The downside of having an abundance of vendors in the patrol car electronics market is that there is no clear standard for all those systems. When one looks at the interiors of patrol cars of different jurisdictions, it’s clear that none of them look alike. An officer from Agency A would need an orientation session in order to operate all of the gadgets in Agency B’s cars. The inner workings of these devices reveal even more dissimilarities, and it’s no small wonder that the technicians in the radio shops and city yards of America suffer continuing frustration at their mission to get all these displays, control heads, black boxes, connectors and cable harnesses to play nice together.
Project54, an initiative under way at the University of New Hampshire (UNH), aims to resolve some of these problems by developing a standard for integration of patrol car electronics. This could have significant benefits for the end users of these products, the cops on the street. First, devices brought under the Project54 umbrella could share wiring, power supplies, and user interfaces, so that they could take up considerably less room and be more ergonomically sound for officers to use. Second, the technologies developed by Project54 have already produced a patrol car with electronics that operate reliably and consistently by voice commands, making it possible for officers to run pursuits, make traffic stops, and conduct license plate inquiries without taking their eyes off of the road or their hands from the wheel.
The voice-commanded patrol car environment has been proposed and even marketed a few times, but with limited success. There were two principal barriers to making the speech recognition systems work reliably. One is that both the software and the hardware had not matured sufficiently. Speech recognition requires some significant computing horsepower which wasn’t available in portable computers until recently. The other problem is that the patrol car is a fairly noisy environment, and in order for speech recognition to work at all, officers had to wear a headset or use some other fixed microphone arrangement that kept the input to the system consistent in volume and clarity. The headset was impractical for officers who had to enter and exit the car often, as is the case with patrol operations.
The Project54 system uses an array microphone mounted on the dashboard of the patrol car, behind the steering wheel. An array microphone is actually a series of microphones mounted horizontally that accept sounds from a relatively narrow “window,” and ignore sounds coming from outside of that zone. Project54 is also unique among most of the other speech recognition systems in that it requires little or no prerecognition training of the software. With other speech recognition applications, users must typically “train” the software by reading extensive passages of text into the system so that the software will associate the words with the sounds the user makes when saying them. Regional accents, gender and age differences, and volume levels can all contribute to errors. Project54 has largely overcome this problem by refining the software for this particular task, and by limiting the vocabulary that the system needs to recognize to the relatively short list of words in the control command set.
The system has been deployed in the patrol cars of the New Hampshire State Police, and has been demonstrated at trade shows. Troopers are eager to demonstrate the capabilities of the patrol car voice command systems, and then have an onlooker jump into the driver’s seat and run the system. For anyone familiar with speech recognition systems, it’s pretty impressive to see a voice command first executed by a middle-aged male trooper, and then immediately thereafter by a 25-year-old female newcomer with no resetting or reprogramming of the equipment.
The speech recognition system is activated by pressing a button on the steering wheel previously dedicated to cruise control (which is seldom used by patrol officers). The driver then speaks a preprogrammed command such as “pursuit” (which activates the car’s emergency lights and siren, while at the same time sending a pursuit notification and a GPS-derived location to the communications center, and puts the patrol video system into “record” mode) or “home” (which changes the radio channel to the primary frequency). When the officer primes the system to accept database inquiries, he can speak the characters of a license plate into the system and have the reply read back by a computer-generated voice over the car’s sound system. Outside of the car, an officer can control the electronics with a Palm-based hand-held computer equipped with an 802.11/Wi-Fi interface.
Other systems deployed that tout similar capabilities have required two and even three separate computers, usually mounted in the trunk, to drive the components of the package. Project54 is driven by a single computer module embedded in the dashboard, usually behind an integrated, touch-sensitive display panel. The display panel doubles as a manual control head for all patrol car electronics, the monitor for the patrol video set, and the display for any traditional computing application such as computer-aided dispatch or word processing software.
A number of police product manufacturers have partnered with UNH to develop consistent interfaces and integration into the Project54 suite, such as Whelen, Decatur, Andrea, Stalker, Kustom, Motorola, Raymarine, Code 3, and Federal Signal. Because the system is being developed by a public institution, licensing costs will be greatly reduced as compared to a private sector enterprise.
The research associated with the Project54 effort has revealed a problem that many agencies didn’t know existed. When electronic devices used in a patrol car are powered by a standard charger or power converter (such as those that attach to a cigarette lighter socket), they often radiate electromagnetic interference (EMI) in the VHF radio bands commonly used by public safety agencies. Use of these devices commonly generates electronic noise at levels of 10 decibels (dB) and higher. Decibels are measured on a logarithmic scale, so an increase in noise from 10 dB to 20 dB does not double the interference— it increases it by a factor of 10. Therefore, a 100-watt radio that could just overcome 10 dB of interference at the fringe of its operating area would have to be increased to 1,000 watts output to get the same performance when the interference was increased to 20 dB. Nearly every jurisdiction has “dead zones” where radio performance is marginal or ineffective. Use of a noisy power source could extend the dead areas substantially by making the radio transmitters less effective.
Many of the interference problems can be remedied by better shielding, grounding or replacement of faulty equipment. The UNH CATlab has developed a system consisting of a battery-operated laptop computer and a computer-controlled radio receiver that plots the signal-to-noise ratio (the proportion of desired electronic emissions to those that are undesired) of patrol car equipment. Suspect devices are turned on and off while transmitting to determine which, if any, is the cause of the problem. The rogue device can then be modified or replaced to restore the transmitter to its previous efficiency.
The University of New Hampshire’s CATlab welcomes inquiries about Project54. They can be reached online at www.catlab.unh.edu.
Tim Dees is a former officer who writes and consults about applications of technology in law enforcement. He can be reached at (509) 585-6704 or by e-mail at firstname.lastname@example.org.