Parameter identification and platoon control for virtually coupled train set with physics-informed neural network dynamics model

Neural dynamics change over time and with physiologic state. Modeling of neural dynamics can thus be understood at two levels: (i) identifying the latent process that governs how and when states change, and (ii) identifying the generative circuit mechanisms within each state. Here, we develop a data-driven modeling method that tackles these two levels simultaneously. We formulate a parametric network model of neural dynamics that embeds state-dependent modulation. The modulation itself is controlled by a latent switching process, modeled as a Hidden Markov Model (HMM). A key challenge is that the model itself has internal states that must be estimated from observed data. This leads to a triune optimization problem, consisting of model parameterization of the HMM and neural dynamics, alongside state estimation. Our method brings together several optimization frameworks alongside estimation-theoretic constructs to solve this problem efficiently, enabling blind identification of state transitions and neural dynamics. We validate this methodology on ground-truth data with known parameters, and find that it accurately infers the transitions in latent state and the dynamics of each state. We demonstrate its capability of inferring changes in brain dynamics from electrophysiological data by testing it on electroencephalography recordings with labeled state transitions. While similar methods exist to infer switches and dynamics on the level of individual neurons, there is no directly comparable method available for mesoscale modeling of neural circuits. Our methodology enables blind modeling of changing neural dynamics allowing for inference of modulatory circuit mechanisms.

A neural dynamics model for structural optimization—Theory

A neural system dynamics modeling platform and its applications in randomized controlled trial data analysis

The purpose of this work is the development of a hybrid guidance and control system for the trajectory tracking task of an underactuated Unmanned Surface Vehicle (USV). The novelty of the guidance system is based on neural dynamics and sliding mode integration. The biological neural dynamics model is designed to smooth the outputs of the sliding mode control, avoid sharp speed jumps and satisfy the thruster control constraint. The outputs of the sliding mode controller are used such as the reference inputs of an internal backstepping control loop for low level dynamics regulation. The hybrid control is focused towards real robot exploitation in such a way to damp sudden changes of the torque commands and consequently, mechanical stress onto the thrust motors. The performance of the proposed control strategy and the comparison between sliding mode guidance, and sliding mode control embeds the neural dynamics model are described by a number of simulative tests. Moreover, measurement noises and a model of the sea currents variation, in direction and velocity, are counted in the simulated trials in order to demonstrate the control robustness with respect to measurement and environmental disturbances that unavoidably affect maritime operations.

Traditional models of vehicle dynamics engineered from physical principles are usually simplified and assumed, resulting in the model cannot accurately reflect the actual dynamic characteristics of the vehicle under some working conditions, affecting the control accuracy and even safety. In view of this, inspired by the single-track model, this paper uses the data-driven methods to establish a new high-performance time-delay feedback neural network vehicle dynamics model. The feedback connection of a network can describe complex dynamics. The multi-time-step input of the state and control can include highly nonlinear and strong coupling characteristics of a vehicle. The test results of modeling accuracy show that the proposed model can achieve higher vehicle dynamics prediction accuracy than nonlinear vehicle model. Different from the traditional vehicle dynamics model, the proposed model has long-term memory cells, which can implicitly predict coefficient of friction and can be applied to different road conditions. Then, the trajectory tracking control algorithm is designed based on the proposed vehicle model. According to the steady-state steering assumption, the feedforward front wheel steering angle is calculated, and the steady-state sideslip angle is integrated into the steering feedback according to the reference path to realize the reference trajectory tracking control. Finally, Simulink/CarSim is used to conduct the simulation analysis under the double lane change conditions to evaluate the proposed control algorithm. The analysis results show that the control algorithm based on the proposed model can achieve an accurate tracking control effect of a vehicle at medium and high speeds, providing high-accuracy track tracking and good lateral stability of intelligent vehicles.

Dynamic neural network controller model of PEM fuel cell system

Ventral visual stream neural responses are dynamic, even for static image presentations. However, dynamical neural models of visual cortex are lacking as most progress has been made modeling static, time-averaged responses. Here, we studied population neural dynamics during face detection across three cortical processing stages. Remarkably,~30 milliseconds after the initially evoked response, we found that neurons in intermediate level areas decreased their responses to typical configurations of their preferred face parts relative to their response for atypical configurations even while neurons in higher areas achieved and maintained a preference for typical configurations. These hierarchical neural dynamics were inconsistent with standard feedforward circuits. Rather, recurrent models computing prediction errors between stages captured the observed temporal signatures. This model of neural dynamics, which simply augments the standard feedforward model of online vision, suggests that neural responses to static images may encode top-down prediction errors in addition to bottom-up feature estimates.

A categorical and Łukasiewicz-Topos framework for Łukasiewicz LM-Algebraic Logic models of nonlinear dynamics in complex functional systems such as neural networks, genomes and cell interactomes is proposed. Łukasiewicz Algebraic Logic models of genetic networks and signaling pathways in cells are formulated in terms of nonlinear dynamic systems with n-state components that allow for the generalization of previous logical models of both genetic activities and neural networks.

BackgroundRestoration of upper limb movements in subjects recovering from stroke is an essential keystone in rehabilitative practices. Rehabilitation of arm movements, in fact, is usually a far more difficult one as compared to that of lower extremities. For these reasons, researchers are developing new methods and technologies so that the rehabilitative process could be more accurate, rapid and easily accepted by the patient. This paper introduces the proof of concept for a new non-invasive FES-assisted rehabilitation system for the upper limb, called smartFES (sFES), where the electrical stimulation is controlled by a biologically inspired neural inverse dynamics model, fed by the kinematic information associated with the execution of a planar goal-oriented movement. More specifically, this work details two steps of the proposed system: an ad hoc markerless motion analysis algorithm for the estimation of kinematics, and a neural controller that drives a synthetic arm. The vision of the entire system is to acquire kinematics from the analysis of video sequences during planar arm movements and to use it together with a neural inverse dynamics model able to provide the patient with the electrical stimulation patterns needed to perform the movement with the assisted limb.MethodsThe markerless motion tracking system aims at localizing and monitoring the arm movement by tracking its silhouette. It uses a specifically designed motion estimation method, that we named Neural Snakes, which predicts the arm contour deformation as a first step for a silhouette extraction algorithm. The starting and ending points of the arm movement feed an Artificial Neural Controller, enclosing the muscular Hill's model, which solves the inverse dynamics to obtain the FES patterns needed to move a simulated arm from the starting point to the desired point. Both position error with respect to the requested arm trajectory and comparison between curvature factors have been calculated in order to determine the accuracy of the system.ResultsThe proposed method has been tested on real data acquired during the execution of planar goal-oriented arm movements. Main results concern the capability of the system to accurately recreate the movement task by providing a synthetic arm model with the stimulation patterns estimated by the inverse dynamics model. In the simulation of movements with a length of ± 20 cm, the model has shown an unbiased angular error, and a mean (absolute) position error of about 1.5 cm, thus confirming the ability of the system to reliably drive the model to the desired targets. Moreover, the curvature factors of the factual human movements and of the reconstructed ones are similar, thus encouraging future developments of the system in terms of reproducibility of the desired movements.ConclusionA novel FES-assisted rehabilitation system for the upper limb is presented and two parts of it have been designed and tested. The system includes a markerless motion estimation algorithm, and a biologically inspired neural controller that drives a biomechanical arm model and provides the stimulation patterns that, in a future development, could be used to drive a smart Functional Electrical Stimulation system (sFES). The system is envisioned to help in the rehabilitation of post stroke hemiparetic patients, by assisting the movement of the paretic upper limb, once trained with a set of movements performed by the therapist or in virtual reality. Future work will include the application and testing of the stimulation patterns in real conditions.

To improve the accuracy of soil heavy metal content prediction, this paper proposes a dynamic neural network optimization model (DNNOM). The model is based on a radial basis function neural network (RBFNN). The weights and bias of the output layer of the RBFNN were generated using the adaptive dynamic genetic optimization algorithm (ADGOA), and the center point of the hidden layer of the RBFNN was determined using an efficient density peak clustering algorithm (EDPC). An adaptive variance measure (AVM) was then used to generate the width vector of RBFNN hidden layer. The model was applied to the prediction soil heavy metal content in six new urban areas in Wuhan. Through comparison with support vector machine(SVM), light gradient boosting machine(LightGBM), RBFNN, and genetic algorithm optimizes the radial basis function neural network(GA-RBFNN), the experimental results demonstrate that the DNNOM is closer to the real value than the other four models, and the four error indicator values are also significantly lower than those of the other comparison models, which have higher prediction accuracy. Especially when compared with RBFNN, the MAPE and SMAPE of DNNOM have dropped by 3.98% and 3.9%, respectively.

Caenorhabditis elegans (C. elegans) is one of the simplest multi-celled organisms whose body is composed of about 1000 cells. This animal has only 302 neurons devoted to information processing, and its interconnecting structure has been completely revealed anatomically. This has given an advantage to using C. elegans to reveal primitive information processing mechanisms in neural circuits. For this reason, various neural models have been proposed; however these previous models did not consider the interactions among the nervous system, body, and environment. Accordingly, the neural mechanisms to direct body movement from acquired environmental information have not yet been completely revealed. Against this background, this paper proposes a simulation platform including environmental, neural, and body dynamics models. We then applied the proposed platform for simulation of chemotaxis. The results confirmed that the proposed models can simulate attractant behavior to NaCl.

Vision-based pose stabilization of nonholonomic mobile robots has received extensive attention. At present, most of the solutions of the problem do not take the robot dynamics into account in the controller design, so that these controllers are difficult to realize satisfactory control in practical application. Besides, many of the approaches suffer from the initial speed and torque jump which are not practical in the real world. Considering the kinematics and dynamics, a two-stage visual controller for solving the stabilization problem of a mobile robot is presented, applying the integration of adaptive control, sliding-mode control, and neural dynamics. In the first stage, an adaptive kinematic stabilization controller utilized to generate the command of velocity is developed based on Lyapunov theory. In the second stage, adopting the sliding-mode control approach, a dynamic controller with a variable speed function used to reduce the chattering is designed, which is utilized to generate the command of torque to make the actual velocity of the mobile robot asymptotically reach the desired velocity. Furthermore, to handle the speed and torque jump problems, the neural dynamics model is integrated into the above mentioned controllers. The stability of the proposed control system is analyzed by using Lyapunov theory. Finally, the simulation of the control law is implemented in perturbed case, and the results show that the control scheme can solve the stabilization problem effectively. The proposed control law can solve the speed and torque jump problems, overcome external disturbances, and provide a new solution for the vision-based stabilization of the mobile robot.

A novel tracking controller is proposed in this paper for real-time navigation of a nonholonomic mobile robot based on a neural dynamics model and a conventional backstepping model. By taking advantage of the features of the neural dynamics in a gated dipole model, the proposed controller is capable of generating smooth, continuous control signals to track a desired path. In addition, it can deal with arbitrarily large initial tracking errors. It resolves the velocity jump problem in conventional backstepping controllers resulted from sudden changes of tracking errors. The proposed approach is not very sensitive to model parameters and converges faster than a previous tracking control model. The effectiveness and efficiency of the proposed approach is demonstrated by a series of simulation and comparison studies.

In this work, a novel point stabilization control strategy for mobile robots which moves on uneven surface is presented. Firstly, sliding mode method is adopted to extend the nonlinear kenimatic control law to dynamic system, so that the robot is driven by torques. Then, to solve the speed and torque jump problem, the neural dynamics model is integrated into the presented controller. In addition, we utilize genetic algorithm (GA) to optimize the controller parameters for obtaining better stabilization performance. The stability of the proposed control system is analyzed by using Lyapunov theory. Finally, simulation results are given to illustrate the effectiveness of the proposed control scheme.

Online set-point optimisation cooperating with predictive control of a yeast fermentation process: A neural network approach