An improved upper bound on the contribution by intersymbol interference to the probability of error in digital systems

In the transfer of digital data through a linear, time-invariant channel, two major sources of distortion are present. These are additive Gaussian noise and intersymbol interference, whose effects are commonly measured by the probability of error in the data transfer. The contribution to the error rate by additive Gaussian noise is well known, but the contribution by intersymbol interference is still in general unknown. This paper considers the contribution to the probability of error by intersymbol interference through an investigation of upper bounds to the probability distribution function associated with this distortion. Previous endeavors in this area are discussed and improvements are demonstrated. Regions of improvement and examples are presented.