An adaptive modular reduction based error-detection algorithm

In this paper, an adaptive modular reduction based error-detection algorithm is proposed, it is based on the research of the modular reduction model and the missing rate of residue code. In the proposed method, more than one moduli are used to detect the input before they being sent into the functional logic, then the best modulus is chosen accordingly to optimizing the error detection performance. The simulation shows that, for the frequently-used unit in signal processing, such as addition and multiplication, fault missing rate of the proposed method is lower than that of the regular method under the same circuit size. And the cost of the detection branch in the proposed method is lower than that in the regular method under the same fault missing rate.