CE 697 Advanced Data Adjustment Notes Summer 2016

• constraint applications: camera clusters
• alfred leick's 82 paper on min. constr. & derivation of in. constr.
• tutorial on mc & ic
• result of monte carlo simulation of global TS, overlaid with chisqr plot
• demo of RANSAC algorithm (data taken from wikipedia web page)
• another demo of RANSAC algo. 4 parameter transformation
• L1.m matlab file for L1 norm minimization
• L2.m matlab file for L2 norm minimization (LS)
• derive kalman filter from conventional LS
• graphical examples of error propagation & problems
• Lecture 01: review direct constraint solutions, elimination method
• Lecture 02: hw1, gen. lin. hyp., cross cov. matrix (see rev.)
• Lecture 03: hints for hw1, constr. w/added params., inner constr. intro.
• Lecture 04: min. constr., inner constr. C matrix w/ eigenvect. & point coords., geom. interp.
• Lecture 05: derive F TS, describe monte carlo technique, derive unified LS (linear)
• Lecture 06: notes HW3, HW2, linear ULS deriv. 1 (w/correct.),2,3,4, NL deriv., counting, sig0^2
• Lecture 07: blunder detect., Wbar, red. num., hypoth. test, type I,II, alpha, beta
• Lecture 08: baarda data snooping, set up global & indiv. hyp. tests, alpha, beta, int./ext. rel.
• Lecture 09: IRLS & RANSAC
• Lecture 10: L1 norm minimization
• Lecture 11: L1_norm_minimization
• Lecture 12: kalman filter, LS&KF, linear KF, nonlinear (extended) KF, (note revisions p. 7,8)
• Lecture 13: revise NLKF notation, dyn. models: lin & nonlin, cont. to discrete, euler, etc. target track
• Lecture 14: discuss HW5 #2 and #3, dynamic model and measurement model
• Lecture 15: unscented transf., unscented KF, trad. EP w/ NL funct., monte carlo vs. unsc. EP