Adaptive Capture Control of a Continuum Manipulator With Self-Powered Sensors

An adaptive deep neural network is used in an inverse system identification setting to approximate the inverse of a nonlinear plant with the aim of constituting the plant controller by copying to the latter the weights and architecture of the converging deep neural network. This deep learning (DL) approach to the adaptive inverse control (AIC) problem is shown to outperform the adaptive filtering techniques and algorithms normally used in adaptive control, especially when in nonlinear plants. The deeper the controller, the better the inverse function approximation, provided that the nonlinear plant has an inverse and that this inverse can be approximated. Simulation results prove the feasibility of this DL-based adaptive inverse control scheme. The DL-based AIC system is robust to nonlinear plant parameter changes in that the plant output reassumes the value of the reference signal considerably faster than with the adaptive filter counterpart of the deep neural network. The settling and rise times of the step response are shown to improve in the DL-based AIC system.

Predicting residue-residue distance relationships (e.g. contacts) has become the key direction to advance protein structure prediction since 2014 CASP11 experiment, while deep learning has revolutionized the technology for contact and distance distribution prediction since its debut in 2012 CASP10 experiment. During 2018 CASP13 experiment, we enhanced our MULTICOM protein structure prediction system with three major components: contact distance prediction based on deep convolutional neural networks, distance-driven template-free (ab initio) modeling, and protein model ranking empowered by deep learning and contact prediction. Our experiment demonstrates that contact distance prediction and deep learning methods are the key reasons that MULTICOM was ranked 3rd out of all 98 predictors in both template-free and template-based structure modeling in CASP13. Deep convolutional neural network can utilize global information in pairwise residue-residue features such as co-evolution scores to substantially improve contact distance prediction, which played a decisive role in correctly folding some free modeling and hard template-based modeling targets. Deep learning also successfully integrated 1D structural features, 2D contact information, and 3D structural quality scores to improve protein model quality assessment, where the contact prediction was demonstrated to consistently enhance ranking of protein models for the first time. The success of MULTICOM system clearly shows that protein contact distance prediction and model selection driven by deep learning holds the key of solving protein structure prediction problem. However, there are still challenges in accurately predicting protein contact distance when there are few homologous sequences, folding proteins from noisy contact distances, and ranking models of hard targets.

Real-time monitoring of rainwater is a critical issue in the development of autonomous vehicles and smart homes, while the corresponding sensors play a pivotal role in ensuring their sensitivity. Here, we study a self-powered intelligent water droplet monitoring sensor based on a solid-liquid triboelectric nanogenerator (SL-TENG). The sensor comprises a SL-TENG, a signal acquisition module, a central processing unit (CPU), and a wireless transmission module, facilitating the real-time monitoring of water droplet signals. It is worth noting that the SL-TENG has self-powering characteristics and can convert the kinetic energy of water droplets into electrical energy. The excellent output performance, with open-circuit voltage of 9 V and short-circuit current of 2 μA without any treatment of the SL-TENG, can provide an effective solution to the problem that traditional sensor need battery replacement. In addition, the SL-TENG can generate stable amplitude electrical signals through water droplets, exemplified by the absence of decay in a short-circuit current within 7 days. More importantly, the sensor is equipped with intelligent analytical capabilities, allowing it to assess rainfall based on variables such as amplitude and frequency. Due to its excellent stability and intelligent analysis, this sensor can be used for roof rainwater monitoring, intravenous administration monitoring, and especially in automobile automatic wipers and other fields.

Continuum manipulators are gaining widespread acceptance in commercial robotics, particularly in the medical field, where their compliance allows a large benefit for patient safety. However, this compliance also makes precise position control of these manipulators quite difficult. This paper presents two closed-loop control implementations applied to a small scale continuum manipulator. These implementations are both based on manipulator tip position feedback from an electromagnetic sensor. The command tracking and disturbance rejection properties of the two control implementations are shown to be approximately equivalent, and provide improved position control when compared to open-loop control, without sacrificing system stability.

A point cloud deep neural network metamodel method for aerodynamic prediction

A miniature self-powered electrochemical sensor for the determination of patulin based on an integrated photocatalytic fuel cell

Introduction The exhaled acetone has been proved as a typical biomarker for Type-I diabetes. Normally, the acetone concentration from the exhale breath of healthy body is lower than 0.8 ppm, while the patient’s breath contains more than 1.8 ppm acetone gas [1]. Therefore, developing a reliable and accurate acetone gas sensor is of great significance. Meanwhile, traditional acetone gas sensors depend on the output power supplies, combining them with self-powered devices can fully utilize the energy of the external environment, which guarantee the normal operation of the sensors under the statues of power failure.Currently, the metal oxide semiconductor (MOS)-based acetone sensors are extensively researched owing to their high response and fast response-recovery times. However, the excessive power consumption restrains their further application. Recently, chitosan (CTS) has been proved to be an acetone sensing material working at room temperature [2]. For enhancing the gas sensing performance, fictionizing chitosan with other materials might be an effective way. Therefore, the chitosan (CTS)/zinc oxide (ZnO) bilayer sensing film based TENG as a self-powered gas sensor has been designed for detecting acetone at room temperature. Results indicated that the CTS/ZnO bilayer sensing film exhibited greater response of acetone than the CTS monolayer sensing film. Experimental ZnO were prepared by a hydrothermal method. Specifically, 3 mmol zinc acetate dihydrate and 5 mmol hexamethylenetetramine were dissolved in 30 mL of deionized water and ultrasonically treated for 30 min to form a clear solution. Next, the solution was transferred into an autoclave, and subsequently heated up to 150℃ for 210 min before cooling naturally. After that, the mixtures were centrifugated and washed by ethanol and deionized water for several times, and dried at 80℃ in an oven to obtain the ZnO.The detailed fabrication process of the self-powered gas sensor (Fig. 1a) are demonstrated as follows: First, 0.1 mm thickness of the flexible polyethylene terephthalate (PET) with the size of 3 cm×3 cm was covered by a gold electrode layer via thermal evaporation. Then, 1mL of CTS solution (4 mg/mL, dissolved in 2 vol.% of acetic acid) and 1.5 mL of ZnO aqueous solution (2 mg/mL) were deposited onto the PET substrate with gold electrode layer via spray coating method, respectively. Meanwhile, the PDMS friction layer was transferred onto the gold electrode layer mentioned above. Finally, the PET substrate with the gas-sensing materials was fixed onto a flexible silica gel film as the upper layer, while the PDMS layer was fixed at the bottom of an acrylic test chamber. The whole device was driven by the periodic force exerted on the silica gel film.All the output signals generated by the sensor were measured by a Keithley 6514 electrometer at room temperature (~25℃). By mixing a standard dry acetone gas (100 ppm, dry air was used as the balance gas) with dry air in proportion with the help of two mass flow controllers (MFC 300), the acetone gas with different concentration was generated and introduced into the test chamber. The total flow rate was kept at 200 sccm for unifying the gas-sensing measurement. The response value was calculated as follows: Response (%) = (Vpp-g−Vpp-a)/ Vpp-a×100%, where Vpp-g and Vpp-a are the peak-peak output voltage exposed in air and tested gas, respectively. Results and Conclusions Fig. 1b illustrates the output voltage of the sensor with the exposure to dry air and different concentration of acetone gas at room temperature. The output voltages increased with the increasing acetone concentration in the range of 2-50 ppm. Meanwhile, the CTS/ZnO sensor is of good linear response at 2-10 ppm (R2=0.9872) and 10-50 ppm acetone (R2=0.9030), of which the responses are 6.46-17.29% and 17.29-32.09%, respectively. Based on the above phenomenon, a possible sensing mechanism has been established: After the introduction of acetone gas, the nucleophilic addition reaction between the CTS and acetone would destroy the hydrogen bond interaction between CTS and ZnO (Eqs. (1)), leading to the interruption of electrical pathways for the effective electrons transfer [2], resulting in the decrease of surface conductivity and eventually increase the output voltage [3].CTS-NH2+CH3COCH3=CTS-N=C(CH3)2+H2O(1)In summary, we have successfully fabricated a novel triboelectric gas sensor based on CTS/ZnO bilayer membrane. After testing procedure, a possible sensing mechanism has been established. The CTS/ZnO gas sensor has a great potential in the applications of acetone detection without additional power supplies, and provides a useful strategy for the researches of novel acetone sensors.

Atrazine, used in large quantities of the triazine class of herbicides, is recognized for its difficulty in degradation, high mobility, and significant environmental residue, classifying it among the environmental hormone-like substances. Its propensity to generate highly toxic metabolites and its potential carcinogenicity underscore the imperative need for sensitive and selective detection methods in environmental matrices. Traditional analytical techniques offer high sensitivity and selectivity but are often limited by their time-consuming nature, high costs, and operational complexity. These traditional methods, while effective, are not always suitable for rapid, on-site monitoring. In contrast, sensors based on various nanomaterials and sensing strategies for electrochemical and photoelectrochemical detection have emerged as promising alternatives. These sensors are distinguished by their sensitive responses, rapid detection capabilities, real-time monitoring, and portable design, which facilitate the swift and precise identification of environmental contaminants. This review focuses on key nanomaterials and sensing strategies, including immunosensors, aptasensors, molecular imprinting sensors, and self-powered sensors, all of which hold significant promise for enhancing the performance of atrazine detection sensors. This review comprehensively compiles the prevalent detection methodologies for atrazine over the past half-decade, delving into their foundational principles, recognition mechanisms, categorization, and a critical evaluation of their merits and demerits. It also encapsulates the advancements in the detection of atrazine residues and offers an in-depth analysis of the materials, recognition elements, and sensing strategies that are frequently employed. Furthermore, the review provides an outlook on emerging trends and prospective directions in novel sensor development.

Protein gamma-turn prediction is useful in protein function studies and experimental design. Several methods for gamma-turn prediction have been developed, but the results were unsatisfactory with Matthew correlation coefficients (MCC) around 0.2–0.4. Hence, it is worthwhile exploring new methods for the prediction. A cutting-edge deep neural network, named Capsule Network (CapsuleNet), provides a new opportunity for gamma-turn prediction. Even when the number of input samples is relatively small, the capsules from CapsuleNet are effective to extract high-level features for classification tasks. Here, we propose a deep inception capsule network for gamma-turn prediction. Its performance on the gamma-turn benchmark GT320 achieved an MCC of 0.45, which significantly outperformed the previous best method with an MCC of 0.38. This is the first gamma-turn prediction method utilizing deep neural networks. Also, to our knowledge, it is the first published bioinformatics application utilizing capsule network, which will provide a useful example for the community. Executable and source code can be download at http://dslsrv8.cs.missouri.edu/~cf797/MUFoldGammaTurn/download.html.

Efficient energy utilization while ensuring occupants’ comfort remains a primary objective in smart home environments. This study presents a novel smart home management methodology that integrates Deep Neural Networks (DNNs) with an optimization algorithm to enhance energy efficiency and user comfort. Unlike traditional approaches, which often focus on individual aspects of energy management or user comfort, our method comprehensively addresses both aspects by considering key environmental factors such as air quality, temperature, and illumination. By employing specialized DNNs, our approach dynamically adjusts home conditions to align with user preferences, thereby enhancing overall comfort and satisfaction. The primary innovation of our methodology lies in the Comfort Index Calculation System, which utilizes advanced DNNs to precisely quantify user comfort based on environmental conditions. This system enables our optimization algorithm to iteratively adjust environmental settings, striking a balance between user comfort and energy consumption. By continuously learning from user feedback and environmental data, our integrated approach ensures that smart homes adapt to occupants’ evolving preferences and external conditions, thereby offering a more personalized and responsive living environment. Our results demonstrate the capability of this integrated approach in improving smart home environments, offering significant implications for environmental living and smart building automation. By bridging the gap between energy efficiency and user comfort, our study contributes to advancing the field of smart home management, ultimately realizing the full potential of energy-efficient and comfortable living environments.

An adaptive matching control method of multiple turboshaft engines

Aiming at the problems of low prediction accuracy and efficiency and poor prediction effect in the current psychological pressure prediction methods, a psychological pressure prediction method for college students based on deep neural network is proposed. The structure and algorithm of depth neural network and gray theory model are analyzed. Using the deep neural network, this paper establishes the sample set data of college students’ psychological pressure prediction and constructs the college students’ psychological pressure prediction model combined with the deep neural network algorithm of gray theory. The physical network information model is formed through the relationship between neurons. According to the dynamic changes of college students’ psychological pressure in each neuron of the physical network, the prediction of college students’ psychological pressure is completed. The experimental results show that the proposed method is effective in predicting college students’ psychological pressure and can effectively improve the accuracy and efficiency of college students’ psychological pressure prediction.

Turbulent mixing noise is a vital component of jet noise, and its rapid, accurate prediction has always been persistently pursued. Recent advancement in machine learning has been applied to jet noise prediction. However, these applications are pure curve fitting and lack physical constraints. In this study, a physics-merged deep neural network (PMNN)-based prediction method was developed for turbulent mixing jet noise by merging the physics of the jet noise. This deep neural network (DNN)-based method employed recent advancements in jet turbulent mixing noise containing large- and fine-scale turbulence structures. Two simple rational functions for large- and fine-scale turbulent noise similarity spectra were proposed to replace the original complex similarity spectra functions and incorporated into the DNN-based prediction method. For comparison, we present two data-driven DNN-based prediction methods (DDNN). The first DDNN method used the sound pressure level (SPL) as the output of neural networks, directly establishing the nonlinear relationship between the input features and SPL. In the second DDNN method, the dominant modes of the jet noise spectra extracted using the proper orthogonal decomposition method were merged with DNN. These DNN-based methods were then trained using a set of experimental data over a wide range of jet operating conditions. Their performance was evaluated and compared. It was evident that all these DNN-based methods were capable of predicting turbulent mixing noise reasonably well. In contrast to the DDNN methods, the PMNN method could provide insights into the jet turbulent mixing noise components. It demonstrates that the turbulent mixing jet noise spectra at the mid polar angle is generated by the large-scale noise component at low-frequency range and by the fine-scale noise component at high-frequency range.

Deep learning-assisted digital pathology has demonstrated the potential to profoundly impact clinical practice, even surpassing human pathologists in performance. However, as deep neural network (DNN) architectures grow in size and complexity, their explainability decreases, posing challenges in interpreting pathology features for broader clinical insights into physiological diseases. To better assess the interpretability of digital microscopic images and guide future microscopic system design, we developed a novel method to study the predictive feature length-scale that underpins a DNN’s predictive power. We applied this method to analyze a DNN’s capability in predicting brain metastasis from early-stage non-small-cell lung cancer biopsy slides. This study quantifies DNN’s attention for brain metastasis prediction, targeting features at both the cellular scale and tissue scale in H&E-stained histological whole slide images. At the cellular scale, the predictive power of DNNs progressively increases with higher resolution and significantly decreases when the resolvable feature length exceeds 5 microns. Additionally, DNN uses more macro-scale features associated with tissue architecture and is optimized when assessing visual fields greater than 41 microns. Our study computes the length-scale requirements for optimal DNN learning on digital whole-slide microscopic images, holding the promise to guide future optical microscope designs in pathology applications and facilitating downstream deep learning analysis.

Adaptive Cruise Control (ACC) is an important part of automotive autonomy. To have a robust Adaptive Cruise Control system, a well-defined sensor fusion system is required. This paper focuses on feature-level sensor fusion system where three fusion techniques are explored: independent Kalman filtering, interacting multiple model (IMM) filtering, and a novel hybrid technique using a combination of both an IMM Filter and deep neural network (DNN). Kalman Filtering is a popular approach for automotive sensor fusion, but heavily relies on the given kinematic models to make state estimations. We explore the IMM filter to capture multi-model motion as well as a DNN approach to build a data-driven model of the system. Our IMM filter is built with several Kalman Filters featuring different kinematic models. The proposed DNN used is a Long Short-Term Memory (LSTM) network trained on radar and camera data to forecast a vehicle's state. To overcome the weaknesses in both IMM filtering and LSTM networks, we propose a hybrid technique of consisting of both the IMM filter and the LSTM network. The proposed hybrid system uses a single filter for each sensor. The filtered sensor data is synchronized and used as the input for a trained LSTM network. Our LSTM network was trained on over 100 simulated highway driving scenarios that might occur in an ACC application. Simulations and code were developed using MATLAB and the Autonomous Driving Toolbox. Our hybrid IMM-LSTM system outperformed the independent Kalman Filtering approach in longitudinal and lateral tracking accuracy (RMSE) by 23 percent and had promising results compared to the IMM filter where the tracking error was decreased by nearly 50 percent in certain driving scenarios.