Path-tracking control of wheeled mobile robot (WMR) has gained a lot of research attention, primarily because of its wide applicability -for example intelligent wheelchairs, explorationassistant remote WMR. Recent increase in remote and autonomous operations/requirements for WMR has led to more and more use of IoT devices within the control loop. Consequently, providing interfaces for malicious interactions through false data injection attacks (FDIA). Moreover, optimization-based FDIAs have been shown to cause catastrophic consequences in feedback control systems while by-passing any residual-based monitoring system. Since these attacks target system measurement process, this paper focuses on the problem of improving the resiliency of dynamical observers against FDIA. Specifically, we propose an attack-resilient pruning algorithm which attempts to exclude compromised channels from being processed by the observer. The proposed pruning algorithm improves attacklocalization precision to 100% with high probability, which correspondingly improves the resiliency of the underlying UKF to FDIA. The improvements due to the developed resilient pruningbased observer is validated through a numerical simulation of a two-layer path-tracking control platform of differential-driven wheeled mobile robot (DDWMR) under FDIA.
NOMENCLATUREThe following notations and definitions are used throughout the whole paper: R, R n , R n×m denote the space of real numbers, real vectors of length n and real matrices of n rows and m columns respectively. R + denotes positive real numbers. Normal-face lower-case letters (e.g. x ∈ R) are used to represent real scalar, bold-face lower-case letter (e.g. x ∈ R n ) represents vectors, while normal-face upper case (e.g. X ∈ R n×m ) represents matrices. 1 represents all-ones vector. Let T ⊆ {1, . . . , n}, then for a matrix X ∈ R n×m , X T ∈ R |T |×m is the sub-matrix obtained by extracting the rows of X corresponding to the indices in T . T c denotes the complement of a set T and the universal set on which it is defined will be clear from the context. The symbol • denotes element-wise multiplication of two vectors and is defined as z = x • y, where z i = x i • y i . The symbol * denotes the convolution operator for vectors. supp(x) donotes the support of the vector x given by the set T = supp(x) = {i|x i = 0}. argsort ↓ (x) denotes a function that returns the sorted indices of vector x in descending order. The space of all square integrable signals is denoted by L 2 . The space of all point-wise bounded signals is denoted by L ∞ .