I attended TransportationCamp East this past weekend, and one of the technology themes which kept coming up was the use of smartphones to provide an inexpensive AVL solution for smaller transit agencies which do not presently have real-time vehicle location information. The concept of operations is simple: the smartphone is loaded with an application that fetches the phone's location at a set interval, and reports it over the phone's cellular link to a central server. Theoretically, it's a good, clean solution. Unfortunately, the reality is not quite as attractive. Conventional AVL systems are expensive, and while I agree that to a certain extent they're more expensive than they need to be, a smartphone isn't an appropriate replacement. Here's why:
- Reliability of positioning: Even though most smartphones have a GPS chip, that doesn't necessarily mean they're using the GPS chip for positioning all the time. Instead, they will almost certainly use some combination of multilateration, Wi-Fi positioning and GPS. Of the three technologies, only GPS provides superior accuracy and autonomous operation. Wi-Fi positioning and multilateration are both dependent on service providers, while GPS requires only the GPS receiver itself. In addition, the GPS receivers in cell phones are not as robust as those found in GPS receivers for AVL applications. A commercial-grade GPS receiver for AVL applications will provide a superior time to first fix and better reliability in urban areas, where GPS has historically been unreliable (the 'urban canyon' effect). GPS receivers for vehicle-borne applications (like the Trimble Placer, used by MTA New York City Transit in their OneBusAway pilot on the B63 line) are designed for improved dead-reckoning performance by interfacing with the vehicle.
- Communications: Many existing AVL systems are integrated into a transit agency's trunking radio system. Not only does this provide a channel for data communication, it can also be integrated into the agency's dispatching system. When a bus operator calls the dispatcher, the location of the bus can be displayed automatically. Even better, the AVL system's user interface can provide a bus operator with options for sending pre-defined messages to the dispatcher: mechanical trouble, emergency requiring police, etc. This provides two benefits: all the bus operator has to do to signal a problem is push a button, and, because the condition is signalled over the data link, they don't have to take up airtime to communicate the problem.
All of this can be done with a home-grown AVL system, but it's harder. You won't get integration with the dispatch console, and you've got to provide a separate data connection for the AVL system. The data rates needed don't require anything faster than a CDPD or Mobitex link, but instead a smartphone will use a 3G link—fast, but not necessarily as reliable. On top of that, you won't be able to take advantage of an agency's existing investment in radio infrastructure, instead having to pay for a commercial 3G link.
- Power, mounting, and durability of equipment: A real AVL system is designed and built to be installed in a bus. It's designed to be wired into the bus, and it's designed to do so properly—so it doesn't draw power when the bus isn't actually running and thus drain the battery on the bus (nobody likes a bus which won't start). Some AVL systems are also designed to integrate with other onboard electronics using the SAE J1708 standard. There are also important mechanical considerations—to start, it's hard to mount a smartphone in a bus. Beyond that, vehicle-borne equipment must contend with extremes of temperature, vibration, and power (don't think that 12-volt devices in a vehicle actually see a regulated 12 volts all the time). If you want to provide a user interface (as opposed to just mounting a box somewhere in the bus), then you've got to find a a way to mount your equipment so that it is accessible to the bus operator, but will not present a distraction or become hazardous in a collision.
- User interface: Unlike a purpose-built device, a smartphone presents unnecessary complexity—you've got to keep the driver from being able to access other applications (like a Web browser), and also ensure that the tracking app is always running. Of course, this assumes that the device is readily accessible. Given that smartphones don't have external GPS antennae, the best option is to somehow mount the phone as close to the exterior of the bus as you can. However, if that means locating the phone somewhere that isn't readily accessible, then you complicate troubleshooting. A commercial AVL system will mount the GPS receiver or its antenna on the roof of the bus, while placing the user interface in a location which is convenient and comfortable for the bus operator. When the smartphone crashes (and it will), you don't want to have to crawl all over the bus to reset it.
Now, it is possible to use COTS hardware to put together an AVL system—the basic building blocks are a GPS receiver, a small embedded PC, and a 3G modem. However, it's still going to cost more than a smartphone. While OneBusAway will provide a good backend and user interface for riders, you'll still have to write the vehicle-borne software and user interface. If you want to add dispatching (using VoIP, for example), that'll take more work (although it's not impossible).
I'm not trying to say that you should just write Orbital a huge check and leave it at that—but if you do want to build your own AVL system, you should understand what goes into it. An industrial PC which can tolerate the environmental and electromagnetic conditions found on a bus may be expensive, but it's worth it. The same goes for GPS receivers and other specialized hardware. Some agencies keep their buses on the road 20 hours a day, operating in challenging conditions, and may have a fleet of hundreds or even thousands of buses. A smartphone-based tracking program may work well as a pilot project, but when it comes to full-scale deployment, the value of equipment which is designed for the environment becomes apparent.