Publisher DOI: 10.1007/978-3-319-99154-2_11
Title: Analytical solution for long battery lifetime prediction in nonadaptive systems
Language: English
Authors: Ivanov, Dmitry 
Larsen, Kim G 
Schupp, Sibylle 
Srba, Jiří 
Issue Date: 2018
Source: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (11024 LNCS): 173-189 (2018)
Journal or Series Name: Lecture notes in computer science 
Abstract (english): Uppaal SMC is a state-of-the-art tool for modelling and statistical analysis of hybrid systems, allowing the user to directly model the expected battery consumption in battery-operated devices. The tool employs a numerical approach for solving differential equations describing the continuous evolution of a hybrid system, however, the addition of a battery model significantly slows down the simulation and decreases the precision of the analysis. Moreover, Uppaal SMC is not optimized for obtaining simulations with durations of realistic battery lifetimes. We propose an analytical approach to address the performance and precision issues of battery modelling, and a trace extrapolation technique for extending the prediction horizon of Uppaal SMC. Our approach shows a performance gain of up to 80% on two industrial wireless sensor protocol models, while improving the precision with up to 55%. As a proof of concept, we develop a tool prototype where we apply our extrapolation technique for predicting battery lifetimes and show that the expected battery lifetime for several months of device operation can be computed within a reasonable computation time.
URI: http://hdl.handle.net/11420/2739
ISBN: 978-331999153-5
ISSN: 0302-9743
Institute: Softwaresysteme E-16 
Type: InProceedings (Aufsatz / Paper einer Konferenz etc.)
Appears in Collections:Publications without fulltext

Show full item record

Page view(s)

12
checked on Jun 16, 2019

Google ScholarTM

Check

Export

Items in TORE are protected by copyright, with all rights reserved, unless otherwise indicated.