Automated Passenger Counting (APC) Systems: How Do They Work?
What Are APCs?
APC systems are electronic machines that count the number of passengers that board and disembark at every bus stop. They, together with AVL systems , form the two most important technologies that every transit system should have. In systems that have them, they replace the schedule checkers that previously collected ridership information manually. When the Federal Transit Administration is satisfied they are calibrated correctly, ridership information they collect can be used to fulfill National Transit Database reporting requirements.
Why Should I Get APCs?
The main advantage of APCs are that, unlike schedule checkers, they collect ridership for as many as every single trip operated, if APC units are installed on 100% of the bus fleet. They also reduce cost, because even if initial startup costs are high in the long term it costs much less to collect ridership information via APC units than it does to hire employees to manually collect it. The main disadvantage is that APC units, while accurate, are sometimes not as accurate as manual collection - APC units collect accurate information from 80 - 95% of the time while manual collection is generally accurate between 90 and 95% of the time. One example of APC accuracy problems occurs when for some reason on a particular trip the number of boardings does not equal the number of alightings. While a manual checker would be able to start the next trip with a load of zero, APC systems may carry over a non-zero end of trip load if not reset by software, thereby passing collection errors on one trip to the following trip.
How Do APCs Work?
Two sets of two sensors each at the same height level are installed at both the front and rear door. When passengers enter or exit the door they break an infrared beam, which causes a computer to record a boarding or alighting depending on the order in which the two beams are broken. The sensors are enough to provide a gross level of ridership; if stop-level ridership is required geographic information must be provided by a GPS system such as an Automated Vehicle Locator (AVL) program. Data is then downloaded to a computer for analysis.
How Much Do APCs Cost?
The actual passenger counting system units cost about $10,000 per bus; if additional AVL equipment is needed to allow for collection of stop level data than the cost will increase. Of course, this cost does not include the development and installation of any software needed to analyze the APC data - figure at least another $250,000 for these costs.
How Many APCs Does My Transit System Need?
To provide enough APC-equipped buses so that every trip is sampled a reasonable amount of times in a given period, 10% of the fleet should have the units. To meet Title VI requirements , the units should be distributed throughout the fleet rather than concentrated in one model year or one geographical area. However, this number assumes that the transit agency has the ability to distribute these vehicles amongst all the blocks so that all trips will eventually get sampled. Assigning buses in this matter may cause additional work for transit supervisors; installing APC units on all vehicles in the fleet - which seems to be the goal of systems with APC devices - avoids this problem.
How Are APCs Used?
APC systems are used to generate robust ridership information on a stop by stop basis. They are the best method of collecting ridership; as discussed earlier manual ride checks, while accurate, are very limited in scope, and ridership from farebox reports, even if accurate, cannot provide information on where passengers left the bus, making it impossible to know the load of the bus and segments of a route that especially high or low ridership. Another way that APC systems are used is APC reports can be used to determine schedule adherence and whether bus routes need more or less running time to get between timepoints. Truly, APC units are an essential ingredient in effective transit planning .
APCs Versus Manual Counting and the Effect on National Transit Ridership
APCs, by allowing for a 100% counting of passenger trips, provide much more accurate information than the old manual counting method does, but the differences go beyond that. In fact, it can be misleading to compare ridership generated by manual counts to ridership generated by APCs. Here's why: NTD ridership generated by manual counts is calculated by multiplying the average number of passengers on a small number of randomly selected trips (as few as 48 per month) by the total number of trips in the calendar month. Of course, if the randomly selected trips include several that have very low or very high ridership the monthly ridership total will be distorted. More importantly, if a transit agency adds trips in a month, its' NTD ridership will almost always increase; and if a transit agency subtracts trips in a month, its' NTD ridership will almost always decrease, because of the NTD formula. Federal formulas cannot take into account the possibility that a transit agency may cut trips that have no passengers at all; in such a case, the NTD ridership would decline (because there would be fewer trips to multiply the average ridership per trip by) while the actual ridership may have no change. In Service Cut Strategies , I noted that CTA and Metro ridership totals were not affected by their service reductions, while Community Transit had a significant decline. Does the fact that both CTA and Metro collect ridership information from APCs while Community Transit uses manual data collection help to explain the change in ridership? At this point, nobody knows.
The installation of Automated Passenger Counting equipment should be one of the top priorities for all agencies that still collect ridership using traditional manual methods. While there may be a significant upfront cost for the installation, this cost is more than offset by continuing operating savings in the future and the wealth of usable data on ridership and on-time performance that APCs provide. Transit agencies should just be aware that there may be a significant set-up period before APCs will fully function; I recommend that hiring of consultants to assist in the set-up.