Structure of the Beam Network

A conclusion is arrived at when you get tired of thinking.

his is a non-technical page that can be read by anyone, and we will devote it to a short description how an automatic transportation system should be built, geographically speaking. It is important for the proper function of the network that it does not follow the traditional star-shaped configuration, which is the dominating structure of metropolitan public transport systems (Figure 1). This would lead to unnecessary congestion in the central town. One should instead spread the nodes out as much as possible.	And, because the nodes are one of the three traffic limiting factors of the network (the other factors are the speed limit on the beams and the safety distances between vehicles), the network capacity stands to gain from having as many nodes as possible. The benefits of a finer mesh Preventing gridlock

The Benefits of a finer Mesh

The reason that traditional public transport systems tend to converge at certain points is because they are mass transit systems, insofar as the passengers are bunched together for transport efficiency, and have to share the cars/buses/whatever with each other, and thus have to be transported to common "exchange points". This need is completely done away with in the automated beam network.

e (at SwedeTrack System) foresee that there might be vehicles taking as many as 32 passengers, although an upper limit should preferably be set at 16 passengers. But those big vehicles would be intended for: Groups and parties that want to travel together, such as school classes, tourists People who are comfortable with travelling along certain routes at certain times, as in today's mass transit. They would be offered this alternative at a considerably reduced fare. Handling many people at once, such as before and after football games During those hours that there is a sufficient demand, there would be scheduled tours along certain routes, since it would (probably) be cheaper for the passenger to travel with one of these vehicles than ordering individual vehicles.		This is (of course) because passenger handling capacity increases with more passengers per vehicle. It is not self-evident that this would be the case in the long run, however. Scheduled tours with large vehicles might disappear altogether, as the beam network grows, becomes more finely meshed and reaches more and more destinations, thus gradually eliminating the need for passengers to walk to the nearest station where they would assemble in huge numbers. Thus, more stops, more widely dispersed, would mean fewer passengers at each individual stop. This reduces the demands on the stops; they would not need to be so large, and it would give the average traveler shorter distance to walk to the nearest stop. The capacity of the network to convey passengers between 2 given points would grow, as a side effect, because the vehicles would be given more alternative routes betwen these 2 points, and this would thus reduce the need for large vehicles. This is illustrated at left.

Summing up, then: 1. The main reason that today´s public transport bunch commuters together in trains and buses is that wages to drivers and conductors are a sizeable part of the transport utilities´ expenses. This consideration does not apply to automatically driven vehicles. 2. The main reason that automatic public transports today also bunch commuters together is to increase transport capacity. This consideration does not apply to a network that is so extensive and finely meshed that cars can still make headway by choosing the right route from a great number of alternative routes.

In the central parts of a town, the beam network would tend to follow the checkerboard pattern of the streets. Outside city centers and in the suburbs, the network would resemble a spider's web (Figure 2). Between nearby cities there would be a net of trunklines, like todays' roads and railways. You can make a comparison between these 3 network layouts on Figure 3. The bottom picture shows the trunk connections between a group of nearby cities or communities. The middle picture is an enlargement of one of these communities, and the top drawing shows an enlargemenet of the central parts of a city. This network structure has every advantage over present public transport structures: It distributes the traffic more evenly, which alleviate the congestion of downtown areas It enables alternative routes of travel, should the shortest route be blocked It provides for shorter connections and travel times by allowing the beamcars to calculate and travel the shortest route It serves suburbs and out-of-the-way places better than today´s buses and trains.		In the city environment, the easily installed beams can be used in many ways: In broad streets, two-way traffic could be run on two beams placed above the middle of the street. If automobil traffic is successively removed, one suburb at a time, trees could be planted in such a way that the beams are mostly hidden behind the greenery. For smaller streets, one-way beam traffic might be more suitable. If these beams are reserved for small vehicles, the beams would not have to be as large as in the bigger streets. You can read more about how well the beams would handle the crowding and the transport needs of inner cities. One should also take due consideration of where to place the shunts, relative to each other.

Preventing Gridlock

nyone who has visited a city with real traffic congestion cannot avoid the conclusion that there are either too many vehicles for the roads available, or the road grid is poorly designed (or both). With a "poor roadgrid" is meant that, although street capacity might be sufficient, intersections might be too far apart, have cumbersome traffic control, be one-way, etc. Conventional road networks are often subject to “gridlock” in which a region of the network is completely saturated and motion only occurs at the area´s periphery. Unless steps are taken to prevent its occurrence, gridlocks are also possible in a non-deterministic PRT network. But; there are some fundamental differences between PRT and the road network. The most basic is that PRT can have a reasonable level of control on the number of vehicles operating on the system. The number of vehicles can be limited, and even a dual-mode system, and a system where privately owned beam vehicles are allowed, may exercise some control on entering vehicles. Also, the trajectory generation algorithms can efficiently utilize guideway space to maximize the static capacity. Here, static capacity refers to the number of vehicles the system (or a loop in the system) can contain before gridlock occurs and it would be measured in vehicles or perhaps vehicles per unit distance. This is different from the normal definition of roadway capacity – termed dynamic capacity – which is a measure of flow (e.g. number of vehicles per hour that pass a certain point, or vehicles times distance/hr for the whole fleet of network vehicles). Asynchronous PRT networks which utilize safe trajectories will experience similar dynamic capacity curves as roadways. Dynamic capacity will initially grow as a function of line speed, up to a point, then peak, and after that drop, as permitted speed increases. Static capacity is of course maximized at rest. Once the system exceeds the capacity of the given line speed, the real speed will drop towards increasing static capacity. Eventually all vehicles will stop. Furthermore, once a region begins to slow down, random events can trigger significant variations in average speeds – the region becomes less stable. This will result in some passenger discomfort, similar to stop-and-go traffic.

The obvious solutions to this are: Not to attempt solving traffic demand by just supplying more beamcars to the network. The network has to grow correspondingly, either geographically or with a finer mesh, so that the number of guideway-kilometers keep pace with the number of beamcars. To ensure that intersections can never be blocked. With PRT, the basic challenge is to ensure that no segment ever reaches static capacity. This capacity is a function of the types of vehicles on any segment and is easily calculated as each new vehicle enters and leaves the segment. This last requirement would, for example, mean that whenever a new vehicle wishes to switch to a diverging branch (this is any output from a diverging intersection), it would require permission to do so. This could not be accomplished in the road network, and in reality it would not function in this manner in an automatically controlled traffic system, either. Suppose the request for diverging a certain way was denied, what would the beamcar do then? Rather, the traffic pattern has to be planned in advance. Figure 10 To reduce stop-and-go conditions in high-density regions, a local controller may track vehicle travel information and give vehicles the information sufficiently far in advance. If a branch is unavailable, the vehicle may be forced onto a different branch. If no branch is available, the vehicle would have to stop prior to the decision point, so that it could proceed onto any branch once able to do so. In low-density traffic areas with reduced infrastructure control, gridlock would generally be unlikely, so it should be sufficient for each vehicle to verify that it can fully enter its desired segment before proceeding into any diverge. This method still permits “gridlock” in the sense that multiple lines or a complete circuit might be blocked. However, by maintaining clear intersections, vehicles will be able to operate at the periphery and the gridlock would not develop to the point where vehicles become immobile.

That method is also applicable to station control. Vehicles would simply not be permitted to diverge into a station area unless they could clear the diverge. This rule could conceivably permit some flexibility. For example, if there are at least one vehicle in the station which is ready to depart, the arriving vehicle may be permitted to temporarily block the line, thus opening up a space for the departing vehicle, and then enter the station. The example in figure 10 below shows 4 vehicles that have left their original berths, and are now queueing to get out; berth number 4 is thus about to be empty. In heavy traffic, the vehicle Ve would normally forge ahead, and, at position A prevent vehicle 1 from leaving. Instead, it slows down at B, creating an empty timeslot that vehicle 1 can use, thus allowing vehicles 2, 3 and 4 to move forward, creating the needed gap so that Ve can enter the station. This whole procedure necessitates that the booking point Bo is so far off, that the node computer has sufficient time for this traffic directing. So, this would require some ahead-planning on the part of the node controller at the station, but it would avoid a detour for the arriving beamcar, improve station throughput, and cause relatively little disruption to line flow. More complex station areas, with varying vehicle and passenger profiles, may require more sophisticated control. These could resemble "holding patterns" for aircraft, waiting to land at airports, or, better known to the everyday motorist; circling the block, waiting for a parking space to become empty. For example, if a station has a single “parking spot” for a disabled/slow user and it is presently in use, an arriving disabled passenger may be required to circle until the berth is clear or at least ready to depart. Otherwise, the arriving passenger might block station operation until the slot was available. Some traffic simulations and traffical flexibility is advisable to weigh the trade-offs of the likely duration of any blockage vs. the cost to the user and the system of circling. If these occurrences are frequent, then clearly some sections of the beam network needs to be expanded. In principle, there should be sufficient resources of all kinds available (such as berthing space and buffer zones for vehicles) so that "holding patterns" should never have to be employed.