Traveler Information > Route Guidance

• Introduction
  • What Is Route Guidance?
  • The Rationale for Route Guidance
• System Description
  • Dynamic versus Static
  • Digital Map Databases
  • Positioning Technologies
  • Map-Matching
  • Route Planning
  • Route Guidance
  • The Human-Machine Interface
  • Communications
  • Vehicle Location and Navigation
• Assessment
  • ADVANCE
• Implementation Challenges
• References

INTRODUCTION

What Is Route Guidance?

History
Route guidance is used in many contexts: in transit and commercial fleets that track vehicles and dispatch drivers using wireless location technologies such as beacons, microwave signals or satellites; on talking buses and trains that announce destinations automatically; in platform and station signs that give riders real-time information; and in train dispatch and control systems. Increasingly, however, route guidance is coming to refer to systems that communicate information about a route to the driver of a vehicle, usually through an on-board device. This report addresses this last type of route guidance, also known as in-vehicle route guidance. (Discussions of transit applications of route guidance can be found here, and discussions of its use in commercial freight fleets are here.)

The first route guidance system dates back to roughly 2600 B.C. in China, when a device called the south-pointing carriage was invented. It used a gear mechanism that is the ancestor of today's differential odometer to keep a mechanical human figure pointed perpetually south no matter what direction the carriage turned. Another Chinese device invented at about the same time was the precursor to the modern odometer. It had a gear that caused two mechanical figures to beat on a drum each time the carriage moved a certain distance. One figure sounded a beat every time the carriage traveled a single unit of measurement, called a "li," and the second one's sounded every 10 lis. (These observations and much of the basic information about route guidance presented here are adapted from the 1997 book, Vehicle Location and Navigation Systems, by Yilin Zhao. Complete bibliographic information about it can be found at the end of the report.)

The first modern route guidance systems were patented in the U.S. in the early decades of the 20th Century, as automobile travel increased, before maps, street naming systems and signage could catch up with drivers' needs. Much like the Chinese invention, the early systems used an odometer to determine the vehicle's location. They then issued instructions, usually in the form of a pre-printed message, in real time. Such systems could only work, however, along pre-specified routes that the machine "knew."

Interest in route guidance devices flagged in the 1920s, as road markings and maps improved. In World War II, the U.S. military developed one of the earliest electronic route guidance systems. It used an on-board computer to calculate a vehicle's path from a set starting point and expressed its ending location in map coordinates.

Starting in the 1970s, technological advances such as increases in computers' power to make calculations, more reliable and robust wireless communications and more expressive visual display devices enabled the development of more sophisticated route guidance systems, with an explosion of interest and ideas starting in the 1990s. Most of the research and deployment in this period has taken place in Europe, Japan and the U.S.

Components

As outlined in Zhao's Vehicle Location and Navigation Systems, route guidance systems are usually comprised of:

• a digital map database;
• a system that synthesizes signals or sensor data to locate the vehicle on a map;
• a route planning function that designs a path before or during a trip, according to pre-selected criteria or preferences;
• a route guidance function, which directs the driver along the planned route;
• a device or devices that serve as an interface between the human user and the system; and
• a one- or two-way wireless communication system.

See our Telecommunications Diagram on In-Vehicle Route Guidance for more information.

The Rationale for Route Guidance

Route guidance systems are intended to enable a driver to take the route that most closely matches his requirements. Usually, this means the one that is shortest, fastest or least congested (or some combination of those). Ideally, the route guidance system will be robust enough to help drivers disperse themselves efficiently so that they fully utilize all available routes. Otherwise, there is a danger of simply transferring the congestion from one portion of the transportation network to another.

SYSTEM DESCRIPTION

Dynamic Versus Static Route Guidance

The earliest and most widely deployed of the modern route guidance systems use a static database of geographic information about the region in which the vehicle is traveling. It is usually pre-loaded on some high-capacity storage device, such as a CD-ROM or DVD, that can be accessed by the driver. The driver can obtain directions by entering his starting point and desired destination, both of which must be already stored in the database. This method is the simplest to deploy, assuming the existence of a reliable, digital geographic database. It is popular in Japan because there is no uniform system of naming streets. However, the market for such systems in the U.S. has not been as strong. U.S. drivers tend to place little value on systems that give directions or location information. The use of such systems has been limited for the most part to rental cars, where they have been tested in different parts of the country.

Most recently, systems have been developed that give drivers directions based on dynamic information about traffic and other road conditions that is conveyed over a continuous, two-way communications link between the vehicle and a central information center. The center automatically tracks the location of the vehicle and also collects and analyzes information from sources such as sensors, loops and probe vehicles to determine congestion, road conditions and other factors that could affect the choice of a "best" route. The analysis and recommendations can be carried out automatically by a central computer or in total or in part by a concierge service staffed by live attendants. As an example of a fully autonomous system, a German company has introduced an in-car navigation system whose on-board device receives constantly updated information about the vehicle's location and road conditions in the area where the vehicle is traveling and determines on its own whether to advise a change in route. By contrast, many of the global-positioning satellite (GPS)-based Mayday and road assistance systems that U.S. automakers are building into their vehicles link the vehicles to a live help desk or communications center staffed around the clock.

Digital Map Databases

• create a graphical display of a region;
• find a location by using a street address or intersection;
• devise a route for a driver;
• carry out map-matching—comparing information from other sources against the map database for route corrections and confirmation;
• steer the driver along a route;
• tell the driver where his vehicle is on the road network or whether it has left the network; and
• supply information about roadside attractions, events, facilities and landmarks.

Maps in route guidance systems are usually either:

• digital pictures of paper maps, which have been created by scanning them into a computer's memory in the form of graphical units, called pixels, and stored as "raster-encoded" files, or
• a conglomeration of data, stored digitally in "vector-encoded" files, which are bits of information that can be put together in many different combinations or can be combined with data from other sources to create many different structures.

Raster-encoded, graphical maps are faster to make, but they take up a lot of computer memory and do not contain any more information than what is visible. Vector-encoded, data-file maps take longer to create, but they require less computer memory and can be used in more ways because the data in their files can interact with data in other databases and with itself. Increasingly, route guidance systems are using both types of files, the digitally scanned maps for visual displays and the vector maps for route guidance functions.

A vector-encoded digital map database can have any degree of cartographic detail, which, because it is stored in digestible amounts, can be used immediately by a route guidance system's location and navigation components. Examples of the cartographic information are:

• types of roads (arterials, highways, or local streets, for example),
• street names,
• address ranges for a certain segment of the region,
• restrictions on vehicle operations such as one-way travel, weight limits and turn restrictions,
• landmarks, and
• the presence of signs or signals.

The digital database map is usually supplied in raw digital form by a vendor. The route guidance system uses a software program called a compiler to shape the raw data into files for use in the route guidance system. However, the raw digital cartographic information must be in format, scale, coordinates and base calculations that are compatible with those that the route guidance system uses and with other sources of digital cartographic information that may also be incorporated into the route guidance system. There are a number of proprietary, commercial digital map databases. Two of the larger vendors are Etak and Tele Atlas BV Navigation Technologies (NavTech). (The rest of this discussion will be concerned only with vector-encoded databases.)

Though much of the work is automated, creating digital map databases is still labor-intensive, and the accuracy and degree of detail of the database is very dependent on the quality of the measurements and calibrations that are added in the final, human-based processing. Accuracy of the database files is extremely important. Errors on a scale as small as 15 meters (approximately 45 feet) can significantly interfere with the function of the entire system. So it is important to use only well-made digital map databases.

The database can be stored in a central computer that is an active part of the system, or it can be kept on a high-capacity storage device such as a CD-ROM or DVD.

Positioning Technologies

Because a route guidance system needs a highly detailed description of the vehicle's position in order to operate smoothly and provide useful information, the component that determines the vehicle's position uses more than one source of information, either sensors of the same type that cover overlapping regions, or a range of different types of sensors in the same region (loop detectors, closed circuit cameras, and probe vehicles, for example). The data from these multiple sources are fused and synthesized by computer programs and mathematical models in order to compensate for inherent biases.

Sensors can be absolute or relative. Absolute sensors locate the vehicle by calculating its position as measured from fixed features of the earth's geography. Magnetic compasses and GPS systems are the two most widely used absolute sensors. Relative sensors track the vehicle's location in comparison to some pre-determined location. An odometer is a widely used relative sensor, which measures the number of rotations of a vehicles' wheels from a certain starting point. Odometers linked to right and left wheels are differential odometers. They can measure the vehicle's heading over the course of a trip by comparing the rotations of the right wheel to the left wheel and factoring in the overall forward progress.

Positioning technologies can use standalone calculations, usually dead reckoning, or calculate positions by communicating with an outside source of information, usually satellite signals in the form of GPS. There are also systems that use radio signals from earth stations. Dead reckoning is the least expensive but the most error-prone. It involves calculating the vehicle's position by measuring the distance it has traveled from a predetermined starting point and the chronological length of the journey. Dead reckoning tends to become less accurate as the journey progresses because slight inaccuracies in ongoing measurements accumulate and gradually distort the calculations. GPS is very accurate, but it relies on costly satellite technology.

Map-Matching

Map-matching compares a vehicle's route to a predetermined set of roads in the vehicle's vicinity. In a pattern-recognition process, it uses algorithms to predict the vehicle's position based on these absolute references, which adjust for errors in dead reckoning systems. This is possible on road networks because cars do not usually stray very far from the designated network. Parking lots, driveways and medians are about as far off the network as they get, which means that the algorithm has a universe of alternatives from which to choose that is small enough for it to make valid selections most of the time.

Route Planning

The route planning function can be carried out for an individual vehicle or for many vehicles on a network. The plan is made according to a set of criteria, commonly known as travel costs. Some drivers may prefer to travel shorter distances, even if the time is longer. Others may prefer to travel a long distance to avoid congestion.

The information used to plan the route can be either static or dynamic. Static information could use historical data to make adjustments for recurring influences such as time of day or day of the week, but it wouldn't incorporate real-time information. Dynamic information can have that capability, which makes it considerably more helpful. However, such information can be difficult to obtain to a degree of accuracy sufficient to earn the confidence of the driver.

Route Guidance

Route guidance can be either pre-trip, usually in the form of a printout or map, or en-route. En-route guidance requires substantial computational power, along with a navigable map database, a positioning system and location and route planning capabilities. A real-time route guidance system must continually update the vehicle's position, its speed, direction and location as compared to the map of the route network in the immediate vicinity. It also must continually calculate and communicate the maneuvers the driver must make to follow the planned route.

The Human-Machine Interface

Telling the driver to make a maneuver usually involves voice commands, with or without a visual aid. The timing of the announcements and their phrasing is critical for the system to be effective and safe to use. Some measures of a system's responsiveness to human needs are: the number of navigation errors the driver makes; the time he needs to complete the route; the time he spends looking at the road ahead and mirrors (the more, the better); his mental workload; and the quality of his driving, which is measured roughly by how well he maintains consistent speeds, headways and lane position.

Some broad design rules of usability are:

• consistency,
• predictability,
• intuitiveness, i.e., a text-based order will read in the left-right/right-left and top-down/bottom-up order in which ordinary text is read in the prevalent culture,
• grouping elements by function,
• making the functions that are used most often or are most important the easiest to do, and
• transparency and immediacy in the way that the information is presented, so that the user doesn't need to remember very much information for very long.

Control devices can be: foot push buttons, joysticks, keyboards, knobs, switches, rockers, toggles, rotary or slide switches, touch screens or voice recognition readers. Display devices need to be visible in different light conditions and from many different locations in the vehicle; they have to be able to show graphics, catch the driver's eye and be comprehensible even when the driver's eyes are busy. The audio output must be audible under different noise conditions (i.e., windows rolled up or down, driving alone or in a crowded car, radio or music player on or off).

Communications

If a route guidance system uses dynamic information—not just data from a library that is stored on-board—it needs to communicate with outside information sources such as a traffic information center, a concierge desk (in the case of a Mayday system), roadside beacons or even other vehicles. These communications systems, which are wireless, are usually either ground-based or satellite-based.

The ground-based systems are:

• paging and other personal communications services, private mobile radio systems (such as those used to dispatch fleets), and cellular communications,
• radio data networks (RDNs), which use unassigned radio frequencies to broadcast data,
• broadcast subcarriers, which use space left over on an allocated frequency and are received by special equipment (commonly used for subscription services such as background music, weather and soundtracks),
• radio data systems (RDS), which broadcast data on an inaudible subcarrier which can be picked up by low-cost receivers (common in Europe),
• the radio broadcast data system (RBDS), a U.S. variant on RDS, that includes RDS and extends it, which is proving popular in Europe and Japan,
• short-range beacons for vehicle-to-roadside communication in which microwave or infrared beacons transmit short bursts of data at high speeds over short distances, typically from roadside furniture and signs; they can be location beacons, which transmit their own location and identifying number; information beacons, which also relay current traffic information that they receive via cable; and communications beacons, which can collect data from the vehicle as well.

Satellite-based systems have earth stations for transmitting or receiving signals. Geosychronous (GEO) satellites, remain over the same spot at very high altitudes (more than 22,000 nautical miles), which requires their earth stations to have large antennas. Low-earth-orbit (LEO) satellites orbit in circular or elliptical patterns at altitudes that rarely are greater than 1,000 nautical miles and require much smaller antennas. GEO satellites are the most widely used, but LEOs may gain in popularity as technology is developed to more fully exploit their advantages.

There a number of important aspects to wireless communications. They are:

• coverage, the area over which a system can transmit and receive data and voice messages, which can be affected by the power of the transmitter, the antenna's height and gain (how well its power stays concentrated in any given direction), the characteristics of the communication channel, the presence of interference, the quality and sensitivity of the receiver, the protocols used in the communications system and the topography of the area;
• capacity and throughput, the amount of data the system can handle;
• costs, which can vary widely, with pager-based systems typically on the low end and satellite-linked telecommunications being the most expensive; and
• connectivity, whether they are capable of one- or two-way communications.

Vehicle Location and Navigation

Simple location systems use stand-alone technologies, the most common being dead reckoning, which was described earlier. Another stand-alone technology for locating a vehicle is computer vision. It requires a pre-existing library of images or landmarks on the route, but it doesn't require communicating with an outside source. If the library of images is sufficiently large, and the machine vision sufficiently acute, it can be very accurate. It can also fill in the gaps caused by interference or distortions in a system that uses two-way communications.

A Global Positioning System (GPS) uses signals from 24 satellites to calculate the vehicle's location. It is a one-way system. Typically, the users only receive data. However, EutelTracs, the European satellite location system, allows two-way communication between a terminal on-board the vehicle and EutelTracs' GEO satellites.

ASSESSMENT

ADVANCE (Advanced Driver and Vehicle Advisory Navigation Concept)

The following narrative about the ADVANCE project is taken from the reports published on the project's Web site, which is listed at the end of this report.

ADVANCE (Advanced Driver and Vehicle Advisory Navigation ConcEpt) is a public/private partnership conceived and developed in 1991. Participants included the Federal Highway Administration (FHWA), the Illinois Department of Transportation (IDOT), the University of Illinois at Chicago and Northwestern University, under the auspices of the Illinois Universities Transportation Research Consortium (IUTRC), Motorola, Inc., and the American Automobile Association (AAA).

In 1995 and 1996, the consortium conducted a test and evaluation of an advanced route guidance system that was designed to use dynamic real-time information about traffic and road conditions delivered to the driver via an interactive map display, which the driver could tailor to his routes. The system carried out route planning and route guidance functions. Reports and evaluations of the test deployment were published in 1996 and 1997 and can be found on the ADVANCE Web site. The consortium is now developing a system of traffic and travel information-gathering and analysis to support a centralized, Web-based traffic information center for the Gary-Chicago-Milwaukee corridor. At the time this report was written, it was in the user-testing phase. To learn about its current status, visit the ADVANCE web site.

The system architecture of ADVANCE incorporates several key concepts: distributed intelligence (all route planning is performed in the vehicle); a hierarchical road network database (with higher-level roadways, such as freeways, having the greatest emphasis); vehicles as traffic probes (for accumulating real-time information); an open (non-proprietary) radio frequency data communications protocol; and a driver interface. The system had four subsystems: a central processing facility, traffic algorithms, in-vehicle route planning and display functions, and a communications system.

The test involved 110 drivers in 80 volunteer households living in the suburbs northwest of Chicago. The drivers used the ADVANCE vehicle guidance system and responded to both baseline (pre-test) and post-test surveys. Thirty-two of these drivers participated in focus groups. Drivers also maintained written logs in which they described any rerouting they did based on advice from the ADVANCE system.

The baseline survey captured driver and household demographics, trip making and driving experience, sources and use of traffic information, experience with common technologies, and personality attributes. The post-test survey asked about the test experience, evaluation of specific ADVANCE features, risk factors associated with ADVANCE, preferences for features in future systems, and the willingness to pay for such systems. The focus group results provided a rich qualitative perspective on driver experiences with ADVANCE and preferences for design features and performance characteristics of future systems.

This test was limited by several important factors. First, the sample size was small and non-random. In this case, the group was generally well-educated and higher income, which is not necessarily representative of the population in the study area or the broader driving population. Second, the test period was short. Third, the ADVANCE system offered limited functionality, and, in particular, very little real-time traffic data. Despite these limitations, a number of findings from the surveys, focus groups, and reroute logs provide consistent, logical, and potentially important directions for the development of future in-vehicle route guidance systems.

Drivers reported that the ADVANCE routes very often tended to be inferior to the ones they chose. This is attributable to the facts that drivers knew the network, congestion patterns and routes that best served their routine trips; the ADVANCE network and travel time database were imperfect; and the ADVANCE route planning algorithm, by policy, placed a priority on suggesting routes along roadways at the top of the hierarchy rather than the neighborhood streets which familiar drivers commonly used for parts of their trips.

Because of their experience-based knowledge, familiar drivers seemed to prefer and felt they would benefit from a substantial degree of control over their choice of routes and route planning criteria. At the same time these drivers expressed a high level of interest in real-time traffic information, particularly information about non-recurring congestion. They were interested in blending such real-time information with their own knowledge to plan their routes.

Familiar drivers in this test seemed to define a different role for the in-vehicle route guidance system than the role underlying the design of ADVANCE: they envisioned an on-board computer guidance system as an intelligent assistant. They were less interested in giving routing control to the computer. Instead, they seemed to envision using the technology to acquire and process real-time data, and to use those data to evaluate driver-provided routes and, where appropriate, to recommend alternatives to those routes.

Drivers perceived and evaluated route guidance systems in two principal dimensions: the route guidance function and its performance; and the driver interface, including data input and information output functions.

Both focus group and survey results revealed patterns of gender and personality differences in both responses to ADVANCE and preferences for future system attributes; these help define the breadth of capabilities that should be considered for future route guidance systems.

Men seemed more inclined than women to follow the guidance system's advice literally and to be more disappointed when it failed or came up short. Women were more likely than men to incorporate their own knowledge when deciding whether to accept the advice. At the same time, they seemed more willing to let the computer make more decisions for them if that meant simpler presentations on the display screen.

In expressing preferences for future systems, both genders placed a high value on dynamic route guidance and real-time information. Women were not as interested in a map showing their locations. They expressed more support than men did for a "help" button, voice instructions, the power to plan trips with multiple stops and the ability to choose destinations by type rather than name. Men had stronger preferences for static information.

People who described themselves as cautious tended to be more forgiving of miscalculations by the guidance unit than drivers who described themselves as confident. The confident drivers expressed stronger preferences for information-rich elements such as congestion maps and customized criteria for choosing a route, but the sample was too small to produce a statistically reliable picture. For complete reports, go to the ADVANCE Web site.

IMPLEMENTATION CHALLENGES

(This section is adapted from Vehicle Location and Navigation Systems, by Yilin Zhao (1997), and other readings, including the ADVANCE reports.)

System Design
Modern route guidance systems rely on sophisticated location and navigation technologies. For these to function properly, they must have access to a large volume of accurate, real-time traffic and road condition data. They must also have a robust, stable architecture and be able to interact seamlessly with numerous different components. The architecture must also allow for frequent system upgrades and evolutions. Another critical challenge is integrating map databases with one another.

Safety and Human Factors
Many human factors issues remain to be resolved concerning the functionality and safety of route guidance systems that drivers use while they are operating a vehicle. Ease of use and safety are closely linked, though it remains a possibility that even the best-designed systems could pose a safety risk. In 1999, the Japanese Ministry of Transportation conducted one of the earliest studies that tried to identify accidents in which the driver's use of in-car navigation systems was a contributing factor, but it was not conclusive because of a lack of data. More research remains to be done. Attempts in the U.S. to determine the importance of drivers' use of cell phones as a contributing factor to accidents have been similarly inconclusive, though many localities have passed laws banning drivers from operating hand-held sets while driving.

Acceptance/Business Models
The market forces driving the design of a navigation system are affected by cultural expectations, environments and driving patterns. Part of the reason for the devices' popularity in Japan is the lack of usable street addresses and names. Travelers typically use landmarks to find their destinations. By contrast drivers in the U.S. and Europe, where locations usually have street addresses, expect and use turn-by-turn instructions. In the U.S. in particular, because street addresses are so reliable, drivers place less value on systems that help them find their destination and more value on those that increase their sense of personal safety and the integrity of their vehicle. That has spurred the growth of Mayday services, which are available in nearly all new U.S. cars, starting with the 2001 model year.

Each of these capabilities requires different parts of the route guidance system to be more robust, making it unlikely that it is economically feasible to develop a single system that can execute them all well. The landmark-based navigation preferred by the Japanese relies more heavily on graphics, scrolling of maps and changing scales of identifiable features, which require graphics-processing power rather than calculating power and speed. By contrast, the turn-by-turn directions preferred in the West require rapid route calculations and quick access to data. The Mayday systems prevalent in the U.S. only require a vehicle location system and a communications link to a service desk.

Clearly, considerable additional research and testing needs to be carried out on nearly every aspect of route guidance technology.

REFERENCES

Vehicle Location and Navigation Systems, Yilin Zhao, 345 pp. Artech House, Inc. Norwood MA. 1997. A clear, complete description of the major concepts. Information is presented in a way that remains timely even in this rapidly changing subject area.

ADVANCE Information Source (home page of Advanced Driver and Vehicle Advisory Navigation Concept). (last visited October, 2001). Extensive materials on the architecture, design, execution and evaluation of the ADVANCE system.

Links

TravInfo Evaluation - San Francisco
http://www.path.berkeley.edu/%7Eloukakos/travinfo.html

Washington State road information
http://www.wsdot.wa.gov/Rweather/

Author: Phyllis Orrick.  Last update: 03/01/02