TY - JOUR

T1 - Min-max Kalman filtering

AU - Yaesh, I.

AU - Shaked, U.

N1 - Funding Information:
This work was supported by the C&M Maus Chair at Tel Aviv University.

PY - 2004/11

Y1 - 2004/11

N2 - The problem of H∞-optimal state estimation of linear continuous-time systems that are measured with an additive white noise is addressed. The relevant cost function is the expected value of the standard H∞ performance index, with respect to the measurement noise statistics. The solution is obtained by applying the matrix version of the maximum principle to the solution of the min-max problem in which the estimator tries to minimize the mean square estimation error and the exogenous disturbance tries to maximize it while being penalized for its energy. The solution is given in terms of two coupled Riccati difference equations from which the filter gains are derived. In the case where an infinite penalty is imposed on the energy of the exogenous disturbance, the celebrated Kalman filter is recovered. In the stationary case, where all the signals are stationary, an upper-bound on the solutions of the coupled Riccati equations is obtained via a solution of coupled linear matrix inequalities. The resulting filter then guarantees a bound on the estimation error covariance matrix. An illustrative example is given where the velocity of a maneuvering target has to be estimated utilizing noisy measurements of the position.

AB - The problem of H∞-optimal state estimation of linear continuous-time systems that are measured with an additive white noise is addressed. The relevant cost function is the expected value of the standard H∞ performance index, with respect to the measurement noise statistics. The solution is obtained by applying the matrix version of the maximum principle to the solution of the min-max problem in which the estimator tries to minimize the mean square estimation error and the exogenous disturbance tries to maximize it while being penalized for its energy. The solution is given in terms of two coupled Riccati difference equations from which the filter gains are derived. In the case where an infinite penalty is imposed on the energy of the exogenous disturbance, the celebrated Kalman filter is recovered. In the stationary case, where all the signals are stationary, an upper-bound on the solutions of the coupled Riccati equations is obtained via a solution of coupled linear matrix inequalities. The resulting filter then guarantees a bound on the estimation error covariance matrix. An illustrative example is given where the velocity of a maneuvering target has to be estimated utilizing noisy measurements of the position.

KW - Kalman filter

KW - Min-max estimation

KW - Stochastic H Filtering

KW - Tracking

UR - http://www.scopus.com/inward/record.url?scp=4644366101&partnerID=8YFLogxK

U2 - 10.1016/j.sysconle.2004.04.008

DO - 10.1016/j.sysconle.2004.04.008

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

AN - SCOPUS:4644366101

SN - 0167-6911

VL - 53

SP - 217

EP - 228

JO - Systems and Control Letters

JF - Systems and Control Letters

IS - 3-4

ER -