heather2

Information about heather2

Published on January 16, 2008

Author: Marietta1

Source: authorstream.com

Content

Spacecraft Rendezvous and Docking with Real-Time, Randomized Optimization:  Spacecraft Rendezvous and Docking with Real-Time, Randomized Optimization Phillips, Kavraki, and Bedrossian, 2003 Presented by Heather Arneson September 19, 20006 Overview:  Overview Base Problem: 6 degree-of-freedom spacecraft rendezvous and docking problem Impulsive control Avoidance of known obstacles Avoidance of plume impingement Minimize fuel use Extension of Base Problem: Real-time obstacle avoidance Plume Impingement:  Plume Impingement Method:  Method Hybrid offline-online technique Offline Probabilistic search is used to generate a path that avoids collision and plume impingement Online Gradient descent method used for real-time obstacle avoidance Feedback control loop to handle error in the initial offline solution Inputs to Algorithm:  Inputs to Algorithm Start and goal configurations of chase vehicle relative to target vehicle Method to check for collisions and violation of other constraints Method to integrate impulsive control Metric to determine cost for a given control Estimate of cost from a particular configuration to the goal configuration Algorithm 1: Randomized Tree Expansion:  Algorithm 1: Randomized Tree Expansion for i = 0 to N do p = CHOOSE_WAYPOINT() n = EXPAND_WAYPOINT(p) ADD_TO_TREE(p,n) end for CHOOSE_WAYPOINT (which waypoint should be expanded?):  CHOOSE_WAYPOINT (which waypoint should be expanded?) Compute weight for each waypoint weight(p) = order(p)/(A*_cost(p) out_degree(p)) out_degree(p) = # times p has been expanded order(p) => more recent nodes chosen more often Randomly select waypoint with probability proportional to it’s weight => Not just following the locally optimal direction => won’t get stuck in local minima EXPAND_WAYPOINT:  EXPAND_WAYPOINT Creates a new waypoint, n, p -> n is likely feasible Apply a control to p and integrate to get n Control direction, magnitude, time to integrate chosen at random Cost computed easily Only feasible movements from a waypoint are considered and explored (no collision check) ADD_TO_TREE:  ADD_TO_TREE p -> n feasible? (check for collision) cost(n) = cost(p) + cost(p -> n) n is added as an outbranch from p estimate cost(n -> goal) n -> goal feasible? (check for collision) (how do they know how to get from n to goal?) cost to goal through n is calculated store lowest cost path Range Searching:  Range Searching Include proximity of waypoint to other waypoints in the weight to expand search at the perimeter of explored space What is close when using kinodynamic constraints? Path Gradient Descent best path refinement (offline):  Path Gradient Descent best path refinement (offline) Find the cost gradient of the path and incrementally perturb the path in that direction Perturb each waypoint in the path in the direction of the gradient of that waypoint Deform waypoints to decrease cost while not violating constraints Algorithm 2: Path Gradient Descent:  Algorithm 2: Path Gradient Descent for i = 0 to N do randomly permute order of waypoint considered for j = 0 to # waypoints do calculate gradient follow gradient e end for end for Dynamic Obstacle Avoidance (online):  Dynamic Obstacle Avoidance (online) Path Gradient Descent technique can be extended to work in a changing environment Can be used as feedback control Account for integration error Dynamic obstacle avoidance Guided Randomized Expansion Tree:  Guided Randomized Expansion Tree A Low Cost Path:  A Low Cost Path Path Gradient Descent:  Path Gradient Descent Initial Path Refined Path Dynamic Obstacle Avoidance:  Dynamic Obstacle Avoidance Avoidance Paths Initial Path Final Path

Related presentations


Other presentations created by Marietta1

Solar Energy Slide Show 1 3
26. 01. 2008
0 views

Solar Energy Slide Show 1 3

SLIDESHOW
08. 05. 2008
0 views

SLIDESHOW

EU
07. 05. 2008
0 views

EU

Cummins AWI update
02. 05. 2008
0 views

Cummins AWI update

00000032 src
02. 05. 2008
0 views

00000032 src

1101e4short
30. 04. 2008
0 views

1101e4short

1184287012505 0
24. 04. 2008
0 views

1184287012505 0

Skretting
22. 04. 2008
0 views

Skretting

qiweichen
21. 04. 2008
0 views

qiweichen

Lecture 5 Ideology and Practice
17. 04. 2008
0 views

Lecture 5 Ideology and Practice

16 Consumer Oriented E Commerce
04. 02. 2008
0 views

16 Consumer Oriented E Commerce

christmas vocabulary
30. 01. 2008
0 views

christmas vocabulary

Lecture4 5
12. 02. 2008
0 views

Lecture4 5

comets
10. 01. 2008
0 views

comets

angelfish
10. 01. 2008
0 views

angelfish

stakeholder mapping spring2007
12. 01. 2008
0 views

stakeholder mapping spring2007

Lecture 8
12. 01. 2008
0 views

Lecture 8

NOO
13. 01. 2008
0 views

NOO

Disparities Final 131125 7
14. 01. 2008
0 views

Disparities Final 131125 7

Outer Solar System
16. 01. 2008
0 views

Outer Solar System

MFC 2005Construction
18. 01. 2008
0 views

MFC 2005Construction

Tech4
18. 01. 2008
0 views

Tech4

afsm100
23. 01. 2008
0 views

afsm100

Fin525Fall2006Week1
24. 01. 2008
0 views

Fin525Fall2006Week1

APCh09
28. 01. 2008
0 views

APCh09

Liver and the Heart
15. 01. 2008
0 views

Liver and the Heart

4101
07. 02. 2008
0 views

4101

100707 Samson
28. 01. 2008
0 views

100707 Samson

Grooming
29. 01. 2008
0 views

Grooming

Session 1b
07. 02. 2008
0 views

Session 1b

Christine Gurnett HSE
07. 02. 2008
0 views

Christine Gurnett HSE

Japanese Theatre Dance
13. 02. 2008
0 views

Japanese Theatre Dance

ERuel Critical Periods
13. 02. 2008
0 views

ERuel Critical Periods

Lecture 1 2005
14. 02. 2008
0 views

Lecture 1 2005

RMSI DisasterManagement
11. 02. 2008
0 views

RMSI DisasterManagement

infrastruct toby brownecooper
08. 03. 2008
0 views

infrastruct toby brownecooper

07 FosteringCreativity
10. 03. 2008
0 views

07 FosteringCreativity

SOME HINTS ABOUT ISTANBUL LIFE
04. 02. 2008
0 views

SOME HINTS ABOUT ISTANBUL LIFE

Prevention
15. 03. 2008
0 views

Prevention

2007 HRPAO Presentation
07. 04. 2008
0 views

2007 HRPAO Presentation

TSG VIVA
27. 03. 2008
0 views

TSG VIVA

Thalia EQs Bldgs
30. 01. 2008
0 views

Thalia EQs Bldgs

electronicfreire
20. 02. 2008
0 views

electronicfreire

Pecan2002
08. 01. 2008
0 views

Pecan2002

outbriefmar07
07. 02. 2008
0 views

outbriefmar07

oilsandscalgary2
23. 01. 2008
0 views

oilsandscalgary2

SO2DB
06. 02. 2008
0 views

SO2DB

S Esseh
11. 01. 2008
0 views

S Esseh

sivan progsem 2003
15. 01. 2008
0 views

sivan progsem 2003

smallNEO
17. 01. 2008
0 views

smallNEO

Slovenia
03. 03. 2008
0 views

Slovenia

200710892712598
29. 01. 2008
0 views

200710892712598

BR New ParentsV2 C 4 0
15. 01. 2008
0 views

BR New ParentsV2 C 4 0

pps 319
14. 02. 2008
0 views

pps 319

Chapter01and02street 08Spring
19. 03. 2008
0 views

Chapter01and02street 08Spring

PRES INGLESE overview
28. 01. 2008
0 views

PRES INGLESE overview

CASIX poster pages
05. 02. 2008
0 views

CASIX poster pages

ATC 2006 Corporate teams
21. 01. 2008
0 views

ATC 2006 Corporate teams