CHAPTER 6. CONCLUSION
71
for both traffic densities.
|
controller |
density |
ATWT |
TWQL |
|
marching |
HD |
304 |
2346 |
|
optim |
HD |
406 |
2467 |
|
request |
HD |
234 |
1294 |
|
phase |
HD |
242 |
1164 |
|
platoon |
HD |
x |
x |
|
marching |
LD |
282 |
1128 |
|
optim |
LD |
332 |
1040 |
|
request |
LD |
45 |
6 |
|
phase |
LD |
116 |
101 |
|
platoon |
LD |
24 |
0 |
Table 6.2: Scenario 2: Best results
The request, phase and platoon controller are better than the optim controller.
The request controller at high traffic density reduces the ATWT with 51% and
the TWQL with 47% compared with the optim controller. For low traffic density
there is a reduction of 86% for ATWT and 99% for TWQL.
For the phase controller the reduction is 50% for ATWT and 52% for TWQL,
at high traffic density. At low traffic density, the controller reduces the ATWT
with 65% and the TWQL with 90%.
For the platoon controller a problem is discovered at high traffic density. When
all roads are congested, and the conditions for platoon is crossing are satisfied,
almost all traffic is blocked. This problem can be solved with an extra rule for the
controller (one of the following):
• the case platoon is crossing is stopped when a maximum time has reached.
• when a possible destination lane is full, the case platoon is crossing is
stopped when a maximum time has reached.
The first rule can cut platoons when no problem occurs, the second rule need
sensors at all inputs and outputs of a junction but will not cut a platoon when no
problem occurs. It is also possible to give a bigger value to busses or other longer
vehicles when the condition platoon is crossing is checked.