Common use of Fault Tolerance Clause in Contracts

Fault Tolerance. Due to the presence of noises and manufacture variations, there may be a difference of CSI measurements hi in the ith sample, denoted as δi. When δ is larger than ε, ∆σˆ begins to incur mismatched bits, which leads to a wrong information delivery. Using multiple samples in a block can ε2 reduce the variance of the represented features. According to ▇▇▇▇▇▇▇▇▇ inequality, we have P{|δ −E(δ)| ≥ ε} ≤ D(δ) . Block-based information delivery can efficiently reduce the variance of average δ, and then reduce the secret bit error rate. Σ TDS extracts the feature of block based on SVD. As afore- mentioned in Section 3.2, the block size is 10 n (typi- cally n = 6). SVD can be expressed as G = U ΣˆV T = Serial 0.421 0.590 0.401 0.530 0.841 0.913 0.885 0.642 Table 2: NIST statistical test results. To pass this test, p-value must be greater than 0.01. i=1 σˆiUiV T , where σˆ is the singular value of G, and Ui, Σ Vi are the ith column vectors of U and V , respectively. The power of noise is PN = β (σw)2, where σw is the ith sin- Σby gular value of noise matrix. TDS uses the second or third singular values σˆ2 and σˆ3 to represent the signal features and discards the singular value smaller than σˆ4 which are mainly relevant to noises. Therefore, the noise is decreased (σw)2 through SVD.

Fault Tolerance. ε2 Due to the presence of noises and manufacture variations, there may be a difference of CSI measurements hi in the ith sample, denoted as δi. When δ is larger than ε, ∆σˆ Δσˆ begins to incur mismatched bits, which leads to a wrong information delivery. Using multiple samples in a block can ε2 reduce the variance of the represented features. According to ▇▇▇▇▇▇▇▇▇ inequality, we have P{|δ −E(δ)| ≥ ε} ≤ D(δ) . Block-based information delivery can efficiently reduce the variance of average δ, and then reduce the secret bit error rate. Σ TDS extracts the feature of block based on SVD. As afore- mentioned in Section 3.2, the block size is 10 n (typi- cally n = 6). SVD can be expressed as G = U ΣˆV T = Index State Environment A Static Indoor C Mobile Indoor D Mobile Outdoor Serial 0.421 0.590 0.401 0.530 0.841 0.913 0.885 0.642 Table 2: NIST statistical test results. To pass this test, p-value must be greater than 0.01. i=1 σˆiUiV T , where σˆ is the singular value of G, and Ui, Σ Vi are the ith column vectors of U and V , respectively. The power of noise is PN = β (σw)2, where σw is the ith sin- Σby gular value of noise matrix. TDS uses the second or third singular values σˆ2 and σˆ3 to represent the signal features and discards the singular value smaller than σˆ4 which are mainly relevant to noises. Therefore, the noise is decreased (σw)2 through SVD.