Crossings and Roundabouts

The early bird gets the worm, but itīs the second mouse that gets the cheese.

	B eams have to cross each othersī path here and there. That situation is usually solved in either of 2 ways: As a shunt (alternative A in the illustration at right) As a 2-level crossing (alternative B). There is, however, an interesting solution to the problem of letting 2 beams cross each other on the same level. We present that solution a bit down on this page. Both of the alternatives above have both advantages and drawbacks. The shunt does not require sloping beams (to achieve different levels), and enables the cars to change direction by switching onto the other beam, if required. But at least one of the beams will need curves, which reduces speed. The 2-level crossing provides for high speed on both beams, and does not require timeslotting, since the cars donīt get in each otherīs way. In order to avoid too steep slopes and too great height over the ground at the stop areas, so that the elevator won't reach all the way down, the beam traffic network should be limited to 2 levels on those parts that are above ground. This limitation need not apply, then, to bridges, tunnels, indoor facilities and garages for the beam cars.	On this page will briefly look at: One-way traffic Two-way traffic Crossings on the same level

1. One-way Traffic

Figure 1 Figure 2 or one-way traffic (figure 1) there should not be a problem to shunt the cars in the desired direction in an intersection (alternative A). One could also have this intersection in one plane by turning it into a combined shunting and diverging node (alternative B i figure 1). There will be no problem in alternative A unless one wants to be able to alter the traffic direction on the beam from time to time. Entrances and exits to stops that are also used as car buffers will have to be planned in accordance with the direction of the traffic flow. Figure 3

n the case of two-way traffic, there is still no problem for cars wanting to turn right (assuming there is right-hand traffic on the beams) or going straight ahead in a 2-plane crossing. A left turn will require a more complex solution, such as a roundabout. The alternative in figure 2 requires quite a lot of space, but can in an urban environment be implemented by, for instance, putting one beam around a neighboring block (B in the figure). The B route will enable cars coming from the north (or up in the figure) to turn east. Figure 2:1

roundabout as shown figure 3 enables shunting in all directions, while keeping the intersection in one plane. Same plane croosings can be used even at high traffic throughputs, for instance by enabling cars turning right to use a separate roundabout, outside the first one, as is shown in figure 2:1. In the case of very high traffic intensity, one will have to use a 2-plane intersection, as is shown in figure 2:2. In this case, the routes marked B will have to go close to the ground, resulting in street-level traffic. Excellent place for a stop. Figure 2:2

The optimal solution to the 2-way, 2-level crossing problem is to do as in figure 2:3, provided that there is space enough. The beamcar at A, coming from north, has the options of going straight ahead (1), to turn west by following path 2, or to turn east by following path 3. The advantage with this solution is that vehicles going straight ahead need not slow down, as they have to do in a roundabout. Figure 2:3

3. Crossings on the Same Level

o let the beams cross each other in the same plane would be practical in many situations. 3 problems would arise out of this arrangement: 1) Problem: The wheels of the propulsion car would have to cross the slit of the intersecting beam. This slit is broad enough to require supports for the wheels, so they won't get stuck or destroyed. Possible solution: The wheels of the propulsion car will push down small bridges as they arrive to the slit, one for each wheel (shown by A in figure 6). These bridges would spring back in a vertical position afterwards, so the slit B would normally be free. This solution requires that the traffic always travel in the same direction on both beams. 2) Problem: The risk for collisions will have to be eliminated. The angle of the scanning radar in front of the beam vehicle would have to be 180 degrees in order to discover crossing traffic. Or, separate radars on the sides of the vehicle have to be used. Naturally, a functioning system controlled by computers would not allow 2 cars to reach the intersection at the same time, but the computerized functions should not be relied upon entirely when it comes to safety. Possible solution: Both nodes closest after the crossing on their respective beams would have to operate as one node of the converging kind. Figure 7 illustrates how this would work. Nodes 1, 2 and 5 are normal divergence nodes, while 3 and 6 are normal convergence nodes.

Suppose (to simplify matters) that nodes 2, 3, 5 and 6 are symmetrically positioned around the crossing. Then 5 and 2 (or, rather a point a bit south of node 2) would be booking points for node 6, and 2 and 5 (almost) would be booking points for node 3. Thus far, everything would function in accordance with the rules for "point synchronous networks as described elsewhere.

ut apart from this, nodes 3 and 6 would have to function as one common convergence node, with booking points at 2 and 5. This whole planning would be less complicated if we were to assign common booking points for the whole intersection. These booking points would then preferably have to be moved to to the next foregoing nodes, according to standard procedure. That is to say, node 1 and a point equally far removed from the intersection, but on the other beam (i.e. point 4) would be booking points.

As a tangible illustration, if the vehicle A is travelling straight ahead, then A would have to let B pass, since B has passed its booking point and A has not. This would apply regardless whether B is about to make a turn or travel straight on. If A however is about to make a right turn, then A, for the same reason, would have to wait for both B and C to pass (considering that B has already passed its booking point).

However, 2 is a divergence node, and is thus informed how B is going to shunt. Consequently, if node 2 reports that B is going to turn left, then A should not have to take B into consideration. Simple, when you think about it, isn't it?

3) Problem: The communication between the vehicles on this part of the beams must be ensured. The nature of this problem is that the wave guides on each section of the beam network are isolated from each other. A section, in this context, being the part of a beam that goes from one node to the next. The nodes then function as mediators, communicationwise, between these sections.

One could say that each nodes "owns" the sections leading up to it, insofar as the vehicles travelling upon it would be handled by this node. In figure 7 the nodes 3 and 6 would be collectively responsible for the intersection, and this could only work if they behave as one node. This problem would be even more accentuated at large terminals and traffic places. Should the problem be solved for these places, it would also automatically be solved for same-plane intersections.