A hybrid dynamical systems approach to intelligent low-level navigation

Abstract

Animated characters may exhibit several kinds of dynamic intelligence when performing low-level navigation (i.e., navigation on a local perceptual scale): They decide among different modes of behavior, selectively discriminate entities in the world around them, perform obstacle avoidance, etc. In this paper, we present a hybrid dynamical system model of low-level navigation that accounts for the above-mentioned kinds of intelligence. In so doing, the model illustrates general ideas about how a hybrid systems perspective can influence and simplify such reactive/behavioral modeling for multi-agent systems. In addition, we directly employed our formal hybrid system model to generate animations that illustrate our navigation strategies. Overall, our results suggest that hierarchical hybrid systems may provide a natural framework for modeling elements of intelligent animated actors. This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of the University of Pennsylvania's products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to pubs-permissions@ieee.org. By choosing to view this document, you agree to all provisions of the copyright laws protecting it.This conference paper is available at ScholarlyCommons: http://repository.upenn.edu/hms/32 A Hybrid Dynamical Systems Approach to Intelligent Low-Level Navigation

An algebra of hybrid systems

“…y y nd pply meet opertion with the set of ll innite trjetories t the outermost level s in ixmple QFWF yftenD it is not neessry to store the ontrol mode in the vlue setD iFeFD V n e hosen s s n insted of M ¢s n F ixmples of this re given in ixmple QFR nd etion TF 4.2 Composition of Hybrid Automata wore omplited hyrid systems rise often y omposing smller systemsF he produt of two nite utomt s well s the prllel omposition re well knownF imilr to these onstrutions renzinger denes produt nd prllel omposition for hyrid utomt PQF sn this setion we disuss their lgeri ounterprtsF Product. pollowing etion RFI nd the denition of produt of hyrid uE tomt PQ we dene for two hyrid utomt A I ; A P @with disjoint sets of modes M I nd M P A the following edge lelling funtions @for jumps nd eventsA @v I ; v P A a 3 @w I ; w P A D df W a I P A I ; a P P A P X v I a 1 3 w I ; v P a 2 3 w P :…”

“…(3) a ¡ b u £p ¡ £q a @a u £pA ¡ @b u £qA C @a u £pA ¡ @b u £p ¡ £qA C @a u £p ¡ £qA ¡ @b u £qA : (4) If additionally p £p holds, the summand @a u £ F pA ¡ @b u £qA can be omitted from the right hand sides of Parts (1) and (3).he lengthy proof n e found in the eppendix of PWF por singleD nite trjetories rt @IA is illustrted in pigure TF rereD the hnge etween propE erties p nd q n our either extly t the omposition point of a nd bD inside a or inside bF ht is why the formul on the right hnd side of rt @IA onsists of three summndsF Figure 6. Composed trajectories satisfying £ F p ¡ £q en pplition of vemm SFIR@IA is to omine sfety requirements of the shpe R l;u F ine £ F p ¡ £q a £p _ £qD sfety requirement of this form gurntees tht the proess £q is tully enteredF e onlude y yet nother equivlent hrteristion of time progressF sn the dethment formuls of this lemmD forming the meet with I performs the projetion into the test lgerD nd we otin our revised opertors y omitting this meetF here is hoie in whih of these two formuls to useF e tke the rst oneD sine it results in more diret trnsltion of the universl quntier A H F essume foolen quntle S nd a; b P SF hen Eb : w a df wb ; Ab : w a df Eb : w a w=b ;…”

Action selection and task sequence learning for hybrid dynamical cognitive agents

“…Transitions between modes are instantaneous, based on guard conditions, and may have discontinuous side effects, encoding discrete dynamics. Hybrid dynamical systems can be apt models for navigating robots or animated agents (e.g., [9,10]), and HDCAs' reactive and deliberative structures naturally correspond to HDS elements. Each task of an HDCA is a reactive behavior, implemented as an HDS mode; deliberation in HDCAs occurs during mode transitions.…”

Preference-Based Trajectory Generation

“…While most behavior-based protocols can be represented with a Markov decision process (MDP), another branch of research looks at hybrid dynamic systems [15,16]. In such systems, discrete events trigger shifts between different continuous dynamics.…”