Reachable Set Estimation for Neural Network Control Systems: A Simulation-Guided Approach
نویسندگان
چکیده
The vulnerability of artificial intelligence (AI) and machine learning (ML) against adversarial disturbances attacks significantly restricts their applicability in safety-critical systems including cyber-physical (CPS) equipped with neural network components at various stages sensing control. This article addresses the reachable set estimation safety verification problems for dynamical embedded serving as feedback controllers. closed-loop system can be abstracted form a continuous-time sampled-data under control controller. First, novel computation method adaptation to simulations generated out networks is developed. reachability analysis class feedforward called multilayer perceptrons (MLPs) general activation functions performed framework interval arithmetic. Then, combination methods developed classes modeled by ordinary differential equations, recursive algorithm over-approximating system. examining emptiness intersection between over-approximation sets unsafe sets. effectiveness proposed approach has been validated evaluations on robotic arm model an adaptive cruise
منابع مشابه
Reachable Set Estimation and Safety Verification for Piecewise Linear Systems with Neural Network Controllers
In this work, the reachable set estimation and safety verification problems for a class of piecewise linear systems equipped with neural network controllers are addressed. The neural network is considered to consist of Rectified Linear Unit (ReLU) activation functions. A layer-by-layer approach is developed for the output reachable set computation of ReLU neural networks. The computation is for...
متن کاملReachable Set Estimation and Verification for Neural Network Models of Nonlinear Dynamic Systems
Neural networks have been widely used to solve complex realworld problems. Due to the complicate, nonlinear, non-convex nature of neural networks, formal safety guarantees for the behaviors of neural network systems will be crucial for their applications in safety-critical systems. In this paper, the reachable set estimation and verification problems for Nonlinear Autoregressive-Moving Average ...
متن کاملchannel estimation for mimo-ofdm systems
تخمین دقیق مشخصات کانال در سیستم های مخابراتی یک امر مهم محسوب می گردد. این امر به ویژه در کانال های بیسیم با خاصیت فرکانس گزینی و زمان گزینی شدید، چالش بزرگی است. مقالات متعدد پر از روش های مبتکرانه ای برای طراحی و آنالیز الگوریتم های تخمین کانال است که بیشتر آنها از روش های خاصی استفاده می کنند که یا دارای عملکرد خوب با پیچیدگی محاسباتی بالا هستند و یا با عملکرد نه چندان خوب پیچیدگی پایینی...
Reachable Set Bounding Estimation for Distributed Delay Systems with Disturbances
The reachable set bounding estimation for distributed delay systems with disturbances is a new problem. In this paper,we consider this problem subject to not only time varying delay and polytopic uncertainties but also distributed delay systems which is not studied fully untill now. we can obtain improved non-ellipsoidal reachable set estimation for neural networks with time-varying delay by th...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE transactions on neural networks and learning systems
سال: 2021
ISSN: ['2162-237X', '2162-2388']
DOI: https://doi.org/10.1109/tnnls.2020.2991090