Development of a Driving Schedule to Mimic Transit Bus Behavior in Mexico City 2006-01-3394

It is difficult to project the emissions performance of a vehicle on a route unless the test cycle used to gain the emissions data reasonably represents that route. A chassis dynamometer emissions measurement test schedule consisting of three modes (congested, non-congested and bus rapid transit (BRT) operation) was developed for use in a program to evaluate transit bus technologies in Mexico City. Existing buses were fitted with global positioning system (GPS) data loggers and, between September 2^nd and 8^th of 2004, 54 hours of speed-time data were collected while the buses were operated over several bus routes in Mexico City. The data set was then broken down into individual micro-trips, each consisting of an idle period followed by the bus traveling some distance, followed by a final deceleration to idle. An initial attempt to categorize each micro-trip according to time-of-day and thus determine whether the micro-trip occurred during defined rush-hour and non-rush-hour periods was unsuccessful due a lack of correlation between congested conditions in Mexico City and time-of-day (i.e. there is no clearly defined and spatially extensive rush-hour). Hence, the researchers defined two operational modes, namely “congested” and “non-congested”, with the distinction that the congested mode would contain micro-trips with average speeds falling in the lower 75^th percentile while the non-congested mode would contain micro-trips falling in the upper 75^th percentile. As a result, two modes would have different average speeds, and permit projection of emissions over links with a variety of average speeds. With these definitions, and a reasonable mode length of 1000 seconds, an iterative computer process randomly combined the individual micro-trips from each mode to form forty driving schedules. Each schedule was then compared to the characteristics of the entire respective (congested, non-congested) data set and the schedule which matched the characteristics of each mode most accurately was selected.