東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Improved Trajectory Planning for On-...

Gu, Tianyu.

Linked to FindBook

Google Book

Amazon

博客來

Improved Trajectory Planning for On-Road Self-Driving Vehicles Via Combined Graph Search, Optimization & Topology Analysis.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Improved Trajectory Planning for On-Road Self-Driving Vehicles Via Combined Graph Search, Optimization & Topology Analysis./
Author:	Gu, Tianyu.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2017,
Description:	148 p.
Notes:	Source: Dissertation Abstracts International, Volume: 78-08(E), Section: B.
Contained By:	Dissertation Abstracts International78-08B(E).
Subject:	Robotics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10258079
ISBN:	9781369618075

Improved Trajectory Planning for On-Road Self-Driving Vehicles Via Combined Graph Search, Optimization & Topology Analysis.
Gu, Tianyu.

Improved Trajectory Planning for On-Road Self-Driving Vehicles Via Combined Graph Search, Optimization & Topology Analysis. - Ann Arbor : ProQuest Dissertations & Theses, 2017 - 148 p.

Source: Dissertation Abstracts International, Volume: 78-08(E), Section: B.

Thesis (Ph.D.)--Carnegie Mellon University, 2017.

Trajectory planning is an important component of autonomous driving. It takes the result of route-level navigation plan and generates the motion-level commands that steer an autonomous passenger vehicle (APV). Prior work on solving this problem uses either a sampling-based or optimization-based trajectory planner, accompanied by some high-level rule generation components. However, these schemes are limited in the following respects: 1. Sampling-based planners yield global resolution-complete optimal trajectories, but suffer from the curse of dimensionality and sampling sub-optimality (i.e., almost never reach a local optimum). Optimization-based planners can find a local optimal trajectory, but suffer from the unawareness of global optimum. 2. Both types of trajectory planner lack explicit tactical or behavior-level reasoning capability. They rely on the high-level behavioral decision-making component to make motion-level decisions that do not necessarily comply with the maneuverability of the planner and the APV.

ISBN: 9781369618075Subjects--Topical Terms:

519753
Robotics.

Improved Trajectory Planning for On-Road Self-Driving Vehicles Via Combined Graph Search, Optimization & Topology Analysis.
LDR:03406nmm a2200313 4500 001 2122635
005 20170922124933.5
008 180830s2017 ||||||||||||||||| ||eng d
020 $a 9781369618075
035 $a (MiAaPQ)AAI10258079
035 $a AAI10258079
040 $a MiAaPQ $c MiAaPQ
100 1 $a Gu, Tianyu. $3 3284602
245 1 0 $a Improved Trajectory Planning for On-Road Self-Driving Vehicles Via Combined Graph Search, Optimization & Topology Analysis.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2017
300 $a 148 p.
500 $a Source: Dissertation Abstracts International, Volume: 78-08(E), Section: B.
500 $a Adviser: John M. Dolan.
502 $a Thesis (Ph.D.)--Carnegie Mellon University, 2017.
520 $a Trajectory planning is an important component of autonomous driving. It takes the result of route-level navigation plan and generates the motion-level commands that steer an autonomous passenger vehicle (APV). Prior work on solving this problem uses either a sampling-based or optimization-based trajectory planner, accompanied by some high-level rule generation components. However, these schemes are limited in the following respects: 1. Sampling-based planners yield global resolution-complete optimal trajectories, but suffer from the curse of dimensionality and sampling sub-optimality (i.e., almost never reach a local optimum). Optimization-based planners can find a local optimal trajectory, but suffer from the unawareness of global optimum. 2. Both types of trajectory planner lack explicit tactical or behavior-level reasoning capability. They rely on the high-level behavioral decision-making component to make motion-level decisions that do not necessarily comply with the maneuverability of the planner and the APV.
520 $a In this thesis, we adapt existing algorithms and propose new methods in the fields of optimal control, graph search and topological analysis to design an improved trajectory planning system. The core contributions of our work are summarized below: 1. A hybrid trajectory planner for on-road autonomous driving that maintains the key advantages of both the sampling-based and the optimization-based planners while reducing their limitations. 2. A novel type of edge-augmented graph that allows sampling-based planners to numerically approximate certain trajectory optimization methods. 3. A novel backward induction method based on topological analysis to perform configuration-space segmentation over a direct acyclic graph (DAG) as an efficient way to explore the topological structure of a global configuration space. 4. A novel maneuver pattern distinction method based on trajectory sampling and topological analysis to distinguish region-based, topology-based and overtaking sequence-based patterns for motion-level tactical reasoning. 5. Adaptation of the linear quadratic regulator (LQR) controller for model-feasible trajectory smoothing, and of the iterative-LQR (1st-order differential dynamic programming) for focused execution-feasible trajectory optimization. 6. Identification of useful principles and functions for constructing cost terms for trajectory evaluation.
590 $a School code: 0041.
650 4 $a Robotics. $3 519753
650 4 $a Automotive engineering. $3 2181195
650 4 $a Artificial intelligence. $3 516317
690 $a 0771
690 $a 0540
690 $a 0800
710 2 $a Carnegie Mellon University. $b Electrical and Computer Engineering. $3 2094139
773 0 $t Dissertation Abstracts International $g 78-08B(E).
790 $a 0041
791 $a Ph.D.
792 $a 2017
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10258079