Neural Network Control System Research Articles

Complete Controllability of Nonlinear Neural Network Control Systems

Complete Controllability of Nonlinear Neural Network Control Systems

Tensor product model transformation-based reinforcement learning neural network controller with guaranteed stability

Despite the success of neural network (NN) controllers, the stability of closed-loop NN control systems is still a major challenge. This paper proposes a constrained reinforcement learning (RL) algorithm to design optimal NN controllers for a class of nonlinear dynamical systems while guaranteeing the stability of the control systems. By representing a controlled system as a tensor product (TP) model, a sufficient stability condition for a closed-loop NN control system is derived, which is then employed as a training constraint for the NN controller throughout the training process to ensure the stability in each training step. By showing the existence of a Lyapunov function, which is also directly obtained during the training process, the stability can be theoretically confirmed. Simulation studies on various nonlinear dynamical systems illustrate that the NN controllers trained with the proposed constrained RL algorithm outperform linear–quadratic–regulator controllers and yield roughly the same performance index values as those trained with an unconstrained RL algorithm, indicating that the imposed constraint does not significantly affect the training performance. Most importantly, while simulation results show that all controllers provide closed-loop stability, only the NN controllers trained with the proposed constrained RL algorithm theoretically guarantee stability.

Using Convolutional Neural Network to Design and Predict the Forces and Kinematic Performance and External Rotation Moment of the Hip Joint in the Pelvis

In order to improve the dynamic and kinematic adaptability of the hip joint, this paper presented a control attitude and kinematics and torque of the hip joint with power based neural network control. The CNN neural network uses input data only from the limb designed by the medical software, and is trained by different natural and artificially altered step patterns of healthy individuals. This type of network has been used for deep learning to realize adaptive speed control, dynamic and motion attitude, as well as prediction of force and torque performance. Detailed movement and torque tests were performed using MIMICS and ANATOMY AND PHYSIOLOGY software, and the obtained data were checked and varied by a healthy person, and finally, the test results showed that the neural network control system was able to control the selection. It has a variable and high speed with proper adaptation in various conditions. Finally, MATLAB software was used to design and predict the data of the problem, and favorable results were obtained.

Provable observation noise robustness for neural network control systems

Abstract Neural networks are vulnerable to adversarial perturbations: slight changes to inputs that can result in unexpected outputs. In neural network control systems, these inputs are often noisy sensor readings. In such settings, natural sensor noise – or an adversary who can manipulate them – may cause the system to fail. In this paper, we introduce the first technique to provably compute the minimum magnitude of sensor noise that can cause a neural network control system to violate a safety property from a given initial state. Our algorithm constructs a tree of possible successors with increasing noise until a specification is violated. We build on open-loop neural network verification methods to determine the least amount of noise that could change actions at each step of a closed-loop execution. We prove that this method identifies the unsafe trajectory with the least noise that leads to a safety violation. We evaluate our method on four systems: the Cart Pole and LunarLander environments from OpenAI gym, an aircraft collision avoidance system based on a neural network compression of ACAS Xu, and the SafeRL Aircraft Rejoin scenario. Our analysis produces unsafe trajectories where deviations under $1{\rm{\% }}$ of the sensor noise range make the systems behave erroneously.

Development of control system of isopentane-isoamylene fraction rectification column using neural network

This article presents the development of a neural network control system of a distillation column for the separation of isopentane-isoamylene fraction in order to increase the efficiency of control of rectification processes using intelligent technologies. The feasibility of using intelligent technologies to control the parameters of the column by using a neural network regulator is justified. The processing of research results was carried out using the software “MatLab.”

Reachability Analysis of Neural Network Control Systems With Tunable Accuracy and Efficiency

Reachability Analysis of Neural Network Control Systems With Tunable Accuracy and Efficiency

Stability of tracking wheel mobile robot with teleoperation fuzzy neural network control system

The stability of the Tracking Wheel Mobile Robot with Teleoperation System and Path Following Method is discussed in this study. The path is to be tracked by the host computer which is the master robot. The response from the robot is captured on camera. As the slave robot approaches the target position, the camera captures the response robot’s position and as well as moving trajectory. The host computer receives all of the images, enabling mobile robot deviation recoveries. The slave robot can use teleoperation to follow the sensor based on the decisions made by the master robot. The Lyapunov function in the Fuzzy Neural Network (FNN) control structure assures the system’s stability and satisfactory performance. It supports a mobile robot’s ability to adhere to a reference trajectory without deviating from it. Finally, the outcome of the simulation demonstrates that our controller is capable of tracking different environmental conditions and maintaining stability.

Local Stability and Convergence Analysis of Neural Network Controllers With Error Integral Inputs.

This article investigates the local stability and local convergence of a class of neural network (NN) controllers with error integrals as inputs for reference tracking. It is formally proved that if the input of the NN controller consists exclusively of error terms, the control system shows a non-zero steady-state error for any constant reference except for one specific point, for both single-layer and multi-layer NN controllers. It is further proved that adding error integrals to the input of the (single- and multi-layers) NN controller is one sufficient way to remove the steady-state error for any constant reference. Due to the nonlinearity of the NN controllers, the NN control systems are linearized at the equilibrium points. We provide proof that if all the eigenvalues of the linearized NN control system have negative real parts, local asymptotic stability and local exponential convergence are guaranteed. Two case studies were explored to verify the theoretical results: a single-layer NN controller in a 1-D system and a four-layer NN controller in a 2-D system applied to renewable energy integration. Simulations demonstrate that when NN controllers and the corresponding generalized proportional-integral (PI) controllers have the same eigenvalues, all control systems exhibit almost the same responses in a small neighborhood of their respective equilibrium points.

Reachability Analysis of Neural Network Control Systems

Neural network controllers (NNCs) have shown great promise in autonomous and cyber-physical systems. Despite the various verification approaches for neural networks, the safety analysis of NNCs remains an open problem. Existing verification approaches for neural network control systems (NNCSs) either can only work on a limited type of activation functions, or result in non-trivial over-approximation errors with time evolving. This paper proposes a verification framework for NNCS based on Lipschitzian optimisation, called DeepNNC. We first prove the Lipschitz continuity of closed-loop NNCSs by unrolling and eliminating the loops. We then reveal the working principles of applying Lipschitzian optimisation on NNCS verification and illustrate it by verifying an adaptive cruise control model. Compared to state-of-the-art verification approaches, DeepNNC shows superior performance in terms of efficiency and accuracy over a wide range of NNCs. We also provide a case study to demonstrate the capability of DeepNNC to handle a real-world, practical, and complex system. Our tool DeepNNC is available at https://github.com/TrustAI/DeepNNC.

Modeling of a neural network-based system for identification and control of technical object parameters

The article analyzes the effectiveness of a neural network control system for main-taining the pH level in the feedwater of a steam boiler. An intelligent control system im-plements the principle of reverse error propagation through a neural emulator. The sub-system model of steam boiler water tube blowing was used as the research object. The neural network controller and neural emulator were trained on a control system model with a PID controller using the expert correction methodology of tuning coefficients: proportionality, integration constant, and differentiation based on the analysis of tran-sient process quality indicators. The analysis of the transient processes obtained from simulation modeling allows us to state that the trained neural network control system successfully compensates for disturbances over a wide range of changes in the object's parameter values via control channels and disturbances (simulating changes in steam load).

Motion Trajectory Error of Robotic Arm Based on Neural Network Algorithm

In order to solve the problems of unstable motion and large trajectory tracking error of the manipulator when it is disturbed by the outside world, the author proposes an adaptive neural network manipulator motion trajectory error method. The author gives the dynamic equation of the manipulator and uses the positive feedback neural network to study the dynamic characteristics of the manipulator. An adaptive neural network control system is designed, and the stability and convergence of the closed-loop system are proved by the Lyapunov function. A schematic diagram of the manipulator model is established, and MATLAB/Simulink software is used to simulate the dynamic parameters of the manipulator. At the same time, it is compared and analyzed with the simulation results of the PID control system. Simulation results show that in robot arm 3, the expected motion trajectory is θ3 = 0.4cos(2πt), the initial condition θ(0) = [000]τ, the control parameter K = diag(40,40),40), the disturbance parameter τ’ = 20cos(πt), robot arm link parameters l1 = 0.62 m, l2 = 0.41 m, l3 = 0.34 m, m1 = 3.5, m2 = 2.5 kg, m3 = 2 kg, g = 9.82 m/s2, under t = 2s, the motion trajectory of the robotic arm is disturbed by the outside world, and the adaptive neural network is used to control the motion trajectory with a small tracking error, input torque ripple is small. Conclusion. The manipulator adopts the adaptive neural network control method, which can improve the control accuracy of the motion trajectory and weaken the jitter phenomenon of the manipulator motion.

Analysis of fault ride through the improvement of PV power plant based on artificial neural network

The increasing adoption of renewable energy sources in industrial areas, driven by the need to both obtain a grid voltage from renewable energy sources and reduce high electricity bills, has led to an increased demand for solutions that can improve the power quality of these systems. A photovoltaic power plant (PVPP) based on a dc-chopper circuit protection system is proposed for the enhancement of the power quality of PVPP. The aim of the presented protection method is to protect the panels in the PVPP from the harmful effects of the high dc voltage that occurs during grid faults. The adaptive neural network control system is applied in both the T-type inverter and the dc-chopper circuit protection system. The proposed protection system is implemented with a 250 kW PVPP to verify its validity with simulation results during grid fault. This approach can potentially improve the reliability and stability of PVPP, and contribute to the maintain of PVPP into the grid.

Robust Neurocontrol for Autonomous Dynamic Soaring

The flight endurance of small unmanned aerial vehicles can be significantly extended through the exploitation of naturally occurring wind phenomena. However, due to the limited computational hardware on board such aircraft and the uncertain, stochastic nature of real-world environments, there is a need for efficient and robust strategies that exhibit generalized behavior. In addressing these objectives, recent efforts have explored the use of artificial intelligence training algorithms and neural networks for the design of autonomous control schemes that exploit such wind phenomena. This study incorporates the Neuroevolution of Augmenting Topologies algorithm with domain randomization to train robust neurocontrollers that can control an aircraft along sustained traveling dynamic soaring trajectories in the presence of uncertainties and disturbances. This work presents the developed strategy for integrating robustness in neural network control systems, provides a method for quantifying and comparing robustness, and introduces an approach for identifying the network characteristics that contribute to the evolved robust behavior.

АНАЛІЗ СИСТЕМИ ПАРАЛЕЛЬНОГО НЕЙРОУПРАВЛІННЯ ДИНАМІЧНИМИ ОБ'ЄКТАМИ

The article analyzes the effectiveness of the neural network control system, which together with the PIDcontroller implements the principle of parallel control of a dynamic object. As a rule, most industrialfacilities are characterized by non-linear dependencies, the presence of uncontrolled noise anddisturbances, frequent changes in equipment operating modes, and the presence of significant non-linearities. The model of the blowing subsystem of a water-tube steam boiler was used as an object ofresearch. The training of the neural network controller (NMC) and neuroemulator (emulator) was carriedout on the ACS model with a PID controller using the method of expert adjustment of tuning coefficients:proportionality, constant integration and differentiation based on the analysis of the quality indicators ofthe transition process. The change in the values of the object model parameters along the control anddisturbance channels corresponded to the dynamic modes of operation of the steam boiler in the range ofsteam load (25-110%) from the nominal one. The analysis of transient processes obtained on the basis ofcomputer modeling allows us to assert that the trained neural network control system compensates fordisturbances over the entire range of changes in the values of the object parameters along the control anddisturbance channels (simulation of changes in the steam load), as well as when the parameter values ofthe models go beyond the range study sample. Thus, the neural network controller can successfully perform the functions of an adaptive circuit tuned tothe most unfavorable disturbances in the ACS of parallel action by a complex production facility. And theimplementation of a neural network system of parallel action together with typical regulators in thetechnological processes of heat energy can reduce emergency situations associated with frequent changesin the steam load of power units caused by military actions in our country.

Fixed-Time Neural Control of Robot Manipulator With Global Stability and Guaranteed Transient Performance

Nowadays, due to many limitations in reality, the optimization of tracking precision and convergence time has attracted the attention of researchers in robotics community. In this article, a fixed-time adaptive neural network (NN) controller is proposed for unknown robot manipulators. A switching mechanism is integrated into the control design such that the semiglobal stability of the conventional NN control systems can be extended to global stability. The time-varying barrier Lyapunov function and the fixed-time control technique are incorporated into the controller design to guarantee the prescribed motion constraints and the fixed-time convergence simultaneously. Compared with some existing fixed-time NN control algorithms, the assumption that NN weight should be upper bounded can be relaxed in our work. Finally, the simulation and experiment studies are respectively carried out based on an unknown 2-degree of freedom robot and a Baxter robot to demonstrate the effectiveness and superiority of the proposed control scheme.

Permanent magnet synchronous motor vector control based on BP neural network

In this paper, the traditional permanent magnet synchronous motor vector control system has problems such as low control precision, low reliability, poor system operation status. To solve these problems, because of the neural network has self-learning and self-adaptive ability to BP algorithm can be applied to parts of speed control system. This paper proposes a PMSM which based on BP neural network algorithm and control systems. Then the system model is established in Simulink environment, and the superiority and feasibility of the strategy are verified by comparing with traditional PID controller.

NExG: Provable and Guided State-Space Exploration of Neural Network Control Systems Using Sensitivity Approximation

We propose a new technique for performing state-space exploration of closed-loop control systems with neural network feedback controllers. Our approach involves approximating the sensitivity of the trajectories of the closed-loop dynamics. Using such an approximator and the system simulator, we present a guided state-space exploration method that can generate trajectories visiting the neighborhood of a target state at a specified time. We present a theoretical framework which establishes that our method will produce a sequence of trajectories that will reach a suitable neighborhood of the target state. We provide a thorough evaluation of our approach on various systems with neural network feedback controllers of different configurations. We outperform earlier state-space exploration techniques and achieve significant improvement in both the quality (explainability) and performance (convergence rate). Finally, we adopt our algorithm for the falsification of a class of temporal logic specification, assess its performance, and show its potential in supplementing existing falsification algorithms.

Evaluation of Neural Network Verification Methods for Air-to-Air Collision Avoidance

Neural network approximations have become attractive to compress data for automation and autonomy algorithms for use on storage-limited and processing-limited aerospace hardware. However, unless these neural network approximations can be exhaustively verified to be safe, they cannot be certified for use on aircraft. An example of such systems is the unmanned Airborne Collision Avoidance System (ACAS) Xu, which is a very popular benchmark for open-loop neural network control system verification tools. This paper proposes a new closed-loop extension of this benchmark, which consists of a set of 10 closed-loop properties selected to evaluate the safety of an ownship aircraft in the presence of a co-altitude intruder aircraft. These closed-loop safety properties are used to evaluate five of the 45 neural networks that comprise the ACAS Xu benchmark (corresponding to co-altitude cases) as well as the switching logic between the five neural networks. The combination of nonlinear dynamics and switching between five neural networks is a challenging verification task accomplished with star-set reachability methods in two verification tools. The safety of the ownship aircraft under initial position uncertainty is guaranteed in every scenario proposed.

Modeling of Mobility as a Service (MaaS) Collaborative Dispatching System of Railway Passenger Transport Hub Based on Neural Network Algorithm

In order to solve the modeling problem of travel as a service (MaaS) collaborative dispatching system of railway passenger transport hub based on neural network algorithm, meet people’s needs, make up for the lack of high traffic travel pressure, and improve people’s living standards, through 30 random questionnaires, it is found that 28 people think that travel convenience improves their quality of life, and 2 people think that owning a car has little effect on travel convenience. Travel has always been people’s basic living needs. In recent years, with the improvement of people’s living standards, there are more and more urban vehicles. At the same time, the increase of vehicles also directly leads to the increasing pressure of urban traffic travel, and the problems of vehicle emission pollution and traffic congestion are becoming more and more obvious. With the national low-carbon environmental protection policies, green transportation has become the theme of the times. With the continuous development of shared transportation mode and intelligent information technology, the new transportation concept of “Mobility as a Service” based on this technical mode will highly integrate the existing transportation modes and travel services. With the support of MaaS system, people’s travel modes will change greatly. Starting with the MaaS cooperative dispatching system of railway passenger transport hub, a multiparameter fuzzy neural network control system dispatching algorithm is proposed to better help the modeling of MaaS cooperative dispatching system of railway passenger transport hub.

Applying distributed ledger technologies in megacities to face anthropogenic burden challenges

Distributed ledger technologies can support a rapid transition to smart cities and provide a high level of urban quality. Despite the large number of approaches to the problem of synthesizing smart city management systems, there is still no universal solution. One of the most promising areas is the construction of neural network control systems. The optimization module for a neurosimulator is developed that can operate in real time. The study of the neurosimulator on various data of anthropogenic load showed the possibility of obtaining high control accuracy.