Optimization and Systems Theory Seminar
October 14, at 11.00, room 3721, Lindstedtsvägen 25, KTH:
Institute of Systems Science,
Chinese Academy of Sciences.
System Identification with Quantized Observations
Quantized systems are widely applied in practical fields, but the
study on the identification of the system is just started. The main
reason is that the system output cannot be measured accurately, and
the information can be applied to identification is quantized
information, which is only the comparison between the system output
and thresholds of quantized sensors. The limit information leades
to great difficulties for quantized system identification.
This talk studies the system identification with quantized
observations under fixed thresholds. Identification algorithm is
proposed for parameter identification, and proved to be both
mean-square and almost surely convergent. In order to develop
the efficient algorithm, the Cramer Rao lower bound of parameter
is obtained, based on which an asymptotically efficient algorithm is
Calendar of seminars
Last update: May 17, 2011.