应急车辆信号优先控制的两阶段模型

1
, . 2000 , 1929 , Bachelder
: 2010−06−10; : : 2011−04−28. (70973032); (2009GXS5D130).
, , ,
[1]
, 1994
643 , . . ,
. . , ,
[2] [3]
, Paniati ; Nelson
; ,
4
:
[4]
trem = C − (t1 + tEV )%C trec = trem + nC − tsti
K
(19) (20) (Байду номын сангаас1) ; t1 , ,
trec =
k=1
C (k)
%
,C
, m, n ; trem ; trec
; C (k )
k
,
,
K.
,
trec , k
4
[7]. trec
2
Q(k) =
j=1 l=1
; (
t′ sti .
, . (11) (12) (13) ; t′ clear ; L′ QS
t′′ green
′ ; SEV
; tswit ; t′ EV ; L′ slow . , ,
. . ,
1)
t′′ green3 t′′ green3
; Tallred1 2)
=
t′ EV
− (Tyellow1 + Tallred1 + Tyellow3 + Tallred3 + t′ clear + tsti )
i=1
gi (k) + Tsun C (k) gi (k) Cmax
i gmax
(24) (25) (26) (27)
Cmin
i gmin K
C (k) = trec =trem + nC − tsti
k=1
4 2
Q(K ) =
j=1 l=1
QI jl (K )
¯ Q
(28) (29)
i i−1 Qi + AVi jl (k ) = max{0, Qjl jl (k ) − DV jl (k )}
4
:
i i AVi jl (k ) = qjl (k )g (k ) i i DVi jl (k ) = µjl λjl g (k ) 4 Q0 jl (k ) = Qjl (k − 1)
497
(30) (31) (32) , , , ,j l i ; Qjl (k ) ;l , 1,
i
2 ; DV jl (k ) i
; Tyellow1 . ,
(14)
′ ′ ′ t′′ green3 = tEV − (Tyellow1 − tyellow1 + Tallred1 + Tyellow3 + Tallred3 + tclear + tsti )
(15)
t′ yellow1
3) (a)
.
,
t′′ green3
(b)
=
LVT = LQS + Lslow LVT
.
2.2
(1) ; Lslow
; LQS
.
tEV = L/SEV tEV ; SEV
, , ;L . , , ,
(2)
494
26
. 1)
[11]
, (3)
tclear = (LQS + Lslow )/Sdis tclear
; Sdis .
tsti tsti tEV − tclear
L5 min t′ yellow3
(tsti + tstartup + (LQS + Lslow )/Sdis + Tyellow3 − t′ yellow3 + Tallred3 )SEV
. , . (10) .
2 3 4 5 L = max{L1 min , Lmin , Lmin , Lmin , Lmin }
80 29 138
63 25 113
87 36 159
52 22 96
70 36 142
VISSIM
3.3(m/s),
3 VISSIM Table 3 Distance between detector and intersection in VISSIM simulation
3
.
/(m) /(m)
Abstract: According to operational characteristics of emergency vehicles, a two-phase model was built to convert and optimize signal. The setting distances of vehicle detector and emergency vehicle detector were analyzed, and the first stage model to guarantee that emergency vehicle gets through intersection in safe and without delay, and the second stage model to minimize queue length of ordinal vehicles were modeled. With VISSIM simulation, the model was validated to emergency vehicle getting through intersection without delay in peak and off-peak traffic, while delay of other ordinary vehicles was reduced. Key words: signal preemption; emergency vehicle; intersection; VISSIM simulation
Qjl (k),
,j
k = 1, 2, · · · , K l
, .
(22)
Qjl (k)
k K
.
K 4 2
k
,
.
4 i i (qjl (k) − µi jl · λjl )g (k )}}
Min{
k=1 j=1 l=1
max{0, Q4 jl (k − 1) +
i=1
(23)
s.t.
I
C (k) =
1 Table 1
. VSSIM 1
Basic data at intersection
, .
,
( /h) ( /h) (s)
230 1 350 23
348 1 450 31
216 1 350 23
305 1 450 31
234 1 300 22
308 1 400 28
198 1 300 23
280 1 400 28
i
, ,
4; j ; AVjl (k )
i
,
1, 2, 3, 4
k , j l i ; g (k )
; µjl
i
k i
i
,j i ; qjl (k)
l k i
i
; λjl
k , j j l
l
k j 1, 0,
; Tsun l
¯ ;Q i l i
,
µi jl =
j
4
4.1
VISSIM
VISSIM .
=
t′ EV
− (Tyellow3 + Tallred3 + t′ clear + tsti )
, , .
(18)
t′′ green3
.
3.2
, . 2 .
,
496
26
2 Fig. 2 Chart of different stages during emergency vehicle signal preemption
(4)
L1 min
2) ,
(tsti + (LQS + Lslow )/Sdis )SEV
(5)
L2 min tstartup
3) .
(tsti + tstartup + (LQS + Lslow )/Sdis )SEV
(6)
, , (a) , .
.
L3 min Tallred3
(b) ,
(tsti + tstartup + (LQS + Lslow )/Sdis + Tallred3 − t′ allred3 )SEV
t′ EV
− (Tyellow3 + Tallred3 + t′ clear + tsti )
,
(16)
′ ′ ′ t′′ green3 = tEV − (2Tyellow3 − tyellow3 + Tallred3 + tclear + tsti )
(17)
(c) ,
t′′ green3
′ Tgreen3
Two-phase model of emergency vehicle signal preemption control at intersection
XIE Bing-lei, HU Zheng, ZHAO Hang
(Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen 518055, China)
141 620
138 609
159 686
142 624
498
26
4.2
1
VISSIM
1
Table 4
4,
4 Simulation data of the first phase real-time model /(m) t/(s) 26 98 115 53 21 87 69 100 91 42 5 8 9 13 16 5 8 14 18 24 /(s) 49 29 43 44 45 42 51 33 25 44 14 0 7 21 0 7 0 13 44 7 28 131 21 35 43 71 80 94 147 122 7 102 66 65 7 80 56 0 21 87 7 125 15 72 29 116 72 100 122 58 80 73 51 14 103 34 22 56 37 21 88 75 73 177 35 14 73 65 58 136 136 58 58 0 102 56 29 94 94 7 109 42 50 89 131 22 58 95 28 116
(9)
2 3 4 5 L1 min , Lmin , Lmin , Lmin , Lmin
(10)
3
3.1
, .
, :
4
:
495
1) 3) ),
; 2) , ; 4) ,
′ ′ t′′ green = tEV − (tswit + tclear + tsti ) ′ t′ EV = L/SEV ′ ′ t′ clear = tstartup + (LQS + Lslow )/Sdis
26 2011
4 8 JOURNAL OF SYSTEMS ENGINEERING
Vol.26 No.4 Aug. 2011
,
(
: , . VISSIM , . : ; : U491.2 ; ; VISSIM :A ,
,
,
. , ,
518055)
: 1000−5781(2011)04−0492−08
493
; Qin , .
[5]
; Mirchandani
[6]
, ,
. ,
, ,
[8−10]
Smooth, add only Dwell . , . , , ;
, , “ , , . ”,
2
2.1
, : , , 1 . ; ;
, .
1 Fig. 1 Vehicle classification at intersection
; t′ allred3 .
(7)
L4 min
′ Tgreen3
′ (tsti + tstartup + (LQS + Lslow )/Sdis + Tgreen3 − t′ green3 + Tyellow3 + Tallred3 )SEV
(8) ;
; t′ green3 . ,
Tyellow3
(c)
.
, ,
4.3
100%,
. 4 5 , 6
5 Results of phase optimization with genetic algorithm /(s) 3 15 21
5.
VISSIM ,
.
Table 5
1 1 2 15 23
2 44 29
4 55 18
6 VISSIM Table 6 Queue length at intersection after the second phase in VISSIM simulation /(m)
,
Table 2
10
,
,
2
.
2 VISSIM Queue length at intersection under normal cycle signal in VISSIM simulation /( )
M STDEV [M + 2STDEV]
63 16 95
101 20 141
70 15 100
98
67
42
65
26
67
58
72
495
7 VISSIM Table 7 Analysis of vehicle stopped delay in VISSIM simulation /(s) /(s) 23.4 0 −100% 53.1 0 −100% 36.9 31.1 −16% 46.3 35.3 −24% 44.3 62.9 42% 55.5 72.1 30% 44.2 38.2 −13% 53.2 47.6 −11% 39.4 46.6 18% 47.8 56.5 18% 38.9 42.7 9% 42.8 50 14% 44.2 50.6 14% 49.3 72.5 47% 45.6 59.5 30% 55.1 63.9 16% 40.4 50.8 26% 47.6 76.2 60%