Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Eco-driving assistant application

Eliasson, Amanda LU and Timan, Robin (2018) MAMM01 20181
Ergonomics and Aerosol Technology
Abstract
The global environmental impact made by cars is substantial and will always be in focus when discussing climate change and our carbon footprint. Eco-driving has long been a given set of general rules for drivers to follow such as smooth acceleration.

We wanted to explore the use of machine learning to identify and learn unknown driving patterns that may affect the fuel consumption. We also wanted to explore how communication between the driver and the application should be designed so that it doesn't disturb the driver while driving and how guidelines should be illustrated to motivate the driver in driving eco-friendlier. We did it by exploring the use of gamification as a motivational tool.

We investigated different machine... (More)
The global environmental impact made by cars is substantial and will always be in focus when discussing climate change and our carbon footprint. Eco-driving has long been a given set of general rules for drivers to follow such as smooth acceleration.

We wanted to explore the use of machine learning to identify and learn unknown driving patterns that may affect the fuel consumption. We also wanted to explore how communication between the driver and the application should be designed so that it doesn't disturb the driver while driving and how guidelines should be illustrated to motivate the driver in driving eco-friendlier. We did it by exploring the use of gamification as a motivational tool.

We investigated different machine learning techniques and explored each method's limitations and possibilities. We also tested different design alternatives for the application to see which design is both suitable for driving and motivates the driver to reach their eco-goal.

We found that there are many ways to apply machine learning for eco-driving purposes and each method has its own set of pros and cons. In this report we provide no single right answer to how to apply gamification and machine learning in a driving environment but rather a proof of concept to follow for further development.

Data-driven development has many applications in real-world problems and eco-driving is one of them. We learned that personalizing feedback and
displaying it with a gamified design encouraged drivers to be motivated into reaching an eco-friendlier driving style. (Less)
Abstract (Swedish)
Bilars globala miljöpåverkan är substantiell och kommer alltid vara i fokus när man diskuterar klimatförändringar och vårt koldioxidavtryck. Eko-körning har länge bara varit ett givet antal av riktlinjer åt förare, om till exempel mjuk acceleration.

Vi ville utforska användningen av maskininlärning för att identifiera och lära okända körmönster som kan påverka bränsleåtgången. Vi ville också utforska hur kommunikationen mellan föraren och applikationen bäst designas så att den inte stör föraren under körning och hur riktlinjer ska illustreras för att motivera föraren att köra mer eko-vänligt. Vi gjorde detta genom att utforska spelifiering som motivationsredskap.

Vi undersökte olika maskininlärningstekniker och utforskade varje... (More)
Bilars globala miljöpåverkan är substantiell och kommer alltid vara i fokus när man diskuterar klimatförändringar och vårt koldioxidavtryck. Eko-körning har länge bara varit ett givet antal av riktlinjer åt förare, om till exempel mjuk acceleration.

Vi ville utforska användningen av maskininlärning för att identifiera och lära okända körmönster som kan påverka bränsleåtgången. Vi ville också utforska hur kommunikationen mellan föraren och applikationen bäst designas så att den inte stör föraren under körning och hur riktlinjer ska illustreras för att motivera föraren att köra mer eko-vänligt. Vi gjorde detta genom att utforska spelifiering som motivationsredskap.

Vi undersökte olika maskininlärningstekniker och utforskade varje metods begränsningar och möjligheter. Vi testade även olika designalternativ för applikationen för att se vilken design som är både lämplig under körning och motiverar föraren att nå sitt eko-mål.

Vi kom fram till att det finns många sätt att applicera maskininlärning för eko-körningssyften och att varje metod har sina för- och nackdelar. I den här rapporten kommer vi inte bistå med ett enda rätt svar på hur man applicerar spelifiering och maskininlärning i en bilmiljö utan snarare ett bevis på ett koncept att följa för vidareutveckling.

Data-driven utveckling har många appliceringsområden i verkliga problem och eko-körning är en av dem. vi lärde oss att genom att personifiera återkoppling och visa upp det med en spelifierad design kan förare motiveras till att uppnå en mer eko-vänlig körstil. (Less)
Please use this url to cite or link to this publication:
author
Eliasson, Amanda LU and Timan, Robin
supervisor
organization
course
MAMM01 20181
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Machine learning, Gamification, Android, eco-driving
language
English
id
8946853
date added to LUP
2018-06-08 10:39:56
date last changed
2018-06-08 11:09:38
@misc{8946853,
  abstract     = {{The global environmental impact made by cars is substantial and will always be in focus when discussing climate change and our carbon footprint. Eco-driving has long been a given set of general rules for drivers to follow such as smooth acceleration.

We wanted to explore the use of machine learning to identify and learn unknown driving patterns that may affect the fuel consumption. We also wanted to explore how communication between the driver and the application should be designed so that it doesn't disturb the driver while driving and how guidelines should be illustrated to motivate the driver in driving eco-friendlier. We did it by exploring the use of gamification as a motivational tool.
 
We investigated different machine learning techniques and explored each method's limitations and possibilities. We also tested different design alternatives for the application to see which design is both suitable for driving and motivates the driver to reach their eco-goal.

We found that there are many ways to apply machine learning for eco-driving purposes and each method has its own set of pros and cons. In this report we provide no single right answer to how to apply gamification and machine learning in a driving environment but rather a proof of concept to follow for further development. 

Data-driven development has many applications in real-world problems and eco-driving is one of them. We learned that personalizing feedback and
displaying it with a gamified design encouraged drivers to be motivated into reaching an eco-friendlier driving style.}},
  author       = {{Eliasson, Amanda and Timan, Robin}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Eco-driving assistant application}},
  year         = {{2018}},
}