| 内容提要: |
Unknown quantities traditionally share little common
ground in control and communication areas. Often times,
unknown quantities in communication community are generally
modeled as stochastic noise for the convenience of leveraging
the well defined concepts such as entropy, Shannon information,
Markovness, independence within probability space, while control
community often operates in safety where the performance
index should be guaranteed at each instance so that the unknown
quantities are usually tacked by some bounded unknowns. Cyberphysical
systems (CPSs) tightly involve control, communication
and computing which urge forces us to cross the sophisticated
scenarios where both uncertain variable and processing noise
could coexist. Part I of this paper focuses on the issue of whether
or not the state of an underlying plant could be exponentially
observed or estimated when subjected to unknowns and noise. It
turns out to compute the maximum capacity by introducing the
notion of memoryless multiple quasi-access channel (M-MQAC).
It shows that the maximum capacity is the convex combination of
the underlying Shannon capacity C and its sibling the zero-zero
capacity C0. Specially, the cases that the plant either subjects
multiple uncertainties or noise are also discussed. Clearly, the
later is the usual multiple access channel within information
theory. |