Saturated Output-Feedback Hybrid Reinforcement Learning Controller for Submersible Vehicles Guaranteeing Output Constraints

Omid Elhaki,Allahyar Montazeri,Khoshnam Shojaei,Declan Shanahan

doi:10.1109/access.2021.3113080

Abstract

In this brief, we propose a new neuro-fuzzy reinforcement learning-based control (NFRLC) structure that allows autonomous underwater vehicles (AUVs) to follow a desired trajectory in large-scale complex environments precisely. The accurate tracking control problem is solved by a unique online NFRLC method designed based on actor-critic (AC) structure. Integrating the NFRLC framework including an adaptive multilayer neural network (MNN) and interval type-2 fuzzy neural network (IT2FNN) with a high-gain observer (HGO), a robust smart observer-based system is set up to estimate the velocities of the AUVs, unknown dynamic parameters containing unmodeled dynamics, nonlinearities, uncertainties and external disturbances. By employing a saturation function in the design procedure and transforming the input limitations into input saturation nonlinearities, the risk of the actuator saturation is effectively reduced together with nonlinear input saturation compensation by the NFRLC strategy. A predefined funnel-shaped performance function is designed to attain certain prescribed output performance. Finally, stability study reveals that the entire closed-loop system signals are semi-globally uniformly ultimately bounded (SGUUB) and can provide prescribed convergence rate for the tracking errors so that the tracking errors approach to the origin evolving inside the funnel-shaped performance bound at the prescribed time.

Highlights

The global offshore remotely operated vehicles (ROVs) industry is predicted to worth up when the area is inaccessible to the human or will impose to $3.5Bn with modern applications including civil and intremendous danger to the operator to work on site
Autonomy of the manipulator when it interacts with the this includes but not limited to finite scape time, multiple surrounding environment without human intervention [3, 4]. equilibrium points, limit cycles, appearing new frequencies, Another useful robot to operate in nuclear applications is bifurcation, chaos, unmodeled dynamics, uncertainties, etc
PPC is provided if ηli(t) ≤ ei(t) ≤ ηui(t) be true where ei is ith element of e, ηli and ηui are the pre-set limits of ei, and ηi is defined as ηi(t) := exp(−ait) + ηi∞ where ai ∈ + is a lower bound on the convergence rate, ηi0 > ηi∞ ∈ + are design factors, ηi∞ is an upper bound

Summary

Utilizing and gives

Which show a strictly increasing relation between ei and ei. using e = q − qd, (2), and (9) givese =Dν + κ,. The input saturation nonlinearity dτ (τ ) = τs − τ , that is a NLIP term and will be compensated in the control design procedure, is given by where D = T Jt, κ( e, ηu, ηl, ηu, ηl, qd, qd) = Ψ−T qd, T =. PPC is provided if ηli(t) ≤ ei(t) ≤ ηui(t) be true where ei is ith element of e, ηli and ηui are the pre-set limits of ei, and ηi is defined as ηi(t) := (ηi0 − ηi∞) exp(−ait) + ηi∞ where ai ∈ + is a lower bound on the convergence rate, ηi0 > ηi∞ ∈ + are design factors, ηi∞ is an upper bound. The PBs can be set as ηli = 228 −αiηi, ηui = βiηi where αi, βi ∈ + are tunable factors. Deal with the dimensionality problem, and the critic agent is an interval type-2 fuzzy neural network that can handle

Prescribed performance Transformation

Kp ko

Proposed controller

Λε e