Editing Personal rapid transit (section)

==System design==
Among the handful of prototype systems (and the larger number that exist on paper) there is a substantial diversity of design approaches, some of which are controversial.

===Vehicle design===
Vehicle weight influences the size and cost of a system's guideways, which are in turn a major part of the capital cost of the system. Larger vehicles are more expensive to produce, require larger and more expensive guideways, and use more energy to start and stop. If vehicles are too large, point-to-point routing also becomes more expensive. Against this, smaller vehicles have more surface area per passenger (thus have higher total air resistance which dominates the energy cost of keeping vehicles moving at speed), and larger motors are generally more efficient than smaller ones.

The number of riders who will share a vehicle is a key unknown. In the U.S., the average car carries 1.16 persons,<ref>Skytran Web Site: See "common sense"</ref> and most industrialized countries commonly average below two people; not having to share a vehicle with strangers is a key advantage of [[private transport]]. Based on these figures, some have suggested that two passengers per vehicle (such as with [[skyTran]], EcoPRT and Glydways), or even a single passenger per vehicle is optimum. Other designs use a car for a model, and choose larger vehicles, making it possible to accommodate families with small children, riders with bicycles, disabled passengers with wheelchairs, or a [[pallet]] or two of freight.

====Propulsion====
All current designs (except for the human-powered [[Shweeb]]) are powered by [[electricity]]. In order to reduce vehicle weight, power is generally transmitted via lineside conductors although two of the operating systems use on-board batteries. According to the designer of Skyweb/Taxi2000, [[J. Edward Anderson]], the lightest system uses [[linear induction motor]] (LIM) on the vehicle for both propulsion and braking, which also makes manoeuvres consistent regardless of the weather, especially rain or snow. LIMs are used in a small number of rapid transit applications, but most designs use [[rotary motor]]s. Most such systems retain a small on-board battery to reach the next stop after a power failure.  CabinTaxi uses a LIM and was able to demonstrate 0.5 second headways on its test track.  The Vectus prototype system used continuous track mounted LIMs with the reaction plate on the vehicle, eliminating the active propulsion system (and power required) on the vehicle.

[[ULTra]] and 2getthere use on-board batteries, recharged at stations. This increases the safety, and reduces the complexity, cost and maintenance of the guideway. As a result, the ULTRa guideway resembles a sidewalk with curbs and is inexpensive to construct.  ULTRa and 2getthere vehicles resembles small automated electric cars, and use similar components.   (The ULTRa POD chassis and cabin have been used as the basis of a shared autonomous vehicle for running in mixed traffic.<ref>{{cite news |title=Westfield Technology Group autonomous POD confirmed for Fleet Live 2019 |url=https://www.fleetnews.co.uk/news/fleet-industry-news/2019/08/01/westfield-technology-group-autonomous-pod-confirmed-for-fleet-live-2019 |access-date=28 June 2021 |date=1 August 2019 |archive-date=28 June 2021 |archive-url=https://web.archive.org/web/20210628030728/https://www.fleetnews.co.uk/news/fleet-industry-news/2019/08/01/westfield-technology-group-autonomous-pod-confirmed-for-fleet-live-2019 |url-status=dead }}</ref>)

====Switching====
Almost all designs avoid [[Railroad switch|track switching]], instead advocating vehicle-mounted switches (which engage with special guiderails at the junctions) or conventional steering. Advocates say that vehicle-switching permits faster routing so vehicles can run closer together which increases capacity.  It also simplifies the guideway, makes junctions less visually obtrusive and reduces the impact of malfunctions, because a failed switch on one vehicle is less likely to affect other vehicles.

Track switching greatly increases headway distance. A vehicle must wait for the previous vehicle to clear the junction, for the track to switch and for the switch to be verified.  Communication between the vehicle and wayside controllers adds both delays and more points of failure.  If the track switching is faulty, vehicles must be able to stop before reaching the switch, and all vehicles approaching the failed junction would be affected.

Mechanical vehicle switching minimizes inter-vehicle spacing or headway distance, but it also increases the minimum distances between consecutive junctions. A mechanically switching vehicle, maneuvering between two adjacent junctions with different switch settings, cannot proceed from one junction to the next. The vehicle must adopt a new switch position, and then wait for the in-vehicle switch's locking mechanism to be verified. If the vehicle switching is faulty, that vehicle must be able to stop before reaching the next switch, and all vehicles approaching the failed vehicle would be affected.

Conventional steering allows a simpler 'track' consisting only of a road surface with some form of reference for the vehicle's steering sensors. Switching would be accomplished by the vehicle following the appropriate reference line – maintaining a set distance from the left roadway edge would cause the vehicle to diverge left at a junction, for example.

===Infrastructure design===
[[File:Diagramatic PRT layout2.png|thumb|250px|right|Simplified depiction of a possible PRT network. The blue rectangles indicate stations. The enlarged portion illustrates a station off-ramp.]]

====Guideways====
Several types of guideways have been proposed or implemented, including beams similar to monorails, bridge-like [[truss]]es supporting internal tracks, and cables embedded in a roadway. Most designs put the vehicle on top of the track, which reduces visual intrusion and cost, as well as easing ground-level installation. An overhead track is necessarily higher, but may also be narrower.  Most designs use the guideway to distribute power and data communications, including to the vehicles. The [[Morgantown Personal Rapid Transit|Morgantown PRT]] failed its cost targets because of the steam-heated track required to keep the large channel guideway free of frequent snow and ice. Heating uses up to four times as much as energy as that used to propel the vehicles.<ref>{{cite web |title=The History of my Involvement in PRT and how it led to ATRA |url=https://www.inist.org/library/2019-10.Anderson.History%20of%20my%20involvement%20in%20PRT%20led%20to%20ATRA.pdf |website=INIST |access-date=3 July 2021 |archive-date=9 July 2021 |archive-url=https://web.archive.org/web/20210709185013/https://www.inist.org/library/2019-10.Anderson.History%20of%20my%20involvement%20in%20PRT%20led%20to%20ATRA.pdf |url-status=dead }}</ref>  Most proposals plan to resist snow and ice in ways that should be less expensive. The Heathrow system has a special de-icing vehicle. Masdar's system has been limited because the exclusive right-of-way for the PRT was gained by running the vehicles in an undercroft at ground-level while building an elevated "street level" between all the buildings.  This led to unrealistically expensive buildings and roads.<ref name="whymasdarscaleback"/>

====Stations====
Proposals usually have stations close together, and located on side tracks so that through traffic can bypass vehicles picking up or dropping off passengers. Each station might have multiple berths, with perhaps one-third of the vehicles in a system being stored at stations waiting for passengers. Stations are envisioned to be minimalistic, without facilities such as rest rooms. For elevated stations, an elevator may be required for accessibility.

At least one system, Metrino, provides wheelchair and freight access by using a cogway in the track, so that the vehicle itself can go from a street-level stop to an overhead track.

Some designs have included substantial extra expense for the track needed to decelerate to and accelerate from stations. In at least one system, Aramis, this nearly doubled the width and cost of the required right-of-way and caused the nonstop passenger delivery concept to be abandoned. Other designs have schemes to reduce this cost, for example merging vertically to reduce the footprint.

===Operational characteristics===

====Headway distance====
Spacing of vehicles on the guideway influences the maximum passenger capacity of a track, so designers prefer smaller [[headway]] distances. Computerized control and active electronic braking (of motors) theoretically permit much closer spacing than the two-second headways recommended for cars at speed. In these arrangements, multiple vehicles operate in "platoons" and can be braked simultaneously. There are prototypes for [[autonomous car|automatic guidance of private cars]] based on similar principles.

Very short headways are controversial. The UK Railway Inspectorate has evaluated the ULTra design and is willing to accept one-second headways, pending successful completion of initial operational tests at more than 2 seconds.<ref>[http://www.advancedtransit.org/wp-content/uploads/2011/08/Sustainable-personal-transport-M.-Lowson.pdf Sustainable personal transport]</ref> In other jurisdictions, preexisting rail regulations apply to PRT systems (see CVS, above); these typically calculate headways for absolute stopping distances with standing passengers. These severely restrict capacity and make PRT systems infeasible.  Another standard said trailing vehicles must stop if the vehicle in front stopped instantaneously (or like a "brick wall").  In 2018 a committee of the [[American Society of Mechanical Engineers]] considered replacing the "brick wall" standard with a requirement for vehicles to maintain a safe  "separation zone" based on the minimum stopping distance of the lead vehicle and the maximum stopping of the trailing vehicle.<ref>{{cite web |title=ASCE APM STANDARDS COMMITTEE ACCEPTS ALTERNATIVE TO BRICK WALL STOP |url=http://www.advancedtransit.org/library/news/asce-apm-standards-committee-accepts-alternative-brick-wall-stop/ |website=Advanced Transit |date=11 May 2018 |access-date=3 July 2021}}</ref>  These changes were introduced into the standard in 2021.

====Capacity====
PRT is usually proposed as an alternative to rail systems, so comparisons tend to be with rail. PRT vehicles seat fewer passengers than trains and buses, and must offset this by combining higher average speeds, diverse routes, and shorter headways. Proponents assert that equivalent or higher overall capacity can be achieved by these means.

=====Single line capacity=====
With two-second headways and four-person vehicles, a single PRT line can achieve theoretical maximum capacity of 7,200 passengers per hour. However, most estimates assume that vehicles will not generally be filled to capacity, due to the point-to-point nature of PRT. At a more typical average vehicle occupancy of 1.5 persons per vehicle, the maximum capacity is 2,700 passengers per hour. Some researchers have suggested that rush hour capacity can be improved if operating policies support ridesharing.<ref>{{cite web
 | url = http://pubsindex.trb.org/document/view/default.asp?lbid=803547
 | title = Doubling Personal Rapid Transit Capacity with Ridesharing
 | last = Johnson | first = Robert E.
 | year = 2005  | access-date = August 30, 2017
 | work = Transportation Research Record: Journal of the Transportation Research Board, No. 1930
}}</ref>

Capacity is inversely proportional to headway. Therefore, moving from two-second headways to one-second headways would double PRT capacity. Half-second headways would quadruple capacity. Theoretical minimum PRT headways would be based on the mechanical time to engage brakes, and these are much less than a half second. Researchers suggest that high capacity PRT (HCPRT) designs could operate safely at half-second headways, which has already been achieved in practice on the Cabintaxi test track in the late 1970s.<ref>{{cite web
 | url = http://faculty.washington.edu/jbs/itrans/big/soa2.pdf
 | title = Emerging Personal Rapid Transit Technologies
 | last = Buchanan | first = M. |author2=J.E Anderson |author3=G. Tegnér |author4=L. Fabian 
 | author5=J. Schweizer
 | year = 2005  | access-date = August 30, 2017
 | work = Proceedings of the AATS conference, Bologna, Italy, 7–8 November 2005
 }}</ref> Using the above figures, capacities above 10,000 passengers per hour seem in reach.

In simulations of rush hour or high-traffic events, about one-third of vehicles on the guideway need to travel empty to resupply stations with vehicles in order to minimize response time. This is analogous to trains and buses travelling nearly empty on the return trip to pick up more rush hour passengers.

[[Grade separated]] light rail systems can move 15,000 passengers per hour on a fixed route, but these are usually fully grade separated systems. Street level systems typically move up to 7,500 passengers per hour. Heavy rail subways can move 50,000 passengers per hour per direction.  As with PRT, these estimates depend on having enough trains.

Neither light nor heavy rail scales operated efficiently in off-peak when capacity utilization is low but a schedule must be maintained.  In a PRT system when demand is low, surplus vehicles will be configured to stop at empty stations at strategically placed points around the network. This enables an empty vehicle to quickly be despatched to wherever it is required, with minimal waiting time for the passenger.  PRT systems will have to re-circulate empty vehicles if there is an imbalance in demand along a route, as is common in peak periods.

=====Networked PRT capacity=====
The above discussion compares line or [[corridor capacity]] and may therefore not be relevant for a networked PRT system, where several parallel lines (or parallel components of a grid) carry traffic. In addition, Muller estimated<ref>{{Cite web |url=http://www.leighfisher.com/trb/657-2-05-0599.pdf |title=Muller et al. TRB |access-date=2006-09-25 |archive-url=https://web.archive.org/web/20060831081723/http://www.leighfisher.com/trb/657-2-05-0599.pdf |archive-date=2006-08-31 |url-status=dead }}</ref> that while PRT may need more than one guideway to match the capacity of a conventional system, the capital cost of the multiple guideways may still be less than that of the single guideway conventional system. Thus comparisons of line capacity should also consider the cost per line.

PRT systems should require much less horizontal space than existing metro systems, with individual cars being typically around 50% as wide for side-by-side seating configurations, and less than 33% as wide for single-file configurations. This is an important factor in densely populated, high-traffic areas.

====Travel speed====
For a given peak speed, nonstop journeys are about three times as fast as those with intermediate stops. This is not just because of the time for starting and stopping. Scheduled vehicles are also slowed by boardings and exits for multiple destinations.

Therefore, a given PRT seat transports about three times as many passenger miles per day as a seat performing scheduled stops. So PRT should also reduce the number of needed seats threefold for a given number of passenger miles.

While a few PRT designs have operating speeds of {{convert|100|km/hour|mph|abbr=on}}, and one as high as {{convert|241|km/hour|mph|abbr=on}},<ref>The concept-level SkyTran system is proposed to travel at up to [https://web.archive.org/web/20120130035414/http://www.skytran.us/index.php?option=com_content&view=article&id=28&Itemid=18 241 km/h (150 mph) between cities]</ref> most are in the region of {{convert|40-70|km/hour|mph|abbr=on}}. Rail systems generally have higher maximum speeds, typically {{convert|90-130|km/hour|mph|abbr=on}} and sometimes well in excess of {{convert|160|km/hour|mph|abbr=on}}, but average travel speed is reduced about threefold by scheduled stops and passenger transfers.

====Ridership attraction====
If PRT designs deliver the claimed benefit of being substantially faster than cars in areas with heavy traffic, simulations suggest that PRT could attract many more car drivers than other public transit systems. Standard mass transit simulations accurately predict that 2% of trips (including cars) will switch to trains. Similar methods predict that 11% to 57% of trips would switch to PRT, depending on its costs and delays.<ref name="EDICT"/><ref name=AndreassonRidership>{{cite web| last=Andreasson| first=Ingmar| title=Staged Introduction of PRT with Mass Transit| url=http://www.princeton.edu/~alaink/Orf467F10/PRT@LHR10_Conf/stagedIntroPRT_Andreasson_paper.pdf| publisher=KTH Centre for Traffic Research| access-date=2013-10-12| archive-url=https://web.archive.org/web/20131014230259/http://www.princeton.edu/~alaink/Orf467F10/PRT@LHR10_Conf/stagedIntroPRT_Andreasson_paper.pdf| archive-date=2013-10-14| url-status=dead}}</ref><ref name=YoderRidership>{{cite web| last=Yoder| title=Capital Costs and Ridership Estimates of Personal Rapid Transit| url=http://faculty.washington.edu/jbs/itrans/yoder.htm| access-date=12 October 2013|display-authors=etal}}</ref>

====Control algorithms====
The typical control algorithm places vehicles in imaginary moving "slots" that go around the loops of track. Real vehicles are allocated a slot by track-side controllers. Traffic jams are prevented by placing north–south vehicles in even slots, and east/west vehicles in odd slots. At intersections, the traffic in these systems can interpenetrate without slowing.

On-board computers maintain their position by using a [[PID controller|negative feedback loop]] to stay near the center of the commanded slot. Early PRT vehicles measured their position by adding up the distance using [[odometer]]s, with periodic check points to compensate for cumulative errors.<ref name="FundOfPRT" /> Next-generation [[GPS]] and radio location could measure positions as well.

Another system, "pointer-following control", assigns a path and speed to a vehicle, after verifying that the path does not violate the safety margins of other vehicles. This permits system speeds and safety margins to be adjusted to design or operating conditions, and may use slightly less energy.<ref name="ControlPRT">{{cite web
 | url = https://www.telenor.com/wp-content/uploads/2012/05/T03_1.pdf
 | title = Control of Personal Rapid Transit Systems
 | pages = 108–116
 | publisher = Telektronikk
 | date = January 2003
 | access-date = August 30, 2017
}}</ref> The maker of the ULTra PRT system reports that testing of its control system shows lateral (side-to-side) accuracy of 1&nbsp;cm, and docking accuracy better than 2&nbsp;cm.

====Safety====
Computer control eliminates errors from human drivers, so PRT designs in a controlled environment should be much safer than private motoring on roads. Most designs enclose the running gear in the guideway to prevent derailments. Grade-separated guideways would prevent conflict with pedestrians or manually controlled vehicles. Other public transit [[safety engineering]] approaches, such as redundancy and self-diagnosis of critical systems, are also included in designs.

The Morgantown system, more correctly described as a [[#Group rapid transit|Group Rapid Transit]] (GRT) type of [[Automated guideway transit|Automated Guideway Transit]] system (AGT), has completed 110 million passenger-miles without serious injury. According to the U.S. Department of Transportation, AGT systems as a group have higher injury rates than any other form of rail-based transit (subway, metro, light rail, or commuter rail) though still much better than ordinary buses or [[car]]s. More recent research by the British company ULTra PRT reported that AGT systems have a better safety than more conventional, non-automated modes.{{Citation needed|date=May 2008}}

As with many current transit systems, personal passenger safety concerns are likely to be addressed through CCTV monitoring,<ref>{{cite journal |last1=Muller |first1=Peter J. |last2=Young |first2=Stanley E. |last3=Vogt |first3=Michael N. |title=Personal Rapid Transit Safety and Security on University Campus |journal=Transportation Research Record: Journal of the Transportation Research Board |date=January 2007 |volume=2006 |issue=1 |pages=95–103 |doi=10.3141/2006-11|s2cid=110883798 }}</ref> and communication with a central command center from which engineering or other assistance may be dispatched.

====Energy efficiency====
The [[energy efficiency in transport|energy efficiency]] advantages claimed by PRT proponents include two basic operational characteristics of PRT: an increased average load factor; and the elimination of intermediate starting and stopping.<ref>{{cite web|url=http://citeseer.ist.psu.edu/594390.html|title=CiteSeerX}}</ref>

Average load factor, in transit systems, is the ratio of the total number of riders to the total theoretical capacity. A transit vehicle running at full capacity has a 100% load factor, while an empty vehicle has 0% load factor. If a transit vehicle spends half the time running at 100% and half the time running at 0%, the ''average'' load factor is 50%. Higher average load factor corresponds to lower energy consumption per passenger, so designers attempt to maximize this metric.

Scheduled mass transit (i.e. buses or trains) trades off service frequency and load factor. Buses and trains must run on a predefined schedule, even during off-peak times when demand is low and vehicles are nearly empty. So to increase load factor, transportation planners try to predict times of low demand, and run reduced schedules or smaller vehicles at these times. This increases passengers' wait times. In many cities, trains and buses do not run at all at night or on weekends.

PRT vehicles, in contrast, would only move in response to demand, which places a theoretical lower bound on their average load factor. This allows 24-hour service without many of the costs of scheduled mass transit.<ref>{{citation
 | author = Anderson, J. E.
 | year = 1984
 | title = Optimization of Transit-System Characteristics
 | publisher = Journal of Advanced Transportation, 18:1:1984, pp. 77–111
}}</ref> <!--This article may be viewed using Google's cache at:
https://scholar.google.com/scholar?hl=en&lr=&safe=off&q=cache:D7SIAd2tkzgJ:www.skywebexpress.com/pdf_files/150k_additional/OTSCpaper.pdf+PRT+%22operating+cost%22+%22demand+responsive%22
-->

ULTra PRT estimates its system will consume 839 BTU per passenger mile (0.55 [[megajoule|MJ]] per passenger km).<ref name="Lowson">{{cite web
 | last = Lowson
 | first = Martin
 | url = http://www.advancedtransit.org/wp-content/uploads/2011/08/A-New-Approach-to-Sustainable-Transport-Systems-M.-Lowson.pdf
 | title = A New Approach to Sustainable Transport Systems
 | year = 2004
 | access-date = August 30, 2017
}}</ref><ref>The conversion is: 0.55 MJ = 521.6 BTU; 1.609 km = 1 mi; therefore, 521.6 x 1.609 = 839</ref> <!-- #####I'm removing the below SkyTran section for now, because the reference link seems to be bad - I only get a page full of dead links. If that page is restored, we can restore this section. -ATren, April 2008.#####
SkyTran, a PRT concept using significantly smaller vehicles than other designs, may require only 11 horsepower (9 KW) to cruise at 160 km/h (100 mph), which translates to 151 BTU/passenger mile or 0.1 MJ per passenger km.  However, SkyTran's predicted energy usage is unconfirmed in real world practice, since no SkyTran system or prototype has yet been built. Also, Skytran's small vehicle does not permit disabled passengers, which would require accommodation using other, less energy-efficient modes.<ref name="Malewicki">{{cite web
 | last = Malewicki
 | first = Douglas
 | url = http://www.skytran.net/18EnergyEff/02Energy.htm
 | title = (doc) SkyTran's Super Energy Efficiency
}} Note that this page presents a comparison of seating arrangements; the actual numbers shown for the planned 2-passenger tandem seating arrangement are 10.65 horsepower and 8.85 kilowatts.  The English unit calculation is 8.85 kW / 2 passengers * 3412 (BTU/hour)/kW / 100 mile/hour = 151.0 BTU/passenger mile.  The metric calculation is 8.85 kW / 2 passengers * 3.6 (MJ/hour)/kW / 160 km/hour = 0.0996 MJoule/passenger km.
</ref>--> By comparison, cars consume 3,496 BTU, and personal trucks consume 4,329 BTU per passenger mile.<ref name="edbk">{{cite web
 | publisher = U.S. Dept. of Energy
 | url = http://cta.ornl.gov/data/chapter2.shtml
 | title = Transportation Energy Databook, 26th Edition, Ch. 2, Table 2-12
 | year = 2004
}}</ref>

Due to PRT's efficiency, some proponents say solar becomes a viable power source.<ref>{{cite web
 | year = 2003
 | url = http://www.solarevolution.com/solutions/presentations/ATRA20061118.xls
 | title = ATRA2006118: Solar PRT, p.89
 | publisher = Solar Evolution
 | format = Xcel Spreadsheet
 | access-date = 18 November 2006
 | archive-date = 30 March 2007
 | archive-url = https://web.archive.org/web/20070330035545/http://www.solarevolution.com/solutions/presentations/ATRA20061118.xls
 | url-status = dead
 }}</ref> PRT elevated structures provide a ready platform for solar collectors, therefore some proposed designs include solar power as a characteristic of their networks.

For bus and rail transit, the energy per passenger-mile depends on the ridership and the frequency of service. Therefore, the energy per passenger-mile can vary significantly from peak to non-peak times. In the US, buses consume an average of 4,318 BTU/passenger-mile, transit rail 2,750 BTU/passenger-mile, and commuter rail 2,569 BTU/passenger-mile.<ref name="edbk"/>