.png){kind=icon}
.png){kind=icon}
Return
Minimum-cost parameter blocking attacks on cyber-physical systems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jfranklin.2026.108930.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper investigates the minimum-cost parameter blocking attacks (PBA) problem, from the attackers’ perspective, to compromise the observability in cyber-physical systems (CPS). We consider two attack scenarios based on the attacker’s capabilities: 1) Minimum-Cost Problem under Output PBA (MCP-OPBA), where the attacker can arbitrarily block columns of the output matrix to disrupt observability, and 2) Minimum-Cost Problem under Mixed PBA (MCP-MPBA), where limited output column blocking is combined with state parameter perturbations when output-only PBAs are insufficient. Regarding MCP-OPBA, we first establish an algorithm by leveraging eigenbasis analysis to identify the optimal attack sets for special cases (e.g., identity output matrices with bounded eigenvalue multiplicities). For general cases, a polynomial-time variant is proposed, which significantly reduces computational complexity via rank-deficient submatrix detection. For MCP-MPBA, the problem is cast as an equality-constrained optimization framework, integrating output column blocking and state parameter perturbations. Using Lagrangian relaxation, we derive necessary optimality conditions and employ a Newton iterative algorithm to solve the resulting non-convex optimization. Theoretical analyses and numerical examples demonstrate the effectiveness of minimizing attack cost while destroying system observability.
Keywords:
Observability
Minimum cost
Parameter blocking attack
Cyber-physical systems
Journal
J
IF:
3.7
Papers:
6.3K
Citations:
1.5W