User:Skybum/PRT

PRT System Design
There are currently no agreed-upon standards in PRT system design. Among the handful of PRT systems that are currently developing hardware -- and the many dozens of PRT designs that exist on paper -- there is a tremendous diversity of design approaches. Not only are the designs diverse, they are also in many cases quite contentious. The following sections provide an overview of the primary different design approaches, and highlights the major disputes, where they occur.

Vehicle Design
PRT vehicles carry only two to four passengers in order to reduce their weight. However, this also increases ridership per vehicle, because during idle times every operating vehicle will have a higher ridership (25-50%) than a mass-transit vehicle such as a bus or train (as low as 2% after midnight, 15% during non-rush hours).

Since the U.S. averages 1.16 persons per automobile in commuter areas, many authorities say that the optimum vehicle size in the U.S. for PRT is either 1 or 2 passengers. Some systems (UniModal, Ford Research's PRISM) have found that the weight and cost difference between these sizes of vehicles is so low that two seats is optimum, with tandem seating and a low-drag shape.

Other authorities question the viability of systems with only two seats. The public's worst-case needs are shown by its choice of automobiles, 85% of which have four seats plus or minus one. Groups of three or four commonly travel together. Families with young children may be reluctant to split up. Also a person in a wheelchair with a companion and luggage may not be accommodated. Some PRT vendors therefore have chosen vehicles accommodating three or four passengers with luggage.

According to designer of Skyweb/Taxi2000 J.E. Anderson (below), the lightest-weight system, and therefore the one with the lowest system cost, is a linear induction motor (LIM) on the car, thrusting against a stationary conductive rail for both propulsion and braking. Loss of traction due to precipitation, ice or sudden braking is therefore not an issue, since a LIM's magnetic interaction with the rail would be unaffected. This aspect contributes to the feasibility of short headways between PRT vehicles. LIMs also minimize the number of moving parts in the car, reducing maintenance costs, and lowers the relative fabrication expense for the rail. It's also easy for an on-board computer to control. A similar system was proposed by Doug Malewicki for Skytran.

The Raytheon and ULTra systems use off-the-shelf rotary electric motors. Matra used a "variable reluctance motor" in Aramis.

PRT vehicles are powered by electricity, so pollution is much less. Most systems plan multiply-redundant power supplies, from track-side batteries or natural-gas-powered generators. Stationary power reduces the vehicles' weight.

Most designers eschew line switching built into the track, because failure of an in-track switch would drastically degrade capacity. Vehicle-mounted switches are preferred so that tracks stay in service, and to allow closer spacing of vehicles since no time delay is needed to allow the track to switch. Vehicle-mounted switches may be mechanical or solid-state electromagnetic. Some systems like 2getthere and ULTra do not need switches since the automated steering system merely chooses which path to follow.

Infrastructure Design
Some PRT systems have had substantial extra expenses from the extra track needed to decelerate and accelerate from the numerous stations. In at least one system, Aramis, this nearly doubled the width and expense of the required right-of-way, and caused the nonstop passenger delivery concept to be abandoned. Other systems have schemes to reduce this cost. Control algorithms can space vehicles to reduce turn-out (siding) lengths (see below). Elevated tracks can "vertically merge"(pp.27,33,37) and keep to a narrow right of way

Since systems have minimal waiting times, embarkation stations are very small (inexpensive) and lack amenities such as seating or restrooms. Usually there's only a fare vending machine, a gate or two, a line of vehicles and a security camera. The stations are usually mounted on poles with the track, but may also be inside buildings or at street level.

About 1/3 of the vehicles can be stored at stations, waiting for passengers. Storage facilities need very little space, because the vehicles are automated and interchangeable. So, no access lanes are needed pick out particular vehicles or to hold vehicles of different sizes.

The debate continues over the best guideway for PRT systems. Most systems' guideways would be incompatible with both each other and existing transportation technologies. No technology has been acknowledged by all authorities as clearly superior.

Structurally, some guideway designs are monorail beams, several are bridge-like trusses supporting internal tracks, and others are just cables embedded in a conventional or narrow roadway that can be elevated.

Some points of agreement exist: it should permit fast switching and good braking, be inexpensive, be capable of being elevated, and pleasant to look-at. Ideally, it should not need to be cleared of dust or snow, and able to be built at ground level. Most systems also use the guideway to distribute power, data, and routing indications to the vehicles.

Fast, reliable switching is a key requirement for PRT that rules out some designs. For example, in most monorails, the rail is so heavy that the switch movement time would increase the time between PRT cars so much that the guideway is no longer competitive with a bus.

Designing a power rail for all weather conditions is difficult. For example, glare ice can almost insulate a rail from a vehicle's brushes.

An elevated track structure scales down dramatically with lower vehicle weights. Therefore, the vehicle's weight budget is critical. The heavier the vehicle, the more costly the track, and the track is the gating system cost. As well, large tracks are visually intrusive, so small vehicles contribute to a more attractive track.

The vehicle weight is so critical to capital costs and visual appearance that exotic aerospace techniques can usefully reduce the cost and size of both the vehicle and track.

Most designs put the vehicle on top of the track, because people prefer it. This also makes the poles shorter, with a smaller silhouette. They are said to be stronger and less expensive. Top mounted vehicles are said to unload the skins of the vehicle, which can therefore be lighter. Vehicles on top of tracks also have simpler line-switching, and in low density areas, can be inexpensively mounted on the ground without poles.

Design teams have used similar justifications for cars suspended from (held below) an overhead track. Cars are said to be stressed in tension, "making a lighter vehicle structure" because many materials are stronger in tension than they are in compression. An overhead track is necessarily higher, and therefore more visible, but also narrower, and therefore creates less shadow, while having a small silhouette.

The least expensive real systems have used wheels with linear electric motors for drive and braking. To save money, the controls and electromagnets are mounted in the vehicles. Tight tolerance requirements in such systems can offset the structural cost savings. Taxi 2000 eliminated vehicle suspensions by making running surfaces adjustable. The least expensive structure for an overhead guideway is a rail suspended from a cable (See the aerobus). The fastest (theoretical) system would use magnetic levitation, which had some breakthroughs in 2000. The lowest-energy real PRT vehicles have used air-cushion suspension and drive. Controlled vehicle speeds can avoid vibrations in the structures. Combinations seem possible.

Routing indicators are often bar codes laser-cut from steel plates, and read by the vehicles with non-contact magnetic sensors. This system is unaffected by dust or wear and gives positions with high precision.

Embarkation stations are on turnouts so vehicles can pass vehicles picking up passengers at full speed. Each station has as many berths as required by demand in its service area, with three berths in a 30-40 foot-long station being common. Systems can embark passengers as fast as they can enter a vehicle and sit down.

Operational Characteristics
Another dispute concerns capacity utilization, which directly affects a transit-system's return on investment.

If the peak speeds of PRT and a train are the same, a well-designed PRT is two to three times as fast for a passenger as a well-designed bus or train route, just because the PRT vehicles do not stop every few hundred yards to let passengers on and off.

Therefore for the same maximum speed, PRT theoretically has two to three times as many trips per seat as a bus or train. So PRT should utilize its average seat 50 to 300% more efficiently. This is contested, of course.

Such high route utilizations would let PRT replace a train or high-capacity bus route. If true, PRT could be used in an intermodal transport system, and then expand from a proof-of-concept project into a network.

PRT automatically diverts vehicles to busy routes and travels nonstop at maximum speeds. Simulations with standard assumptions show that at these high speeds, vehicles can be recycled for new trips as often as several times per hour, even during busy periods, even in low-density cities. This yields more trips per hour per vehicle, increasing ridership substantially during rush hour. In simulations of rush hour or high-traffic events like professional sports events, about 1/3 of vehicles on the guideway need to be empty to get the best response time.

At idle times fast speeds do not increase capacity, because no-one wants to travel. However, during rush hour, higher speed allows a smaller fleet serve the same number of passengers. The result is therefore to reduce the absolute fleet size, and the number of idled vehicles during idle times.

The spacing of PRT vehicles on the guideway sets the rate at which the guideway, PRT's major system expense, can be depreciated by traffic. Designers therefore attempt to minimize the headway, the distance between vehicles.

Some PRT designers have planned for very short headways, which can allow a single guideway to carry the same number of passenger miles as four freeway lanes. This dramatically increases the capacity utilization of a heavily-used guideway, and substantially speeds trips through the center of a city, by permitting more use of direct routes. It also permits a PRT guideway to achieve carrying capacities similar to light rail.

Proponents argue that short headways are safe and practical. PRT vehicles normally operate on unshared guideways, on a separate grade from other traffic. This means that emergency braking for side traffic is not required. Since the front vehicle will also be braking, the minimum safe distance between the vehicles will be set by the reaction time of the following vehicle. In most cases, this consists of the brake's mechanical reaction time, and the reaction time of the electronic control system. If the front vehicle electronically signals the following vehicle, and both vehicles use brakes with the same reaction time, the headway might be cut to the on-board computer's response time, which can easily be less than a fiftieth of a second. Even with large safety margins, this permits much closer spacing than the two-second headways recommended for cars.

Very short headways are very controversial. Some regulators (e.g. the British Rail inspectorate, regulating ULTra) are willing to accept two second headways. In these systems, a PRT guideway carries the same number of passenger-miles as a lane of freeway traffic. Most authorities say that regulators may be willing to reduce headways with increased operational PRT experience.

Rail regulations legally apply to PRT systems in some places (See CVS, above). Therefore, some persons calculate headways in terms of absolute stopping distances, using vehicle decelerations taken from rail lines, and then prove that PRT systems are impossibly uneconomic. This method of calculation is traditional in heavy rail systems, because heavy rail normally shares its grade with other traffic and has poor reaction times and very poor brakes compared to vehicle weight. These conditions do not apply to PRT. Some authorities argue that even when used for heavy rail, calculating headways from absolute stopping distance is too conservative.

Simulations with standard assumptions show that PRT, which should be substantially faster than autos in areas with traffic jams, should attract between 35% and 60% of automobile users. In contrast, new light rail systems and bus lines normally attract between 2% automobile users, both in reality, and in similar simulations. In some regions with a long history of rail transit and very high densities (New York is the almost the only U.S. location), new rail and bus systems can attract up to 30% of auto commuters.

Some PRT systems (See Unimodal) plan speeds substantially faster than automobiles achieve on empty expressways. In simulations, these attract even more traffic than slower, conservative PRT designs.

The ridership simulations are disparaged, but have been repeated many times. If true, the high riderships would substantially decrease the cost per rider of PRT compared to trains and buses.

One successful algorithm places vehicles in imaginary moving "slots" that go around the loops of track. Real vehicles are allocated a slot by track-side controllers. The on-board computers maintain their position by using a negative feedback loop to stay near the center of the commanded slot. The vehicles keep track of their position in the slot with on-board speedometers. These have slight measurement errors (about 1%), so to keep the vehicles from bumping, vehicles' position and speed estimates are adjusted as they pass control points on the tracks. The track-side controllers have to keep synchronized with each other, also. The controllers assure that every two moving slots have one vehicle. At intersections "merge" logic manages the four possible combinations.

A slight variation places vehicles on North-South tracks in odd-numbered slots, while East-West vehicles use even-numbered slots. This permits rapid automatic merges and crossing of traffic at intersections. On the straight-aways, adjacent vehicles spread-out, or close-up to reestablish the every-other-slot relation. The alternating slots double the stopping distance in most situations, increasing safety.

Another style of algorithm assigns a trajectory to a vehicle, after verifying that the trajectory does not violate the safety margins of other vehicles. This system permits system parameters to be adjusted to design or operating conditions. This has succeeded in full-scale simulations and small test tracks, and uses slightly less energy. (algorithms are from J.E. Anderson's article, below)

The turn-outs to slow down or speed up for stops can almost double the length of track. Designers often increase the distance between vehicles to trade off lower guideway capacity for shorter, cheaper turnouts. Another trick to reduce turn-out lengths (and expense) is to keep vehicles in bunches (sometimes called "platoons"), and then widen the gap behind a slowing vehicle, and speed up (from a stop) into the end of a bunch.

Vibrations in the guideway can add unnecessary mechanical stress, increasing the cost. Most real systems use vehicle speeds that minimize vibrations in the guideway. Some theoretical designs have explored the use of vehicles' motors to actively damp vibrations in the guideway.

The maker of the ULTra PRT system reports that testing of its control system shows lateral (side-to-side) accuracy of 1 cm, and docking accuracy better than 2 cm.

Safety engineers at PRT companies say that travel via PRT systems should be ten thousand to one million times safer than via cars because of basic design improvements. Computer control is said to be more reliable than drivers. Grade-separated guideways prevent collisions with pedestrians or manually-controlled vehicles. Most PRT systems enclose the running gear in the guideway to prevent derailments. Vehicles usually have computer-diagnosed, dual-redundant motors and electronics. In the event of a total failure, a car can be pushed to a repair facility by another PRT vehicle.

Tracks and vehicles are timed to "miss" at intersections. Careful engineering at several projects has shown that less-expensive one-way, single-level loops can operate as safely and almost as quickly as systems with far more expensive dual-direction clover-leaf intersections.

The Morgantown PRT system (considered group rapid transit by PRT experts) has now completed 110 million injury-free passenger-miles. By comparison, regular transit injures about a hundred people on average in that many passenger miles.

Some systems plan to group vehicles to carry large groups. This also can reduce aerodynamic drag. Groups (called "platoons" or "trains") could share an intercom and destination.

Some systems plan multiple types of vehicles. The smallest vehicles seat two, the largest six. Two has the lowest-per-mile track cost, and handles most trips (average ridership in cars is 1.16 persons per vehicle in the U.S.) Most systems provide for wheel-chair users, bicyclists and light cargo vehicles, sometimes with special vehicles.

Most systems have buttons in a vehicle, such as "let me talk to the operator," "take me to the nearest stop," "take me to the hospital," "take me to the police for help," and "this vehicle is too filthy to use."

Vandalism could be investigated from video of the car, reviewed when the button "this vehicle is too filthy to use." is pressed. The Morgantown System reports very little vandalism, and credits the short wait times and cctv monitoring of the stations.

Cost Characteristics
Many PRT advocates claim that it will have a per-passenger trip costs between $0.01 and $0.10/mile ($0.006 and $0.06/km) -- somewhat cheaper to operate than a moped.

Estimates of guideway cost range between $0.8 million and $22 million per mile., p.89; Although these estimates were realized in ULTra's prototype (which were $1.5 million per mile), many transportation planners disbelieve these unprecedentedly-low estimates.

The capital expense and the rate at which it is paid by passengers is critical, since PRT systems are capital-intensive with low operating costs compared to other technologies.

In all transit systems, vehicles are depreciated on a schedule that accounts for the average number of empty seats per vehicle, and the number of trips per day. This becomes a number called "capacity utilization." When it is higher, fares cover more of the costs of the transit equipment and operators.

Standard transit-planning assumptions concerning overhead per vehicle are said to fail in PRT systems. One major operating expense of bus and light rail systems is the operators' and mechanics' salaries. Additionally, some systems require transit police as well.

PRT systems eliminate driver salaries by automating guidance and fare-collection. Repairs should be less per vehicle because PRTs have electric motors, with one moving part (on most the only moving parts are wheels and the door), versus hundreds for an internal combustion engine.

Transit police are not required because riders are not forced to share a cabin, and criminals cannot easily predict where vehicles will go, and so cannot wait for commuters.

The WVU PRT project failed commercially (though succeeding technically) due to the cost of heating its track to eliminate snow. Some systems in which the vehicles ride atop the track therefore enclose the track to keep precipitation or debris away from the track. Weather is better handled by configurations that suspend vehicles below tracks. Well-designed PRT systems have much lower costs for clearing snow than conventional streets and vehicles.

As for fuel, PRT systems can be powered from the track, and purchase power from the cheapest electric utility. Unlike trains and electric buses, PRTs only accelerate and stop once per passenger, saving substantial energy. Electric motors are non-polluting, and convert power to motion with a total system efficiency of 40 to 90%. The typical automobile is 30% efficient; hybrid cars are 30 to 40% efficient.

Some PRT designers report or estimate very low energy usage. For example, ULTra reports 0.55 megajoules per passenger km, less than 900 BTU per passenger mile. This compares favorably to the standard 3,200 BTU per passenger-mile for rail transit (U.S. Dept. of Energy, Table 2.11).

Still, it is well-known from U.S. federal data that operations and maintenance costs (O&M) are nearly constant per seat for a wide variety of systems: buses, trains, aircraft and private automobiles, which of course lack paid operators.

Some authorities say that even if PRT has the same O&M costs, the increased load factor (O&M/passengers per destination) of PRT (about 0.33) should reduce costs per passenger mile compared to those of other public transit (which are about 0.15).

Planners dispute the cost-estimates of PRT rights-of-way. In modern metropolitan areas, rights-of-way for light rail cost as much as $50 million per mile ($30 million/km). However, a typical light-rail right-of-way is 100 to 300 feet (30 to 100 m) wide, and (naturally) goes through the highest-density, most valuable part of the city. When the railway tunnels to conserve the surface, it becomes even more costly.

PRT rights of way may even cost less than a conventional road system. Proponents say that if auto- and bus-based transit systems include the costs of the roadways needed for buses and automobiles (US $10x10^6 per mile, or $6x10^6 per km), PRT systems are substantially cheaper than bus and automobile systems.

The larger number of vehicles does not increase costs. Costs of transit vehicles are relatively constant per passenger. While larger vehicles enclose more space, they are nearly hand-built. A fleet of smaller vehicles can be mass-produced, as the auto industry shows.

Urban Integration
In mass transit with scheduled service, this "ridership" factor is generally calculated for an entire system, then applied to all vehicles. On most trips of most routes, vehicles are 85% to 95% empty, and only rush-hour trips on important central routes approach vehicle (and route) capacities. The low ridership of bus and trains therefore often causes a substantial cash drain through depreciation and the salaries paid for operators and mechanics. Further, the drain cannot be offset by fares.

In PRT, automated fare collection, driving and security reduce the variable costs of capacity. Also, PRT vehicles will only move in response to demand, or in timely expectation of demand (i.e., waiting in stations; moving to occupy empty station berths and then waiting). This idling of PRT vehicles that are in-service but not in use should save energy compared to scheduled transit modes, in that the proportion of the number of seats in moving vehicles should be proportional to the number of people using the system at a particular moment. In contrast, a bus or train moves a large proportion of empty seats during non-peak periods.

The surprisingly cheap per-mile estimates of PRT designers (see above) depend on dual-use rights of way. By mounting the transit system on narrow poles, placed on an existing street, PRT designers hope to use land very economically. Small PRT vehicles with passengers can weigh as little as 1,000 pounds (450 kg), while conventional rail systems with many passengers often weigh tens or hundreds of thousands of pounds.

In some circumstances, such as at airports, PRT's small size can reduce the volume of its tunnel to less than a quarter of that required for an automated people mover (APM). Even accounting for two PRT guideways to match the capacity of one APM guideway, the tunnel volume (hence cost) will be less than half.

In the U.S., systems must provide service to disabled persons. Some advocates say that a bus system to provide free disabled service is cheaper than elevators at each embarkation station, and this meets legal requirements, but this is an untested legal theory.

Dual mode systems use existing roads, as well as special-purpose PRT guideways. The particular advantage is that they can reduce the initial expense of the guideway network, without losing convenience because of network effects. In some cases, the guideway is just a cable buried in the street. Proponents of single-mode (i.e., the vehicle operates only on the guideway) say there is a low threshold to realize network benefits, because capital costs per unit are already low (see Advantages).

There are several concerns about the appearance of a PRT system.

People near the guideway are most affected by its shadows. In this view, more sunlight is better, because the sunlight falling on the guideway is useless to people. So, guideways should have minimal horizontal structure.

Another view says that the guideway is ugly and its visibility is most apparent in long sight lines. In this view, the silhouette of the guideway should be minimized.

Most planners assume that a competent industrial design will provide an attractive appearance for the PRT vehicle.

Theoretically, parking lots can be far smaller for shopping centers, universities, stadiums and convention centers, freeing valuable land. Roads or rails are required for heavy transport.