Development of Drive Cycle using Fleet Data for Two-Wheelers in Indian Market

Abstract

<div class="section abstract"><div class="htmlview paragraph">Generally, to produce reliable two-wheelers, manufacturers resort to intense engineering efforts to make sure the two-wheeler can withstand the most harsh testing conditions and requirements. This sometimes leads to a higher cost in realizing such outlier requirements. Thus, a drive cycle matching the actual riding characteristics will enable better understanding of the requirements and an optimized engineering effort. There have been several attempts by governmental and non-governmental organizations to realize a real drive cycle for various cities and countries, trying to capture the typical riding style in those regions. But the drive patterns observed in most representative cycles do not match with the scenario in India with frequently dense traffic, constrained roads and slow driving speeds.</div><div class="htmlview paragraph">To understand the driving pattern in India, a drive cycle generation algorithm is developed which uses real time on-road data captured from a fleet of vehicles in India and creating a database of micro-trips. These micro-trips are first categorized based on their average speeds. The algorithm concatenates these micro-trips to make a drive cycle, such that the average speed of the resulting drive cycle matches closely to the average speed of the captured on-road data. The algorithm then iterates different sequencing of these micro-trips in the drive cycle to minimize the error in various parameters like average acceleration, time percentage of acceleration, & deceleration, time percentage of idle, between the resulting drive cycle and the captured on-road data. Representative cycles of different cities and regions have been developed and described in this paper.</div><div class="htmlview paragraph">This paper aims in explaining the approach of extracting a drive cycle from the data collected from a fleet of two-wheelers on-road in the Indian market and comparing the different riding patterns found in different regions. The algorithm developed can be extended to any level of data, ranging from a particular city to even combining different countries together.</div></div>