This paper focuses on a problem of adaptive visual tracking control for an uncalibrated image-based servoing manipulator system with actuator fuzzy dead-zone constrain and unknown dynamic. Without prior knowledge the system, logic systems are employed to approximate unmodeled nonlinear dynamics external disturbances. By using recursive Newton-Euler method, total number rules can be reduced significantly as compared traditional system. defuzzifying slope k̅ model deterministic value k̅, novel...

This paper proposes a deep reinforcement learning (DRL) method that combines random network distillation (RND) and long short-term memory (LSTM) to address the tracking control problem, while leveraging inherent symmetry in robotic arm movements eliminate need for or knowing system’s dynamic model. In general, complexity strong coupling of manipulators make trajectory extremely challenging. Firstly, prediction fixed are jointly trained using RND method. The difference output values between...

The use of Raman spectroscopy for real-time gas monitoring has the advantages corresponding speed, high sensitivity and low cost. However, due to overlap each peak position, the...

The real-time detection and recognition of pitaya fruit is an important prerequisite for automatic picking. We combined with the current deep learning method good accuracy to realize identification fruit. Firstly, we collected a large number pictures labeling, completed production data sets Pitaya Then use YOLOV3, YOLOV3-tiny MobileNet-YOLO network models train. After training, test performance trained model on set. experimental results show that improved has better speed than YOLOV3 model,...

The ultra-low hardware consumption feature of stochastic decoding has made it a potential candidate for the implementation low-density parity-check(LDPC) decoders. However, existing LDPC decoders still suffer from performance degradation and relatively high cycles caused by correlation among bit streams. In this paper, we propose Hybrid Stochastic(HS) decoding, which achieves performance, throughput, efficiency jointly using our proposed novel check node(CN) Two's Complement(TCS) variable...

In uncertain environments with robot input saturation, both model-based reinforcement learning (MBRL) and traditional controllers struggle to perform control tasks optimally. this study, an algorithmic framework of Curiosity Model Policy Optimization (CMPO) is proposed by combining curiosity approach, where tracking errors are reduced via training agents on gains for model-free controllers. To begin with, a metric judging positive negative proposed. Constrained optimization employed update...

The sleeve grouting connection is the most common form of vertical for prefabricated shear walls. However, during construction, this type prone to defects such as insufficient anchorage length reinforcement, deviation and amount grouting, which significantly impact integrity seismic performance wall structure. finite element analysis with still based on solid modeling. This method has disadvantages complex models low computational efficiency. In paper, a simplified modeling walls considering...

In this study, the tracking control and saturation nonlinearities problems of a flexible mobile vehicle-manipulator with disturbance uncertainty are investigated. Based on differential equations model, novel iterative learning (ILC) algorithm is designed, where two functions hyperbolic tangent utilized to deal actuator constraint an adaptive law adopted reduce impact uncertainty. With ILC algorithm, target vibration attenuation manipulator can come true. Simulations given verify...

This paper investigates the stabilization and trajectory tracking problem of wheeled mobile robot with a ceiling-mounted camera in complex environment. First, an adaptive visual servoing controller is proposed based on uncalibrated kinematic model due to operation Then, derived provide solution uncertain dynamic control for subject parametric uncertainties. Furthermore, controllers can be applied more general situation where parallelism requirement between image plane no needed. The...

Abstract This paper investigates the coordinated behaviour of multiple Euler–Lagrange systems under diverse interactions with time delays. Specially, in case undirected interconnection, proportional plus damping control strategy is proposed and sufficient conditions bipartite consensus are derived via energy function based approach. Subsequently, a velocity‐free controller further developed by introducing novel second‐order auxiliary system to each agent. In directed network, fully...

Hybrid low-density parity-check (LDPC) decoding combines conventional Belief-Propagation (BP) algorithm with stochastic to achieve high performance and low complexity simultaneously. However, lossy inefficient stochastic-to-binary (S2B) conversion brings extra degradation latency. In this paper, a bit-serial updating based hybrid (BSSU-HD) is proposed, which employs fully correlated (FCS) check nodes (CNs) probability tracers assisted variable (VNs) accomplish accurate efficient S2B...

In order to solve the problem of variable steady-state operation nodes and poor coordination control effect in photovoltaic energy storage plants, strategy plants based on ADP is studied. Establish power station model including system model, super capacitor battery model; Set maximum limit active change as constraint condition for coordinated station; The optimal multi voltage reactive resource fully considered, established by using algorithm, obtained online learning collected dynamic...

Abstract In general, the trajectory tracking of robotic manipulator is exceptionally challenging due to complex and strongly coupled mechanical architecture. this paper, precise track control formulated as a dense reward problem for reinforcement learning(RL). A deep RL(DRL) approach combining soft actor-critic (SAC) algorithm ensemble random network distillation (ERND) proposed address manipulator. Firstly, an ERND model designed, consisting module list multiple RND models. Each obtains...

In this paper, a deep reinforcement learning (DRL) approach based on generative adversarial imitation (GAIL) and long short-term memory (LSTM) is proposed to resolve tracking control problems for robotic manipulators with saturation constraints random disturbances, without the dynamic kinematic model of manipulator. Specifically, it limits torque joint angle certain range. Firstly, in order cope instability problem during training obtain stability policy, soft actor–critic (SAC) LSTM are...

Summary This paper addresses the control issue for cooperative visual servoing manipulators on strongly connected graph with communication delays, in which case that uncertain robot dynamics and kinematics, uncalibrated camera model, actuator constraint are simultaneously considered. An adaptive image‐based approach is established to overcome difficulty arising from nonlinear coupling between model agents. To estimate coupled camera‐robot parameters, a novel strategy developed its...