We introduce the class of multiply constant-weight codes to improve the reliability of certain physically unclonable function (PUF) response. We extend classical coding methods to construct multiply constant-weight codes from known q-ary and constant-weight codes. Analogues of Johnson bounds are derived and are shown to be asymptotically tight to a constant factor under certain conditions. We also examine the rates of the multiply constant-weight codes and interestingly, demonstrate that these rates are the same as those of constant-weight codes of suitable parameters. Asymptotic analysis of our code constructions is provided.

No abstract

No abstract

Coupled switched capacitance causes crosstalk in ultra deep submicron/nanometer VLSI fabrication, which leads to power dissipation, delay faults, and logical malfunctions. We present the first memoryless transition bus-encoding technique for power minimization, errorcorrection, and elimination of crosstalk simultaneously. To accomplish this, we generalize balanced sampling plans avoiding adjacent units, which are widely used in the statistical design of experiments. Optimal or asymptotically optimal constant weight codes eliminating each kind of crosstalk are constructed.

No abstract