Erratum to "Anonymous Tracking using RFID tags" (INFOCOM 2007)

Generally, the following accuracy metric is used for the counting estimators: probability that the estimation error is in a given interval, with a certain confidence; i.e., P [ ≤ θ] ≥ α, = ∣∣∣ t−t̂ t ∣∣∣, where t is the true and t̂ is the estimated number of tags, respectively. The accuracy metric used in [1] is Zα √ δ0 ≤ β t (see [1], page 1222), where Zα = √ 2 erf−1(2α−1) is the α-quantile of unit normal distribution, δ0 is the estimation variance. In the following, we provide the corrected form of this relation. The estimation t̂ was shown to have a Gaussian distribution with mean t and variance δ0. The left-hand-side of the accuracy relation can then be written as: