Speeding up single-query sampling-based algorithms using case-based reasoning

•Motion planning is an essential task for an autonomous vehicle.•Traditional motion planning algorithms ignore experience when serving new queries.•Experience-based algorithms use multiple thread execution for serving queries.•Motion planning belongs to P-SPACE hard problem class.•Case-based reasoni...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Expert systems with applications Ročník 114; s. 524 - 531
Hlavní autoři: Abdelwahed, Mustafa F., Saleh, Mohamed, Mohamed, Amr E.
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York Elsevier Ltd 30.12.2018
Elsevier BV
Témata:
Algorithms
Artificial intelligence
Case-Based reasoning
Complexity theory
Experience-based algorithms
Exploitation
Reasoning
Response time (computers)
Sampling
Sampling-based algorithms
Case-Based reasoning
Experience-based algorithms
Sampling-based algorithms
Artificial intelligence
ISSN:0957-4174, 1873-6793
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:•Motion planning is an essential task for an autonomous vehicle.•Traditional motion planning algorithms ignore experience when serving new queries.•Experience-based algorithms use multiple thread execution for serving queries.•Motion planning belongs to P-SPACE hard problem class.•Case-based reasoning is an AI approach for problem-solving using stored experiences. We present an extension to the single-query sampling-based algorithm for improving its response time using Case-Based Reasoning (CBR) technique. Unlike traditional experience-based planners, CBR depends on a single thread execution which reduces the required computation power. Additionally, it is always biased towards exploration rather than exploitation to overcome experience-based algorithms drawbacks. Results indicate that CBR extension has significantly improved sampling-based response time for similar served queries.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0957-4174
1873-6793
DOI:10.1016/j.eswa.2018.08.035