[ Home ] [ Products - Index ] [ Services ] [ Next ] [ Previous ] [ Contact ]

7.0 Error Correction in Passenger Counting 

Iris-GmbH has introduced four error definitions for the objective evaluation of passenger counting systems:

• the passenger error,
• the balanced boarding error and the balanced alighting error and
• the unbalanced error.

The IRMA passenger counting system counts passengers with an error of <= 5% or an accuracy of >= 95%, respectively.  So what are the practical implications of this accuracy?

Let us turn to an example:  A vehicle transports 1,000 passengers a day. As IRMA counts the passengers with an error of <= 5%, the value measured for the passengers will be between 950 (=1,000 passengers - 5%) and 1,050 (=1,000 passengers + 5%).The number of passengers counted by IRMA (1 passenger = 1 boarding person + 1 alighting person) is calculated using the means of persons boarding and persons alighting.  IRMA counts boarding passengers and alighting passengers with an error of <= 10% or an accuracy of >= 90%, respectively.

If a vehicle transports 1,000 passengers a day, the value measured by IRMA for boarding passenger will be between 900 (=1,000 boarding passengers -10%) and 1,100 (=1,000 boarding passengers +10%) and the value measured for alighting passengers between 900 (=1,000 alighting passengers -10%) and 1,100 (=1,000 alighting passengers +10%).

Iris-GmbH guarantees the following values for the IRMA passenger counting system (for the raw data, i.e., the data before any statistical processing):

The prerequisites for reaching this level of accuracy are:

• the sensors have been installed immediately next to the doors and been adjusted accurately;
• stops where a passenger change with a buggy takes place will not be considered in the calculation, and
• a minimum of 700 passengers will be counted.

In daily IRMA operations, it is observed:

Explanations and formulas for the calculation of errors, meaning of the errors

s = number of stops considered

d = number of vehicle doors

Meaning of the passenger error, the balanced boarding passenger error and the balanced alighting passenger error

The passenger error, the balanced boarding passenger error and the balanced alighting passenger error are of some significance for the statistical evaluation of the counts, of the degree of load and of passenger miles.  As no automatic passenger counting system works error-free, in most cases the number of boarding passengers counted does not match. the number of alighting passengers counted.  Therefore, the raw data (the non-processed figures for boarding passengers and alighting passengers) produced by the passenger counting system undergo subsequent statistical processing, which serves to determine the number of passengers out of the boarding passengers and alighting passengers counted, to check the data for plausibility and makes slight corrections to them.

In order to attain as good results as possible in the statistical processing of the raw data, IRMA usually counts

• a slightly higher number of boarding passengers and a slightly lower number of alighting passengers
• (or a slightly lower number of boarding passengers and a slightly higher number of alighting passengers).

So, the number of passengers (the means of boarding passengers and alighting passengers) calculated in the course of statistical processing) will be of a higher accuracy than the actual raw data for boarding passengers and alighting passengers.

The significance of the unbalanced error

The unbalanced error describes the number of absolute errors and is therefore of importance for the comparison of different installations and for the comparison of different passenger counting systems.