Observing a Moving Agent

We addrcss tllc problcin of observing a moving agent.. 11i particlrlar, wc ploposc a system for ol~scrving a. manil)l~labioti plorrss , wlicrc a r o l ~ o t I ~ a n d maiii])ulal,cs an object. A tliscrctc cvcnl dynamic sysI.cm (DEIIS) frainc w o ~ k is devclo~tctl for the hand/objcct inlcracbion ovcr t i~r lc ant1 a sta1)iIizing obscrvcr is const,rucbcd. Low-lcvcl m o d ~ ~ l c s arc dcvclopctl for rrcognizlng kllc "cvcnts" 1ha.t caliscs s ta te transibioiis wibhitl thc dynamic manipulal ion sysl ,c~n. 'I'llc work rsamil i rs closcly tllc possibilibics lor errors, mistakes a nrl i~nctrl ,ai~lt , ics in blic manilnrIa.l,ion systcm, obsctvcr constrtictio~i j)ioccss atitl cvcrib itlr~~l,ification n~cc l lan is~ns . T h e s y s t c ~ n utilizcs diffcrcnt, trackitig l,cclir~iclr~cs i n ortlcr to obscrvc I , l~r task in a n oc/7vc, ndaptivr: ant1 gorrl-(lit-cctcrl Inalincr. IZcywords : Cornpl~bcr Vision, Control 'l'l~coiy, IIccisio~i hllaliing, Discrctc Event 1)yna.mic Syslcms, Firiile S t a t c Al~l,oiriata, Irnagc I'roccssil~g, Manrr iacburit~g Sys t c~ns , hlIotion Esl,itnntion, 11.o1~olics. A c k ~ i o w l e d g c i ~ ~ c i l t s : This rescarcl~ \vas s~~ppor(.ctl in 11art 11y Air Forcc AFOSR. Gran1.s 88-0211, 880296; Army/DAAT, Grant. 03-89-C-003JPRI; NSF Cra~ils CISE/(';DA 88-22719, 1111 89-06770; 11A111'A Craiil. N0011-88-0630 and D~iPonl. Corl~or;ll.ion.

[1]  Ruzena Bajcsy,et al.  Models for motion perception , 1987 .

[2]  Narendra Ahuja,et al.  3-D Motion Estimation, Understanding, and Prediction from Noisy Image Sequences , 1987, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Edward S. Deutsch,et al.  On the Quantitative Evaluation of Edge Detection Schemes and their Comparison with Human Performance , 1975, IEEE Transactions on Computers.

[4]  David Malah,et al.  A study of edge detection algorithms , 1982, Comput. Graph. Image Process..

[5]  Seungryoul Maeng,et al.  Real-Time Estimation of 2-D Motion for Object Tracking , 1990, Other Conferences.