hh.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Probabilistic Lane Change Prediction in Highway Scenarios: Lane change Maneuver Prediction in Highway Scenarios
Halmstad University, School of Information Science, Computer and Electrical Engineering (IDE). (Embedded and Intelligent Systems)
2011 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

The concern for more energy efficient vehicles has grown in the last years; the increase of oil prices and the need to curb greenhouse gas emissions has called for the search of solutions and alternatives to the fossil-fuel usage. 

On one hand there is the search for better fuel sources while on the other, the development for better driver assistance systems These systems should be capable of maximizing the driver's security and comfort while minimizing the fuel consumption. This can be achieved by the understanding of the environment surrounding the vehicle. Therefore the assistance should use this knowledge of its environment to alert the driver from any situation of risk so that they can be prevented. Moreover, reducing the number of accidents and unnecessary maneuvers will help reduce the fuel consumption by avoiding emergency breaking and other unnecesary situations. 

This thesis explains the development of a maneuver prediction system for highway scenarios. The system should identify lane changing vehicles with enough time to allow the driver to take the necessary precautions and maintain an efficient and more secure driving. To achieve this gial a set of features that describe the environment around the Ego vehicle is obtained. Afterwards, these Features are analyzed using state of the art data mining techniques and their performance is evaluated using a set of classification algorithms (Linear Discriminant Analysis, Neural networks and Random Forests). 

According to the present set of input data, it is possible to identify a left lane change from a right lane change with low misclassification error. But it is not possible to identify with the same efficiency a lane change from a no lane change. The best performance was obtained with a random forest where half of the lane changes are recognized while at the same time achieving a low number of false alarms; in another test, using a sub set of data by filtering noisy observations, it was possible to recognize left and right lane changes with more than 70\% efficiency. 

Place, publisher, year, edition, pages
2011. , 72 p.
Keyword [en]
Maneuver Prediction Data Analysis Artificial Intelligence Data Mining
National Category
Electrical Engineering, Electronic Engineering, Information Engineering Computer Systems
Identifiers
URN: urn:nbn:se:hh:diva-16486Local ID: IDE1150OAI: oai:DiVA.org:hh-16486DiVA: diva2:450863
Subject / course
Computer Systems Technology
Presentation
2011-09-27, B231, Högskolan Halmstad, Halmstad, 21:42 (English)
Uppsok
Technology
Supervisors
Examiners
Available from: 2011-11-18 Created: 2011-10-23 Last updated: 2011-11-18Bibliographically approved

Open Access in DiVA

Probabilistic Maneuver Prediction Montemayor11(1207 kB)407 downloads
File information
File name ATTACHMENT01.pdfFile size 1207 kBChecksum SHA-512
cbe47f29282aa932b8f76d7e1057f8ba473452a29b0c74ab147f3b76e9f82a9994638eae199c3813d4255cc3e4fc0a4ed0ca1a6e6fbcb7c2cedede0904686a20
Type attachmentMimetype application/pdf

Search in DiVA

By author/editor
Montemayor Ayala, Hector
By organisation
School of Information Science, Computer and Electrical Engineering (IDE)
Electrical Engineering, Electronic Engineering, Information EngineeringComputer Systems

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 83 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf