A Data-Based Feedback Relearning Algorithm for Uncertain Nonlinear Systems

In this paper,a data-based feedback relearning algorithm is proposed for the robust control problem of uncertain nonlinear systems.Motivated by the classical on-policy and off-policy algorithms of reinforcement learning,the online feedback relearning(FR)algorithm is developed where the collected data includes the influence of disturbance signals.The FR algorithm has better adaptability to environmental changes(such as the control channel disturbances)compared with the off-policy algorithm,and has higher computational efficiency and better convergence performance compared with the on-policy algorithm.Data processing based on experience replay technology is used for great data efficiency and convergence stability.Simulation experiments are presented to illustrate convergence stability,optimality and algorithmic performance of FR algorithm by comparison.

Tracking control for air-breathing hypersonic cruise vehicle based on tangent linearization approach

This paper is focused on developing a tracking controller for a hypersonic cruise vehicle using tangent linearization approach.The design of flight control systems for air-breathing hypersonic vehicles is a highly challenging task due to the unique characteristics of the vehicle dynamics.Motivated by recent results on tangent linearization control,the tracking control problem for the hypersonic cruise vehicle is reduced to that of a feedback stabilizing controller design for a linear time-varying system which can be accomplished by a standard design method of frozen-time control.Through a proper model transformation,it can be proven that the tracking error of the designed closed-loop system decays exponentially.Simulation studies are conducted for trimmed cruise conditions of 110000 ft and Mach 15 where the responses of the vehicle to step changes in altitude and velocity are evaluated.The effectiveness of the controller is demonstrated by simulation results.

Analytical reentry guidance framework based on swarm intelligence optimization and altitude-energy profile

Aimed at improving the real-time performance of guidance instruction generation,an analytical hypersonic reentry guidance framework is presented.The key steps of the novel guidance framework are the parameterization of reentry guidance problems and the optimization of parameters.First,a quintic polynomial function of energy was designed to describe the altitude profile.Then,according to the altitude-energy profile,the altitude,velocity,flight path angle,and bank angle were obtained analytically,which naturally met the terminal constraints.In addition,the angle of the attack profile was determined using the velocity parameter.The swarm intelligent optimization algorithms were used to optimize the parameters.The path constraints were enforced by the penalty function method.Finally,extensive simulations were carried out in both nominal and dispersed cases,and the simulation results showed that the proposed guidance framework was effective,high-precision,and robust in different scenarios.

基于安全自适应强化学习的自主避障控制方法

障碍规避是无人机等自主无人系统运动规划的重要环节,其核心是设计有效的避障控制方法.为了进一步提高决策优化性和控制效果,本文在最优控制的设定下,提出一种基于强化学习的自主避障控制方法,以自适应方式在线生成安全运行轨迹.首先,利用障碍函数法在代价函数中设计了一个光滑的奖惩函数,从而将避障问题转换为一个无约束的最优控制问题.然后,利用行为–评价神经网络和策略迭代法实现了自适应强化学习,其中评价网络利用状态跟随核函数逼近代价函数,行为网络给出近似最优的控制策略;同时,通过状态外推法获得模拟经验,使得评价网络能利用经验回放实现可靠的局部探索.最后,在简化的无人机系统和非线性数值系统上进行了仿真实验与方法对比,结果表明,提出的避障控制方法能实时生成较优的安全运行轨迹.