Optimal Tampering Attack Strategy for FIR System Identification With Multi-Level Quantized Observations

This paper considers the optimal tampering attack strategy in system identification of Finite Impulse Response (FIR) systems with multi-level quantized observations under data tampering attacks. First, the data tampering attack model based on a multi-level quantization system is established in a conditional probability manner according to the features of the quantization system. Second, a multi-parameter-based system parameter estimation algorithm is designed and its convergence consistency is proved. Then, according to the convergence of the designed identification algorithm under a tampering attack, the infinite paradigm of the difference between the converged value and the actual parameter after the attack is used as the attack index, and the optimal tampering attack strategy is designed to destroy the consistency of the recognition algorithm so as to make the best attack effect achieved. Finally, numerical simulation experiments under different conditions are used to verify the result.