicable to UASNs. Recently, various trust models have been investigated as effective tools towards improving the security of UASNs. However, the existing trust models lack flexible trust update rules, particularly when facing the inevitable dynamic fluctuations in the underwater environment and a wide spectrum of potential attack modes. In this study, a novel trust update mechanism for UASNs based on reinforcement learning (TUMRL) is proposed. The scheme is developed in three phases. First, an environment model is designed to quantify the impact of underwater fluctuations in the sensor data, which assists in updating the trust scores. Then, the definition of key degree is given; in the process of trust update, nodes with higher key degree react more sensitively to malicious attacks, thereby better protecting important nodes in the network. Finally, a novel trust update mechanism based on reinforcement learning is presented, to withstand changing attack modes while achieving efficient trust update. The experimental results prove that our proposed scheme has satisfactory performance in improving trust update efficiency and network security.

Due to the flammability, explosiveness and toxicity of continuous objects (e.g., chemical gas, oil spill, radioactive waste) in the petrochemical and nuclear industries, boundary tracking of continuous objects is a critical issue for industrial wireless sensor networks (IWSNs). In this article, we propose a continuous object boundary tracking algorithm for IWSNs – which fully exploits the collective intelligence and machine learning capability within the sensor nodes. The proposed algorithm first determines an upper bound of the event region covered by the continuous objects. A binary tree-based partition is performed within the event region, obtaining a coarse-grained boundary area mapping. To study the irregularity of continuous objects in detail, the boundary tracking problem is then transformed into a binary classification problem; ahierarchical soft margin support vector machinetraining strategy is designed to address the binary classification problem in a distributed fashion. Simulation results demonstrate that the proposed algorithm shows a reduction in the number of nodes required for boundary tracking by at least 50 percent. Without additional fault-tolerant mechanisms, the proposed algorithm is inherently robust to false sensor readings, even for high ratios of faulty nodes ($\approx 9\%$).

The diffusion of toxic gas, biochemical material, and radio-active contamination – known as continuous objects – endangers the safe production of the petrochemical and nuclear industries. To mitigate these well known hazards, the new paradigm ofindustrial wireless sensor networks(IWSNs) shows great potential in monitoring evolving hazardous phenomena in unfriendly industrial fields. In order to prolong the lifetime of these networks, existing research focuses on energy-efficient boundary nodes selection. However, sensor state cannot be scheduled proactively, due to the difficulty in predicting the spatiotemporal evolution of diffusive hazards. In this article, we propose amotion behavior learning predictive tracking(MBLPT) algorithm for continuous objects in IWSNs. Considering the relatively unpredictable patterns exhibited by continuous objects, the MBLPT uses a data-driven approach for motion state recognition, and then utilizesBayesian model averaging(BMA) for future boundary prediction. The prediction of the MBLPT provides the knowledge for establishing a wake-up zone, in which standby nodes are activated in advance to participate in tracking the upcoming boundary. Simulation results demonstrate that the MBLPB achieves superior energy efficiency while keeping effective tracking accuracy.

As an effective security mechanism, trust models have been proposed to estimate the reliability of the individual nodes in Underwater Acoustic Sensor Networks (UASNs) during adverse attacks. However, existing trust models neglect the relative importance of the different nodes within the network topology. Further, few trust models study the effects of defective recommendation trust filtering. In this work, we propose an adaptive trust model based on the Long Short-Term Memory (LSTM) network model for UASNs, which we term LTrust. The LTrust is composed of two stages: trust data collection and trust evaluation. In the first stage, the characteristics of the network topology are leveraged towards evaluating direct trust evidence, by aggregating the communication trust and environment trust metrics; a defective recommendation filtering method is designed for broadcasting accurate trust recommendations among the nodes. In the second stage, an adaptive trust model is designed based on the LSTM model, to identify anomalous nodes by evaluating their trust value. The LTrust model has been tested under both hybrid attack and single-mode attack scenarios. Simulation results demonstrate that the LTrust achieves effective performance, as compared to other approaches proposed in the literature, in terms of trust value, accuracy and error rate.

In recent years, developments in data collection schemes based on multipleautonomous underwater vehicles(AUVs) are facilitating the realization of the so-calledunderwater acoustic sensor networks(UASNs). As yet, the lack of suitable collaboration mechanisms among multiple AUVs, which are based on functional or resource distributions, prevents effective information sharing and yields increased data collection delays, thus reducing the capacity of the networks. In this article, to address these shortcomings, we propose astate prediction-based data collection(SPDC) algorithm for UASNs. The principle of operation is as follows. First, some cluster pairs named observation clusters obtain and exchange the state information about AUVs between the adjacent subregions. Based on the shared information, the AUVs predict each other’s status and adjust their data collection areas. Then, the AUVs use a heuristic strategy to complete the path planning based on the updated access area. Finally, a scheduling data forwarding mechanism reduces the diving number of the AUVs, by reasonably allocating the overlapped data unloading intervals between the AUVs and a mobile sink. Experimental results prove that the proposed algorithm shows satisfactory performance in reducing data collection delays and in improving the total network lifetime.

In this paper, a reliable online destination prediction methodology is presented. The destination prediction methodology consists of a novel sequential complete diameter distance limited clustering method and an ensemble of random forest classifiers employing a one-vs-rest binarization strategy. Through the use of a novel OvR Uncertainty metric, predictions with high uncertainty could be withheld, thus increasing the overall reliability of the predictions made. The methodology was validated on 778 journeys from two real non-commuter vehicles based in the UK. These datasets allowed the methodology to be tested on real, yet challenging-to-predict journeys and irregular driver behavior. The sequential complete diameter distance limited clustering method was found to be a fast and effective method for sequentially clustering GPS coordinates into clusters that correspond to geographical locations. Prediction results showed that while only an overall mean prediction accuracy of 52% and 34% could be achieved on the two datasets, mean prediction accuracy could be significantly increased to over 90% and 73% respectively by only providing predictions with low uncertainty.

Industrial Internet of Things inevitably leads to the implementation of highly data-intensive devices, where the associated sensing nodes accelerate the energy consumption rate, which ultimately produces an energy bottleneck. To address this issue, this article proposes adynamic collaborative charging algorithmthat acts on both the mobile nodes and the static nodes in a sensing node network. The proposed scheme is to design a collaborative group of charging robots that can rendezvous with the sensing nodes. The group includes aerial charging vehicles (ACVs)—able to charge the underpowered mobile nodes, and terrestrial charging vehicles (TCVs), which charge their targeted static nodes. The aim of this study is to optimize the charging effect and the energy cost in the rendezvous process. This approach consists of two subalgorithms: 1) a charging algorithm for mobile nodes (CAMNs) and 2) a charging algorithm for static nodes (CASNs). The CAMNs is designed so that each underpowered mobile node can be charged by a dedicated ACV. For this purpose, a deep learning model is trained to divide the underpowered mobile nodes into appropriate clusters, each of which is equipped with a mobile base station. The rendezvous process is then constructed as a mixed continuous/discrete optimization problem, which is solved by using the firefly algorithm. In addition, the CASNs ensures that the TCVs traverse their routes, charging static nodes as they proceed. This traversing process was formulated as a multiobjective optimization problem, solved by using genetic algorithm. Through various experiments and case studies, the results have demonstrated both the feasibility and the efficiency of the proposed algorithms.

As a result of the increasing deployment of Industrial-Internet-of-Things (IIoT) architectures, large volumes of multidimensional data are continuously generated. An important issue with these data is that higher dimensionality increases the degree of fragmentation. Furthermore, data sets collected by IIoT nodes often display outliers, which are usually caused by anomalous events or errors. These outliers contain considerable valuable information, which prevent the normal operation of the system. Thus, methodologies are able to quantify the obtained information to protect the high priority IIoT nodes, are crucial. This study aims at developing such a method driven by sixth-generation (6G) networks. The proposed algorithm uses a multidimensional data relationship diagram to characterize the spatiotemporal correlations among heterogeneous data. Then, an autoregressive exogenous model is used to eliminate the effects of noise on sensor data, and to help in detecting anomalies. Finally, the algorithm produces a Cumulative Coefficient of Value (CCoV), to identify high-value sensing devices and enable massive Internet of Things (IoT) with 6G-using the characteristic patterns hidden within the data. The experimental results demonstrate that the proposed method can effectively handle the effects of the ubiquitous interference noise in complex industrial environments. Moreover, the method yields effective anomaly detection and compensates for some of the shortcomings in traditional methods.

The fifth-generation of wireless communication (5G) is a promising paradigm toward massive interconnectivity within Internet-of-Things (IoT) networks. However, because the data traffic throughput sharply increases with the number of IoT devices, a tremendous burden on the backhaul links and core networks results. With this in mind, mobile edge caching is an effective method that can relieve stress of the backhaul links, while decreasing the service latency. The purpose of this study is to analyze the problem of jointly optimizing cooperative edge caching and radio resource allocation in 5G-enabled massive IoT networks. For that, a joint optimization long-term nonlinear integer programming problem is posed. This class of problems is known to be NP-hard; thus, to reduce the problem complexity, a divide and conquer scheme will be applied—the task at hand will be divided into two subproblems: 1) cooperative edge caching and 2) radio resource allocation. The cooperative edge caching subproblem is formulated as a constrained Markov decision process. Herein, a deep reinforcement learning method to optimize the caching decisions for all the edge nodes. Then, based on the resulting optimal caching decisions, the radio resource allocation subproblem for each edge node is posed as an NLIP problem, and an improved branch-and-bound method is proposed to yield the optimal radio resource allocation decisions for each edge node. Extensive simulations were performed to confirm that the proposed methods have the capability of enhancing the content caching hit ratio, while lessening the content retrieving delays for 5G-enabled massive IoT networks—improving over various baseline algorithms.

The traditional production paradigm of large batch production does not offer flexibility toward satisfying the requirements of individual customers. A new generation of smart factories is expected to support new multivariety and small-batch customized production modes. For this, artificial intelligence (AI) is enabling higher value-added manufacturing by accelerating the integration of manufacturing and information communication technologies, including computing, communication, and control. The characteristics of a customized smart factory are: self-perception, operations optimization, dynamic reconfiguration, and intelligent decision-making. The AI technologies will allow manufacturing systems to perceive the environment, adapt to the external needs, and extract the process knowledge, including business models, such as intelligent production, networked collaboration, and extended service models. This article focuses on the implementation of AI in customized manufacturing (CM). The architecture of an AI-driven customized smart factory is presented. Details of intelligent manufacturing devices, intelligent information interaction, and construction of a flexible manufacturing line are showcased. The state-of-the-art AI technologies of potential use in CM, that is, machine learning, multiagent systems, Internet of Things, big data, and cloud-edge computing, are surveyed. The AI-enabled technologies in a customized smart factory are validated with a case study of customized packaging. The experimental results have demonstrated that the AI-assisted CM offers the possibility of higher production flexibility and efficiency. Challenges and solutions related to AI in CM are also discussed.

While full automation of road vehicles remains a future goal, shared-control and semiautonomous driving-involving transitions of control between the human and the machine-are more feasible objectives in the near term. These alternative driving modes will benefit from new research toward novel steering control devices, more suitably where machine intelligence only partially controls the vehicle. In this article, it is proposed that when the human shares the control of a vehicle with an autonomous or semiautonomous system, a force control, or nondisplacement steering wheel (i.e., a steering wheel which does not rotate but detects the applied torque by the human driver) can be advantageous under certain schemes: tight rein or loose rein modes according to the H -metaphor. We support this proposition with the first experiments to the best of our knowledge, in which human participants drove in a simulated road scene with a force control steering wheel (FCSW). The experiments exhibited that humans can adapt promptly to force control steering and are able to control the vehicle smoothly. Different transfer functions are tested, which translate the applied torque at the FCSW to the steering angle at the wheels of the vehicle; it is shown that fractional order transfer functions increment steering stability and control accuracy when using a force control device. The transition of control experiments is also performed with both: a conventional and an FCSW. This prototypical steering system can be realized via steer-by-wire controls, which are already incorporated in commercially available vehicles.

In this article, we propose a hierarchical data collection scheme, toward the realization of unmanned aerial vehicle (UAV)-aided industrial wireless sensor networks. The particular application is that of agricultural monitoring. For that, we propose the use of hybrid compressed sampling through exact and greedy approaches. With the exact approach-to model the energy-optimal formulation-an improved linear programming formulation of the minimum cost flow problem was utilized. The greedy approach is based on a proposed balance factor parameter, consisting of data sparsity, and distance from cluster head to normal nodes. To improve node clustering efficiency, a hierarchical data collection scheme is implemented, by which nodes in different layers are adaptively clustered, and the UAV can be scheduled to perform energy-efficient data collection. Simulation results show that our method can effectively collect the data and plan the path for the UAV at a low energy cost.

The aim of this article is to develop a methodology for measuring thedegree of unpredictabilityin dynamical systems with memory, i.e., systems with responses dependent on a history of past states. The proposed model is generic, and can be employed in a variety of settings, although its applicability here is examined in the particular context of an industrial environment: gas turbine engines. The given approach consists in approximating the probability distribution of the outputs of a system with a deep recurrent neural network; such networks are capable of exploiting the memory in the system for enhanced forecasting capability. Once the probability distribution is retrieved, theentropyormissing informationabout the underlying process is computed, which is interpreted as the uncertainty with respect to the system's behavior. Hence, the model identifies how far the system dynamics are from its typical response, in order to evaluate the system reliability and to predict system faults and/ornormal accidents. The validity of the model is verified with sensor data recorded from commissioning gas turbines, belonging to normal and faulty conditions.

In this paper, a methodology for assessing the unpredictability of systems with memory was developed. The proposed approach consists in approximating the probability distribution exhibited by the response of a system, understood as a stochastic process, with a deep recurrent neural network; such networks offer increased forecasting capability by exploiting an accumulative register of previous system states. Once the probability distribution is computed, the uncertainty or entropy of the underlying process is measured. This measure determines the degree of regularity in the system, and identifies how atypical the system dynamics are. The proposed model was validated by identifying industrial gas turbine engine faults from recorded sensor data.

In this paper, a new steering control model is introduced, motivated by several characteristics of human driving. The model uses as input an optical variable portraying visual information directly accessible to the driver: the splay error, representing the lane positioning aspect of driving. The splay error is regulated through a multiplicative control model; this approach displays similar statistical properties to those found in human compensatory control. Further, multiplicative control exhibits steering pulse behavior related to human steering. A second input variable in the model, the critical normalized yaw rate, reflects the information from the far region of the road. The parameters of the model are optimized for low and high vehicle speeds through a genetic algorithm. With the fitted parameters, the response of the model is compared to driver behavior by means of steering workload measurement, and validated with naturalistic driving data.

The aim of this paper is to estimate and monitor the human response delay in manual control tasks. A probability distribution analysis is applied on the response delay, based on experimental data collected from human subjects controlling a dynamic system and responding to visually perceived errors via joystick or steering wheel. The distribution analysis includes firstly a sliding segment method, to extract the delay time for each slice of data. Then, probability distributions of the delay time are fitted by using a bootstrap based goodness-of-fit test. For both manual-control cases, with a joystick and a steering wheel respectively, the experimental data can be explained reasonably by a Gamma distribution. Consequently, the Gamma distribution parameters for different human subjects are compared. Based on these findings, an online monitoring method of the level of attention in the human-operator - or applied workload - is proposed, which could be of interest for relevant shared-control applications.

A non-linear, yet simple, multiplicative human-control model is developed by studying the statistical properties of human subjects' motor response to a visual input; statistical analysis of the magnitude of the steering angle, from subjects performing a tracking task with a steering wheel, shows that the data is consistent with a log-normal distribution. Thus the possibility of modelling human-control as a multiplicative process, that replicates the statistical properties found in the human-operator is considered. The proposed multiplicative controller is contrasted with real data and with the Crossover Model. This research has potential applications in a wide range of fields, from human performance modelling to the development of human-machine interfaces, particularly in the application of ground vehicle automation.

The objective of this paper is to estimate the compressor discharge temperature measurements on an industrial gas turbine that is undergoing commissioning at site, using a data-driven model which is built using the test bed measurements of the engine. This paper proposes a Bayesian neuro-fuzzy modelling (BNFM) approach, which combines the adaptive neuro-fuzzy inference system (ANFIS) and variational Bayesian Gaussian mixture model (VBGMM) techniques. A data-driven compressor model is built using ANFIS, and VBGMM is applied in the set-up stage to automatically select the number of input membership functions in the fuzzy system. The efficacy of the proposed BFNM approach is established through experimental trials of a sub-15MW gas turbine, and the results, from the model that is built using test bed data, are shown to be promising for estimating the compressor discharge temperatures on the gas turbine during commissioning.

In this paper, the manipulative control actions of human operators interacting with fractional order plants are studied. Experimental data were recorded from subjects using a joystick or a steering wheel and responding to plants with different fractional dynamics. From the data it is established that human operators can identify and learn to respond to fractional order plants. Moreover, it is proven that the classical Crossover model is not valid to represent man-machine systems with fractional order dynamics. A generalized Fractional Crossover model is proposed for systems with memory effect. The model is validated with experimental data.

This paper presents a novel methodology for modeling human lane keeping control by characterizing a unique concept of elementary steering pulses, which are motor primitives in man-vehicle systems. The novelty of the paper is the introduction of elementary steering pulses that have been evidently extracted from naturalistic driving data through machine learning techniques (data-driven modeling), and are incorporated into an alternative steering control scheme. This newly proposed hybrid-open-closed-loop control scheme starts an elementary steering pulse with an open-loop steering actuation, representing real human's reflex responses triggered by human lane keeping errors, and adjusts it back with the traditional closed-loop control. This shows a significant improvement on both the stability and the matching performance to real driving events. Online measurement of the key metrics in the steering process provides a new tool for monitoring driver states, and the biofidelic steering model may provide human-like qualities for future automated lane keeping systems. Both will add to the array of tools available for achieving autonomous and semiautonomous driving systems, which greatly benefits the current vehicle industry.