Self-powered soft robot in the Mariana Trench.

The deep sea remains the largest unknown territory on Earth because it is so difficult to explore1-4. Owing to the extremely high pressure in the deep sea, rigid vessels5-7 and pressure-compensation systems8-10 are typically required to protect mechatronic systems. However, deep-sea creatures that lack bulky or heavy pressure-tolerant systems can thrive at extreme depths11-17. Here, inspired by the structure of a deep-sea snailfish15, we develop an untethered soft robot for deep-sea exploration, with onboard power, control and actuation protected from pressure by integrating electronics in a silicone matrix. This self-powered robot eliminates the requirement for any rigid vessel. To reduce shear stress at the interfaces between electronic components, we decentralize the electronics by increasing the distance between components or separating them from the printed circuit board. Careful design of the dielectric elastomer material used for the robot's flapping fins allowed the robot to be actuated successfully in a field test in the Mariana Trench down to a depth of 10,900 metres and to swim freely in the South China Sea at a depth of 3,224 metres. We validate the pressure resilience of the electronic components and soft actuators through systematic experiments and theoretical analyses. Our work highlights the potential of designing soft, lightweight devices for use in extreme conditions.

SF-YOLOv5: A Lightweight Small Object Detection Algorithm Based on Improved Feature Fusion Mode.

In the research of computer vision, a very challenging problem is the detection of small objects. The existing detection algorithms often focus on detecting full-scale objects, without making proprietary optimization for detecting small-size objects. For small objects dense scenes, not only the accuracy is low, but also there is a certain waste of computing resources. An improved detection algorithm was proposed for small objects based on YOLOv5. By reasonably clipping the feature map output of the large object detection layer, the computing resources required by the model were significantly reduced and the model becomes more lightweight. An improved feature fusion method (PB-FPN) for small object detection based on PANet and BiFPN was proposed, which effectively increased the detection ability for small object of the algorithm. By introducing the spatial pyramid pooling (SPP) in the backbone network into the feature fusion network and connecting with the model prediction head, the performance of the algorithm was effectively enhanced. The experiments demonstrated that the improved algorithm has very good results in detection accuracy and real-time ability. Compared with the classical YOLOv5, the mAP@0.5 and mAP@0.5:0.95 of SF-YOLOv5 were increased by 1.6% and 0.8%, respectively, the number of parameters of the network were reduced by 68.2%, computational resources (FLOPs) were reduced by 12.7%, and the inferring time of the mode was reduced by 6.9%.

Improved GBS-YOLOv5 algorithm based on YOLOv5 applied to UAV intelligent traffic.

As the road traffic situation becomes complex, the task of traffic management takes on an increasingly heavy load. The air-to-ground traffic administration network of drones has become an important tool to promote the high quality of traffic police work in many places. Drones can be used instead of a large number of human beings to perform daily tasks, as: traffic offense detection, daily crowd detection, etc. Drones are aerial operations and shoot small targets. So the detection accuracy of drones is less. To address the problem of low accuracy of Unmanned Aerial Vehicles (UAVs) in detecting small targets, we designed a more suitable algorithm for UAV detection and called GBS-YOLOv5. It was an improvement on the original YOLOv5 model. Firstly, in the default model, there was a problem of serious loss of small target information and insufficient utilization of shallow feature information as the depth of the feature extraction network deepened. We designed the efficient spatio-temporal interaction module to replace the residual network structure in the original network. The role of this module was to increase the network depth for feature extraction. Then, we added the spatial pyramid convolution module on top of YOLOv5. Its function was to mine small target information and act as a detection head for small size targets. Finally, to better preserve the detailed information of small targets in the shallow features, we proposed the shallow bottleneck. And the introduction of recursive gated convolution in the feature fusion section enabled better interaction of higher-order spatial semantic information. The GBS-YOLOv5 algorithm conducted experiments showing that the value of mAP@0.5 was 35.3[Formula: see text] and the mAP@0.5:0.95 was 20.0[Formula: see text]. Compared to the default YOLOv5 algorithm was boosted by 4.0[Formula: see text] and 3.5[Formula: see text], respectively.

Mechanical Performance and Failure Analysis of a 3D-Printed "Continuous Layer-Lattice Layer-Continuous Layer" Sandwich Structure.

Sandwich structures are engineered with continuous layers surrounding the inner lattices, which combines the advantages of the high strength of the continuous layer and the light weight of the lattice layer. They are widely employed in weight-critical energy-absorbing engineering fields such as aerospace, automobile, and robotics. However, the application of sandwich structures made of polymer matrix composites is still limited due to lack of essential performance investigation and adequate reference data. The following innovative works are accomplished in this paper: (i) Continuous long glass fiber (CGF) is employed within the continuous layer of the sandwich structure, with composite short carbon fiber/polyamide (SCF/N) applied within the lattice layer. (ii) Sandwich structures with different cell types and orientations of the lattice infills are designed and prepared by additive manufacturing. (iii) The basic mechanical properties of the sandwich structures, i.e., the bi-directional tension/compression compound performance, failure modes and mechanisms in characteristic directions, are analyzed systematically. (iv) The effects of geometric features on the three-point bending properties of L-shaped sandwich structures are investigated and compared with those of pure SCF/N structures. The results show that the bending resistance per unit weight was up to 54.3% larger than that of pure SCF/N, while the weight could be decreased by 49%, and the bending flexibility before fracture could be increased by 44%. These studies contribute fundamental research data to the application of sandwich structures prepared by fiber reinforced polymer matrix composites.

A new detection algorithm for alien intrusion on highway.

In recent years, highway accidents occur frequently, the main reason is that there is always foreign body invasion on the highway, which makes people unable to respond to emergencies in time. In order to reduce the occurrence of highway incidents, an object detection algorithm for highway intrusion was proposed in this paper. Firstly, a new feature extraction module was proposed to better preserve the main information. Secondly, a new feature fusion method was proposed to improve the accuracy of object detection. Finally, a lightweight method was proposed to reduce the computational complexity. We compare the algorithm in this paper with existing algorithms, the experimental results showed that: On the Visdrone dataset (small size targets), (a) the CS-YOLO was 3.6% more accurate than the YOLO v8. (b) The CS-YOLO was 1.2% more accurate than the YOLO v8 on the Tinypersons dataset (minimal size targets). (c) CS-YOLO was 1.4% more accurate than YOLO v8 on VOC2007 data set (normal size).

Speaker-turn aware diarization for speech-based cognitive assessments.

Introduction: Speaker diarization is an essential preprocessing step for diagnosing cognitive impairments from speech-based Montreal cognitive assessments (MoCA). Methods: This paper proposes three enhancements to the conventional speaker diarization methods for such assessments. The enhancements tackle the challenges of diarizing MoCA recordings on two fronts. First, multi-scale channel interdependence speaker embedding is used as the front-end speaker representation for overcoming the acoustic mismatch caused by far-field microphones. Specifically, a squeeze-and-excitation (SE) unit and channel-dependent attention are added to Res2Net blocks for multi-scale feature aggregation. Second, a sequence comparison approach with a holistic view of the whole conversation is applied to measure the similarity of short speech segments in the conversation, which results in a speaker-turn aware scoring matrix for the subsequent clustering step. Third, to further enhance the diarization performance, we propose incorporating a pairwise similarity measure so that the speaker-turn aware scoring matrix contains both local and global information across the segments. Results: Evaluations on an interactive MoCA dataset show that the proposed enhancements lead to a diarization system that outperforms the conventional x-vector/PLDA systems under language-, age-, and microphone-mismatch scenarios. Discussion: The results also show that the proposed enhancements can help hypothesize the speaker-turn timestamps, making the diarization method amendable to datasets without timestamp information.

A lightweight YOLOv3 algorithm used for safety helmet detection.

YOLOv3 is a popular and effective object detection algorithm. However, YOLOv3 has a complex network, and floating point operations (FLOPs) and parameter sizes are large. Based on this, the paper designs a new YOLOv3 network and proposes a lightweight object detection algorithm. First, two excellent networks, the Cross Stage Partial Network (CSPNet) and GhostNet, are integrated to design a more efficient residual network, CSP-Ghost-Resnet. Second, combining CSPNet and Darknet53, this paper designs a new backbone network, the ML-Darknet, to realize the gradient diversion of the backbone network. Finally, we design a lightweight multiscale feature extraction network, the PAN-CSP-Network. The newly designed network is named mini and lightweight YOLOv3 (ML-YOLOv3). Based on the helmet dataset, the FLPSs and parameter sizes of ML-YOLOv3 are only 29.7% and 29.4% of those of YOLOv3. Compared with YOLO5, ML-YOLOv3 also exhibits obvious advantages in calculation cost and detection effect.

A Novel Treatment of Postpartum Depression and Review of Literature.

Early-onset postpartum depression has been shown to have a unique neurobiological basis compared to major depressive disorder, implying a need for targeted treatments such as the recent Food and Drug Administration (FDA)-approved brexanolone. In this case report, a woman with a past medical history of major depressive disorder was diagnosed with postpartum depression due to worsening mood with suicidal and homicidal ideations. She was treated with vilazodone and aripiprazole with good effect after consideration of her past medication trials. Her regimen is unique in clinical practice and not reported in current literature for the treatment of postpartum depression. It may represent a safe and effective medication choice, especially in the context of current first-line treatments that have a high treatment failure rate. More research is needed to find treatments that address the unique challenges of postpartum women.

Force and Microstructure Variation of SLM Prepared AlMgSc Samples during Three-Point Bending.

Lightweight parts manufactured by metal selective laser melting (SLM) are widely applied in machinery industries because of their high specific strength, good energy absorption effect, and complex shape that are difficult to form by mechanical machining. These samples often serve in three-dimensional stress states. However, previous publications mainly focused on the unidirectional tensile/compressive properties of the samples. In this paper, AlMgSc samples with different geometric parameters were prepared by the SLM process, and the variation of force and microstructure during three-point bending were systematically investigated. The results demonstrate that the deformation resistance of these samples has good continuity without mutation in bending, even for brittle materials; the bending force-displacement curves exhibit representative variation stages during the entire bending process; the equivalent bending strength deduced from free bending formula is not applicable when compactability is less than 67%. The variations of grain orientation and size of the three representative bending layers also show regularity.

Stable and Fast Planar Jumping Control Design for a Compliant One-Legged Robot.

Compliant bipedal robots demonstrate a potential for impact resistance and high energy efficiency through the introduction of compliant elements. However, it also adds to the difficulty of stable control of the robot. To motivate the control strategies of compliant bipedal robots, this work presents an improved control strategy for the stable and fast planar jumping of a compliant one-legged robot designed by the authors, which utilizes the concept of the virtual pendulum. The robot was modeled as an extended spring-loaded inverted pendulum (SLIP) model with non-negligible torso inertia, leg inertia, and leg damping. To enable the robot to jump forward stably, a foot placement method was adopted, where due to the asymmetric feature of the extended SLIP model, a variable time coefficient and an integral term with respect to the forward speed tracking error were introduced to the method to accurately track a given forward speed. An energy-based leg rest length regulation method was used to compensate for the energy dissipation due to leg damping, where an integral term, regarding jumping height tracking error, was introduced to accurately track a given jumping height. Numerical simulations were conducted to validate the effectiveness of the proposed control strategy. Results show that stable and fast jumping of compliant one-legged robots could be achieved, and the desired forward speed and jumping height could also be accurately tracked. In addition to that, using the proposed control strategy, the robust jumping performance of the robot could be observed in the presence of disturbances from state variables or uneven terrain.

Symmetric Convolutional and Adversarial Neural Network Enables Improved Mental Stress Classification From EEG.

Electroencephalography (EEG) is widely used for mental stress classification, but effective feature extraction and transfer across subjects remain challenging due to its variability. In this paper, a novel deep neural network combining convolutional neural network (CNN) and adversarial theory, named symmetric deep convolutional adversarial network (SDCAN), is proposed for stress classification based on EEG. The adversarial inference is introduced to automatically capture invariant and discriminative features from raw EEG, which aims to improve the classification accuracy and generalization ability across subjects. Experiments were conducted with 22 human subjects, where each participant's stress was induced by the Trier Social Stress Test paradigm while EEG was collected. Stress states were then calibrated into four or five stages according to the changing trend of salivary cortisol concentration. The results show that the proposed network achieves improved accuracies of 87.62% and 81.45% on the classification of four and five stages, respectively, compared to conventional CNN methods. Euclidean space data alignment approach (EA) was applied and the improved generalization ability of EA-SDCAN across subjects was also validated via the leave-one-subject-out-cross-validation, with the accuracies of four and five stages being 60.52% and 48.17%, respectively. These findings indicate that the proposed SDCAN network is more feasible and effective for classifying the stages of mental stress based on EEG compared with other conventional methods.

Severe Anxiety Post-COVID-19 Infection.

COVID-19 infection is linked to increased risk of neuropsychiatric symptoms such as psychosis and suicidal ideation/behavior. After further review of the literature, there is not a large body of data on anxiety following COVID-19 infection. Most literature found is related to fear/anxiety of contracting and dying from COVID-19. We illustrate a case of a 27-year-old male with no previous psychiatric treatment history or symptomology, who developed severe anxiety with intrusive thoughts of self-harm via firearm after COVID-19 infection. Given the severe nature of the anxiety and intrusive thoughts, the patient feared for his safety and sought acute inpatient admission. The patient was effectively treated with group therapy and psychotropic medications and was able to be discharged in a timely manner with outpatient psychiatric follow-up. Much is still unknown of COVID-19. With this case report, we discuss a potential relationship between anxiety and COVID-19 infection.

Edge-Semantic Learning Strategy for Layout Estimation in Indoor Environment.

Visual cognition of the indoor environment can benefit from the spatial layout estimation, which is to represent an indoor scene with a 2-D box on a monocular image. In this paper, we propose to fully exploit the edge and semantic information of a room image for layout estimation. More specifically, we present an encoder-decoder network with shared encoder and two separate decoders, which are composed of multiple deconvolution (transposed convolution) layers, to jointly learn the edge maps and semantic labels of a room image. We combine these two network predictions in a scoring function to evaluate the quality of the layouts, which are generated by ray sampling and from a predefined layout pool. Guided by the scoring function, we apply a novel refinement strategy to further optimize the layout hypotheses. Experimental results show that the proposed network can yield accurate estimates of edge maps and semantic labels. By fully utilizing the two different types of labels, the proposed method achieves the state-of-the-art layout estimation performance on the benchmark datasets.

A recognition method for extreme bradycardia by arterial blood pressure signal modeling with curve fitting.

OBJECTIVE: The aim of this study is to investigate the potential of arterial blood pressure (ABP) signal for the detection of the subjects with life-threatening extreme bradycardia (EBr). APPROACH: The steps of the proposed method include ABP signal preprocessing, ABP wave segmentation, model parameter estimation, and EBr subject detection. First, the noise, interference and abnormal segments are eliminated in the pre-processing. Then, the ABP signal is segmented into a series of ABP waves by cardiac cycles. The pulse decomposition analysis (PDA) approach is presented to quantitively describe the changes in ABP waves. The back-propagation neural network, probabilistic neural network and decision tree (DT) are engaged to design the classifiers to discriminate the EBr subjects from healthy subjects by the parameters of PDA models. The international physiological signal databases of Fantasia for healthy subjects and 2015 PhysioNet/CinC Challenge for EBr subjects are exploited to validate the proposed method, and 79 310 ABP waves of healthy subjects and 4595 ABP waves of EBr subjects are extracted. MAIN RESULTS: We obtain the average PDA models of healthy subjects and EBr subjects and derive their changes. The two-sample Kolmogorov-Smirnov test result shows that all model parameters are markedly different (H= 1, P < 0.05) between the healthy and EBr subjects. The classification results show that the DT has the best performance with specificity of 99.74% ± 0.07%, sensitivity of 93.12% ± 1.24%, accuracy of 99.37% ± 0.10% and kappa coefficient of 93.92% ± 0.92%. SIGNIFICANCE: The proposed method has the potential to detect EBr subjects by the ABP signal.

Further Results on Adaptive Stabilization of High-Order Stochastic Nonlinear Systems Subject to Uncertainties.

This paper concerns the adaptive state-feedback control for a class of high-order stochastic nonlinear systems with uncertainties including time-varying delay, unknown control gain, and parameter perturbation. The commonly used growth assumptions on system nonlinearities are removed, and the adaptive control technique is combined with the sign function to deal with the unknown control gain. Then, with the help of the radial basis function neural network approximation approach and Lyapunov-Krasovskii functional, an adaptive state-feedback controller is obtained through the backstepping design procedure. It is verified that the constructed controller can render the closed-loop system semiglobally uniformly ultimately bounded. Finally, both the practical and numerical examples are presented to validate the effectiveness of the proposed scheme.

Whole genome sequencing for the genetic diagnosis of heterogenous dystonia phenotypes.

INTRODUCTION: Dystonia is a clinically and genetically heterogeneous disorder and a genetic cause is often difficult to elucidate. This is the first study to use whole genome sequencing (WGS) to investigate dystonia in a large sample of affected individuals. METHODS: WGS was performed on 111 probands with heterogenous dystonia phenotypes. We performed analysis for coding and non-coding variants, copy number variants (CNVs), and structural variants (SVs). We assessed for an association between dystonia and 10 known dystonia risk variants. RESULTS: A genetic diagnosis was obtained for 11.7% (13/111) of individuals. We found that a genetic diagnosis was more likely in those with an earlier age at onset, younger age at testing, and a combined dystonia phenotype. We identified pathogenic/likely-pathogenic variants in ADCY5 (n = 1), ATM (n = 1), GNAL (n = 2), GLB1 (n = 1), KMT2B (n = 2), PRKN (n = 2), PRRT2 (n = 1), SGCE (n = 2), and THAP1 (n = 1). CNVs were detected in 3 individuals. We found an association between the known risk variant ARSG rs11655081 and dystonia (p = 0.003). CONCLUSION: A genetic diagnosis was found in 11.7% of individuals with dystonia. The diagnostic yield was higher in those with an earlier age of onset, younger age at testing, and a combined dystonia phenotype. WGS may be particularly relevant for dystonia given that it allows for the detection of CNVs, which accounted for 23% of the genetically diagnosed cases.

Detection of multipoint pulse waves and dynamic 3D pulse shape of the radial artery based on binocular vision theory.

BACKGROUND AND OBJECTIVE: Pulse signals contain a wealth of human physiological and pathological information. How to get full pulse information is especially challenging, and most of the traditional pulse sensors can only get the pulse wave of a single point. This study is aimed at developing a binocular pulse detection system and method to obtain multipoint pulse waves and dynamic three-dimensional pulse shape of the radial artery. METHODS: The proposed pulse detection approach is image-based and implemented by two steps. First, a new binocular pulse detection system is developed based on the principle of pulse feeling used in traditional Chinese medicine. Second, pulse detection is achieved based on theories and methods of binocular vision and digital image processing. In detail, the sequences of pulse images collected by the designed system as experimental data are sequentially processed by median filtering, block binarization and inversion, area filtering, centroids extraction of connected regions, to extract the pattern centroids as feature points. Then stereo matching is realized by a proposed algorithm based on Gong-shape scan detection. After multipoint spatial coordinate calculation, dynamic three-dimensional reconstruction of the thin film shape is completed by linear interpolation. And then the three-dimensional pulse shape is achieved by finding an appropriate reference time. Meanwhile, extraction of multipoint pulse waves of the radial artery is accomplished by using a suitable reference origin. The proposed method is analyzed from three aspects, which are pulse amplitude, pulse rate and pulse shape, and compared with other detection methods. RESULTS: Analysis of the results shows that the values of pulse amplitude and pulse rate are consistent with the characteristics of pulse wave of the radial artery, and pulse shape can correctly present the shape of pulse in space and its change trend in time. The comparison results with the other two previously proposed methods further verify the correctness of the presented method. CONCLUSIONS: The designed binocular pulse detection system and proposed algorithm can effectively detect pulse information. This tactile visualization-based pulse detection method has important scientific significance and broad application prospects and will promote further development of objective pulse diagnosis.

An Efficient Method for Traffic Sign Recognition Based on Extreme Learning Machine.

This paper proposes a computationally efficient method for traffic sign recognition (TSR). This proposed method consists of two modules: 1) extraction of histogram of oriented gradient variant (HOGv) feature and 2) a single classifier trained by extreme learning machine (ELM) algorithm. The presented HOGv feature keeps a good balance between redundancy and local details such that it can represent distinctive shapes better. The classifier is a single-hidden-layer feedforward network. Based on ELM algorithm, the connection between input and hidden layers realizes the random feature mapping while only the weights between hidden and output layers are trained. As a result, layer-by-layer tuning is not required. Meanwhile, the norm of output weights is included in the cost function. Therefore, the ELM-based classifier can achieve an optimal and generalized solution for multiclass TSR. Furthermore, it can balance the recognition accuracy and computational cost. Three datasets, including the German TSR benchmark dataset, the Belgium traffic sign classification dataset and the revised mapping and assessing the state of traffic infrastructure (revised MASTIF) dataset, are used to evaluate this proposed method. Experimental results have shown that this proposed method obtains not only high recognition accuracy but also extremely high computational efficiency in both training and recognition processes in these three datasets.

Chronic renal failure rats are highly sensitive to aristolochic acids, which are nephrotoxic and carcinogenic agents.

Aristolochic acid (AA), a component of some Chinese herbal medicines, may cause Chinese Herbs Nephropathy (CHN) and multi-systemic tumors by the formation of AA-DNA adducts. In this study, we established an animal model to further characterize the mechanisms of AA-induced diseases. Our results indicated that AA significantly inhibited rat growth in terms of weight gain. By measuring the serum creatinine levels, AA resulted in considerable damage to the rat renal system, not only for those in which chronic renal failure (CRF) was induced but also for normal healthy rats. Mutation-specific polymerase chain reaction (PCR) and XbaI restriction fragment length polymorphism (RFLP) revealed the CAA-->CTA transversion mutation at codon 61 of the H-ras proto-oncogene from the stomach tissues of CRF rats fed with AA, but not from other tissues of rats in the same experimental group. In addition, no such mutations were found in the tissues of CRF rats without AA treatment or healthy rats fed with AA. Our results strongly demonstrated that AA was in fact nephrotoxic and carcinogenic, especially to those CRF rats.

Development of a robotic device for facilitating learning by children who have severe disabilities.

This paper presents technical aspects of a robot manipulator developed to facilitate learning by young children who are generally unable to grasp objects or speak. The severity of these physical disabilities also limits assessment of their cognitive and language skills and abilities. The CRS robot manipulator was adapted for use by children who have disabilities. Our emphasis is on the technical control aspects of the development of an interface and communication environment between the child and the robot arm. The system is designed so that each child has user control and control procedures that are individually adapted. Control interfaces include large push buttons, keyboards, laser pointer, and head-controlled switches. Preliminary results have shown that young children who have severe disabilities can use the robotic arm system to complete functional play-related tasks. Developed software allows the child to accomplish a series of multistep tasks by activating one or more single switches. Through a single switch press the child can replay a series of preprogrammed movements that have a development sequence. Children using this system engaged in three-step sequential activities and were highly responsive to the robotic tasks. This was in marked contrast to other interventions using toys and computer games.