In the process of planting crops, detection diseases in leaf parts is one key links to prevention and control crop diseases. This paper takes tomato leaves as experimental objects, uses deep learning method extract disease features on surface, including three most common species (Spot blight, Late blight Yellow curl disease). After continuous iterative learning, network can predict category each picture. For diseases, 1000 pictures were selected, divided into 900 for training set (2700...

In this article, two robust adaptive control schemes are investigated for a class of completely non-affine pure-feedback non-linear systems with input non-linearity and perturbed uncertainties using radial basis function neural networks (RBFNNs). Based on the dynamic surface (DSC) technique quadratic Lyapunov function, explosion complexity in traditional backstepping design is avoided when gain signs known. addition, unknown virtual dealt Nussbaum functions. Using mean value theorem Young's...

Few-shot object detection (FSOD) aims to detect novel objects with limited annotated examples. Mainstream methods suffer from the data scarcity of classes insufficient intra-class variations, which makes trained model biased base classes. Actually, there are massive unlabeled instances in dataset and their adequate utilization will enhance discriminability This paper proposes a semi-supervised few-shot method, utilizes teacher pre-trained detector guide learning student through adaptive...

Abstract Porous hydrogels have been developed to be highly efficient interfacial solar vapor generation materials for obtaining affordable freshwater supplies. However, realizing hydrogel with portability, adequate water supply, and durable mechanical properties remains challenging. Herein, a scalable portable 3D‐shaped recovery, transport, robust stability is reported, which prepared by foaming polyvinyl alcohol (PVA) modification of phosphotungstic acid (PTA). Due the interconnected porous...

Abstract Current approaches to defect detection and segmentation make essential use of machine learning methods. To develop lightweight models is one key tasks for many applications. In this work, we present a trilateral parallel feature extraction with multi-feature aggregation network (TriMFANet) surface segmentation. TriMFANet, the top lateral feature-rich used capture detailed information. The other two laterals, efficient semantic (ESFE) reverse ESFE, leverage Hadamard product attention...

In this technical note, H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> control problem of continuous Markovian jump systems is investigated. A linear feedback scheme, combined with a state observer design, proposed in the form matrix inequalities, which can ensure systems' mean-square stability performance. Then, multi-step transition conditional probability function introduced for process, used to define system's probability....

In this brief, sufficient conditions are proposed for the existence of compact sets in neural network controls. First, we point out that set a classical control scheme is unsolved and its result incomplete. Next, as simple case, derive condition first-order systems. Finally, propose neural-network-based backstepping high-order nonlinear The theoretic illustrated through simulation example.

A novel approach to transport a large object by multiple mobile manipulators in unknown environments with obstacles is proposed this paper. The main moving direction of the system as well its width and leading manipulator are first assigned manually or obtained automatically according geometric layout manipulators. Then, three transportation modes, i.e., default mode, shrink incline defined based on constraints from In distributed around be transported initial layout. Mobile mode required...

Summary In this article, observer‐based decentralized adaptive neural control is proposed for uncertain interconnected nonlinear systems with input quantization and time‐varying output restrictions. Decentralized hysteresis quantizer employed to handle signal. The unmeasured states are estimated by designing K‐filters. A Lyapunov description used dispose of state unmodeled dynamics. constrained transformed into novel without constraints constructing invertible mappings. Dynamic surface...

Summary In this paper, we are concerned with the problem of adaptive output‐feedback tracking control for nonlinear systems input quantization, unmodeled dynamics, and output constraints. A novel quantizer advantages hysteresis uniform is introduced to handle signals. barrier Lyapunov function employed solve The state dynamics solved by using a description, neural networks used approximate unknown smooth functions produced in design process. controller simplified combining new dynamic...

The high precision measurement of binocular camera depends on the calibration accuracy. Mainstream methods focus intrinsic parameters and pose relationship two cameras in sequence. However, transmission error will decrease quality. It is still challenging especially case an unknown imaging model due to intervention baffle multimedia refraction. In this article, a general method for vision system with proposed. Based one-to-one mapping between 4-D image coordinates 3-D Cartesian spatial...

A novel distributed hunting approach for multiple autonomous robots in unstructured mode-free environments, which is based on effective sectors and local sensing, proposed this paper. The visual information, encoder sonar data are integrated the robot's frame, sector introduced. task modelled as three states: search state, round-obstacle corresponding switching conditions control strategies given. form of cooperation will emerge where interact only locally with each other. evader, whose...

Bird's-eye-view (BEV) perception has gained popularity since it provides a 3D world representation with scale consistency. Although existing camera-based solutions achieve excellent performance, the BEV positions related to features are still less accurate. In this article, semantic segmentation framework two-stream compact depth transformation and feature rectification is proposed. To balance conflict that maps ensemble tends use two temporal frames long interval, while shorter more...

This paper mainly studies stability and optimal control for networked systems with the real-time setup of time-driven sensor, event-driven controller actuator, assumption that network-induced delay is no longer than certain known times sampling period. The modeling this class given. Then, preliminary stochastic analysis it presented. Finally, based on theory an LQG scheme provided. Much work needs to be continued future research.

In this paper, an adaptive dynamic surface control (DSC) method is proposed for the flexible robotic system with unmodeled dynamics and time-varying output constraints. The disturbances are effectively dealt by introducing a signal. unknown continuous functions approximated using radial basis function neural networks (RBFNNs). An asymmetric barrier Lyapunov (BLF) employed to ensure constraint satisfaction. By theoretical analysis, closed-loop shown be semi-globally uniformly ultimately...

This paper explores the relationship between system stability conditional probability and sliding mode control for second order continuous Markovian jump systems. By using stochastic process theory, multi-step state transition function is proposed time discrete process. A scheme utilized to stabilize The derived. It indicates that a monotonically bounded non-decreasing non-negative piecewise right of parameter. numerical example given show feasibility theoretical results.