IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
1
2
Communication Information Structures and Contents
for Enhanced Safety of Highway Vehicle Platoons
3
4
Lijian Xu, Student Member, IEEE, Le Yi Wang, Fellow, IEEE, George G. Yin, Fellow, IEEE, and
Hongwei Zhang, Senior Member, IEEE
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Abstract—Highway platooning of vehicles has been identified
as a promising framework in developing intelligent transportation systems. By autonomous or semi-autonomous vehicle control
and intervehicle coordination, an appropriately managed platoon
can potentially offer enhanced safety, improved highway utility,
increased fuel economy, and reduced emissions. This paper is
focused on quantitative characterization of the impact of communication information structures and contents on platoon safety.
By comparing different information structures that combine front
sensors, rear sensors, and wireless communication channels, as
well as different information contents such as distances, speeds,
and drivers’ actions, we reveal a number of intrinsic relationships
between vehicle coordination and communications in platoons.
Typical communication standards and related communication latency are used as benchmark cases in this paper. The findings of
this paper provide useful guidelines in sensor selections, communication resource allocations, and vehicle coordination.
22
Index Terms—Autonomous vehicles, communication latency,
23 communication systems, highway platoons, vehicle safety.
I. I NTRODUCTION
24
H
IGHWAY platooning of vehicles has been identified as
a promising framework in developing intelligent transportation systems [1], [2]. By autonomous or semi-autonomous
vehicle control and intervehicle coordination, an appropriately
29 managed platoon can potentially offer enhanced safety, im30 proved highway utility, increased fuel economy, and reduced
31 emissions. In a platoon formation and maintenance, high32 level distributed supervisors adjust vehicle spatial distributions
33 based on intervehicle information such that roadway utilization
34 is maximized whereas the risk of collision is minimized or
35 avoided. Controllers at vehicle levels, sensors, and communi36 cation systems interact intimately in vehicle platoon formation
37 and control. This paper investigates several key issues in such
38 interactions.
25
26
27
28
Platoon control has drawn substantial attention lately [3], 39
[4]. During the 1990s, there were substantial contributions on 40
platoon control, including PATH projects [5], [6], and FleeNet, 41
among others. Intelligent platoon control algorithms were in- 42
troduced with demonstration and experimental validation [7], 43
[8]. The most common objectives in platoon control are safety, 44
string stability, and team coordination [9], [10]. Early studies 45
of platoon control were not communication focused, due to less 46
advanced communication systems at that time. In our recent 47
work [11], [12], a weighted and constrained consensus control 48
method has been introduced to achieve platoon formation and 49
robustness. Currently, onboard front radars are used in vehi- 50
cle distance measurements. In [12], convergence rates were 51
employed as a performance measure to evaluate the benefits 52
of different communication topologies in improving platoon 53
formation, robustness, and safety.
54
Communication channels insert new dynamic subsystems 55
into control loops. The impact of communication systems on 56
feedback loops can be treated as added uncertainty, such as 57
delays and errors [13]–[15]. In terms of coordination of con- 58
trol and communication systems in a platoon, some intrinsic 59
questions arise: 1) How much improvement of safety can be 60
achieved by including communication channels; 2) what infor- 61
mation should be communicated and what are the values of 62
such information; and 3) how will communication uncertain- 63
ties, such as latency, packet loss, and error, affect safety?
64
This paper aims to answer these questions with quantita- 65
tive characterization. To facilitate this exploration, we con- 66
sider various information structures: 1) front radars only and 67
2) combined radars and wireless communications. In addition, 68
we investigate the information contents: 1) distances only, 69
2) distance and speed, and 3) additional early warning of the 70
driver’s braking action. Typical communication standards, such 71
as IEEE 802.11p and related communication latency, are used 72
as benchmark cases in this paper. The findings of this paper will 73
be useful to guide design of information infrastructures, infor- 74
mation contents, control strategies, and resource allocations in 75
platoon control problems.
76
The rest of this paper is organized as follows. Section II 77
introduces the basic platoon control problem and safety issues. 78
Section III defines control strategies and sets up evaluation 79
scenarios for comparative studies of different information struc- 80
tures and contents. Our studies start with safety analysis in 81
Section IV. Under some simplified scenarios, basic relations are 82
derived, including speed–distance relationship for safe stopping 83
distance and collision avoidance, distance progression in a pla- 84
toon, and delay–distance functions for communication latency. 85
IE
E
Pr E
oo
f
1
Manuscript received October 12, 2013; revised December 18, 2013; accepted
March 8, 2014. This paper was supported in part by the U.S. National Science
Foundation under Grant CPS-1136007. The review of this paper was coordinated by Dr. D. Cao.
L. Xu and L. Y. Wang are with the Department of Electrical and Computer Engineering, Wayne State University, Detroit, MI 48202 USA (e-mail:
[email protected]; [email protected]).
G. G. Yin is with the Department of Mathematics, Wayne State University,
Detroit, MI 48202 USA (e-mail: [email protected]).
H. Zhang is with the Department of Computer Science, Wayne State University, Detroit, MI 48202 USA (e-mail: [email protected]).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TVT.2014.2311384
0018-9545 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
2
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
Section V details typical communication scenarios. Communication latency characterization and related experimental data
are presented. Section VI investigates the impact of infor89 mation structure by comparing radar-based distance sensing
90 and communications. Front radars are the current commercial
91 automotive technology. By expanding information structures
92 to include wireless communication networks, improvement
93 on safety is quantitatively studied. The roles of information
94 contents are explored in Section VII, in which improvements
95 on safety by including more information on vehicle speeds
96 and drivers’ actions are studied. Section VIII investigates the
97 impact of communication latency and uncertainties on vehicle
98 safety. Typical scenarios of communication latency, radar reso99 lution, and Doppler frequency shifting are considered. Finally,
100 Section IX discusses implications of the results of this paper
101 and points out some potential extensions.
86
87
88
102
103
Information structures.
II. V EHICLE DYNAMICS AND P LATOON
I NFORMATION S TRUCTURE
This paper is concerned with intervehicle distance control
105 in a highway platoon. For clarity of investigation, we use
106 simplified, generic, but representative vehicle dynamic models
107 from [16]
Fig. 2. Three main information structures. (a) Only front distance information
is available for vehicle control. (b) Both front and rear distances are available.
(c) Additional information is transmitted between vehicles.
IE
E
Pr E
oo
f
104
mv̇ + f (v) = F
(1)
where m (in kilograms) is the consolidated vehicle mass (including vehicle, passengers, etc.), v is the vehicle speed (in
meters per second), f (v) is a positive nonlinear function of
111 v representing resistance force from aerodynamic drag and
2
112 tire/road rolling frictions, and F (in Newtons or kg · m/s ) is
113 the net driving force (if F > 0) or braking force (if F < 0)
114 on the vehicle’s gravitational center. Typically, f (v) takes a
115 generic form f (v) = av + bv 2 , where the coefficient a > 0 is
116 the tire/road rolling resistance, and b > 0 is the aerodynamic
117 drag coefficient. These parameters depend on many factors such
118 as the vehicle weight, exterior profile, tire types and aging,
119 road conditions, and wind strength and directions. Conse120 quently, they are determined experimentally and approximately.
121 This paper focuses on longitude vehicle movements within a
122 straight-line lane. Thus, the vehicle movement is simplified into
123 a 1-D system.
124
Vehicles receive platoon movement information by using
125 sensors and communication systems. We assume that radars
126 are either installed at front or rear of the vehicle. The raw data
127 from the radars are distance information between two vehicles.
128 Although it is theoretically possible to derive speed information
129 by signal processing (derivatives of the distances), this paper
130 works with the direct information and leaves signal processing
131 as part of control design. As a result, radar information is
132 limited to distances. In contrast, a communication channel from
133 vehicle i to vehicle j can transmit any information that vehicle
134 i possesses. We consider the following information contents
135 for transmission: 1) vehicle i’s distance that is measured by
136 its front sensor; 2) vehicle i’s speed, which is available by
137 its own speedometer; and 3) vehicle i’s braking action. Infor138 mation structures are shown in Fig. 1. A vehicle may receive
139 information from its front distance sensor (on its distance to
108
109
110
Fig. 1.
Fig. 3.
Grouping vehicles.
the front vehicle) or its rear sensor (on its distance to the ve- 140
hicle behind it), or wireless communication channels between 141
two vehicles. The wireless communication channels may carry 142
different information contents such as distance, speed, driver’s 143
action, etc.
144
For concreteness, we use a basic three-car platoon to present 145
our key results. Although this is a highly simplified platoon, 146
the main issues are revealed clearly in this system. Three 147
information structures are studied, shown in Fig. 2. “Informa- 148
tion Structure (a)” employs only front sensors, implying that 149
vehicle 1 follows vehicle 0 by measuring its front distance d1 , 150
and then vehicle 2 follows vehicle 1 by measuring its front 151
distance d2 . For safety consideration, this structure provides a 152
baseline safety metric for comparison with other information 153
structures. “Information Structure (b)” provides both front and 154
rear distances. Then, “Information Structure (c)” expands with 155
wireless communication networks.
156
Although we employ a three-car platoon for simplicity, it 157
forms a generic base for studying platoon safety issues for more 158
general platoons. This is graphically explained in Fig. 3. Here, 159
the vehicles in between the leading vehicle and the vehicle of 160
interest are grouped as one subplatoon. We treat this subplatoon 161
as one vehicle, and this leads to the generic structure of Fig. 2. 162
This also implies that the communication distance between the 163
two vehicles may be high.
164
XU et al.: ENHANCED SAFETY OF HIGHWAY VEHICLE PLATOONS
3
Fig. 5. Braking functions based on speed information.
Fig. 4.
165
Braking functions based on distance information.
The platoon in Fig. 2 has the following local dynamics:
⎧
v̇0 = m10 F0 − a0 + b0 v02
⎪
⎪
⎪
⎪
⎨ v̇1 = m11 F1 − a1 + b1 v12 (2)
v̇2 = m12 F2 − a2 + b2 v22
⎪
⎪
⎪
˙
⎪
⎩ d1 = v 0 − v 1
d˙2 = v1 − v2
170
171
IE
E
Pr E
oo
f
where F0 is the leading vehicle’s driving action. F1 and F2 are
local control variables. Since the vehicle lengths are fixed and
can be subtracted from distance calculations, in this formula169 tion, a vehicle is considered a point mass without length.
166
167
168
III. C ONTROL AND E VALUATION S CENARIOS
A. Feedback Control
For safety consideration, the intervehicle distances d1 and
d2 have a minimum distance dmin > 0. To ensure that
vehicles 1 and 2 have sufficient distances to stop when the
175 leading vehicle 0 brakes, a cruising distance dref is imposed.
176 Apparently, the larger dref , the safer the platoon, under any
177 fixed control strategy. However, a larger dref implies more
178 occupation of the highway space and less efficiency in highway
179 usage. As a result, it is desirable to use as small dref as possible
180 without compromising the safety constraint.
181
There are numerous vehicle control laws, which have been
182 proposed or commercially implemented [17], [18]. Since the
183 focus of this paper is on impact of information structures and
184 contents rather than control laws, we impose certain simple and
185 fixed control laws. For safety consideration, we concentrate on
186 the case when the distance is below the nominal value d < dref .
187 The control law involves a normal braking region (small slope)
188 and an enhanced braking region of a sharp nonlinear function
189 toward the maximum braking force, as shown in Fig. 4. We
190 denote this function as F = g1 (d).
191
Similarly, if vehicle i’s speed information is transmitted to
192 another vehicle j (behind i), the receiving vehicle can use this
193 information to control its braking force. This happens when
194 vj > vi . The larger the difference, the stronger the braking
195 force. This control strategy may be represented by function
196 F = g2 (vj − vi ), which is shown in Fig. 5.
172
173
174
197
B. Evaluation Scenarios
To investigate the impact of information structures and con199 tents on platoon safety, we need a reasonable platform to
198
Fig. 6. Braking function for Example 4.
comparative studies. Since vehicle safety involves so many 200
factors, we must define a highly simplified platform in which 201
only key elements are represented. For this reason, we define 202
the following basic scenarios.
203
We use some typical vehicle data from [16]. Under the meter, 204
kilogram, and second systems of units, vehicle mass m has the 205
range of 1400–1800 kg, and aerodynamic drag coefficient b has 206
the range of 0.35–0.6 kg/m. During braking, a (as the rolling 207
resistance) is changed to tire/road slipping, which is translated 208
into the braking force F (negative value in Newton). As a result, 209
a is omitted.
210
Three identical cars form a platoon as in Fig. 2. The vehicle 211
masses are m0 = m1 = m2 = m = 1500 kg. The aerodynamic 212
drag coefficients b0 = b1 = b2 = 0.43. The nominal intervehi- 213
cle distance dref = 40 m. The cruising platoon speed is 25 m/s 214
(about 56 mi/h). The road condition is dry, and the maximum 215
braking force is 10 000 N. This implies that, when the maxi- 216
mum braking is applied (100% slip), the vehicle will come to 217
a stop in 3.75 s. The braking resistance can be controlled by 218
applying controllable forces on the brake pads.
219
The feedback control function F = g1 (d) is shown in Fig. 6. 220
The actual function is
221
(3)
max k1 (d − dref ) + k2 (d − dref )3 , −Fmax
where dref = 40 m, k1 = 50, k2 = 4, and Fmax = 10 000 (N). 222
The function applies smaller braking force when the distance 223
is only slightly below the reference value, but it increases 224
4
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
inversely proportional to the radius of the loaded tire. The 262
rolling resistance of a normal car weighing 1500 kg on concrete 263
with a rolling coefficient of 0.01 can be estimated as follows: 264
Fr= 0.01(1500 kg)(g)= 0.03(1500 kg)(9.81 m/s2 ) = 147 (N).
(4)
When b = 0.43 and v = 25 m/s, the aerodynamic drag force 265
is 268.75 N. This is only 8.3% of the braking force. In the 266
subsequent development, we omit the aerodynamic drag force 267
in our derivations but include it in all simulation studies.
268
Assuming that the platoon cruising speed is v0 (0) = v1 (0) = 269
v2 (0) = 25 m/s and the leading vehicle brakes at t = 0 with 270
F0 = −α, where α is a constant (for the fast braking, α = 271
5000 N, and for the worst case, α = Fmax = 10000 N). Brak- 272
ing function (3) is used. It follows that the dynamics of the 273
three-car platoon are as follows:
274
Fig. 7. Distance trajectories under slow braking.
the braking force more dramatically in a nonlinear function
when the distance reduces further until it reaches the maximum
braking force. We comment that, if one views the braking
function purely from safety aspects, it is desirable to impose
229 the maximum braking as soon as the distance drops. This, how230 ever, will compromise drivability and smoothness of platoon
231 operation. In fact, the braking function of Fig. 6 is already on
232 the aggressive side.
233
To see this, consider the slow braking condition: Suppose
234 that the leading vehicle applies a braking force of 1000 N,
235 which brings it to a stop from 25 m/s in 37.5 s. The distance
236 trajectories of d1 and d2 are shown in Fig. 7. In this case, the
237 minimum distances are 30.9 m for d1 and 24.2 m for d2 . This
238 is acceptable for safety. On the other hand, the transient period
239 shows oscillation, indicating that the braking action has been
240 aggressive already under normal driving conditions.
241
For evaluations, we will use the fast braking scenario defined
242 as follows.
243
Fast Braking: The leading vehicle uses a braking force of
244 5000 N. If the cruising speed of the platoon is 25 m/s, then this
245 braking force brings the leading vehicle to a stop from 25 m/s
246 in 7.5 s.
247
In some derivations, we also use the extreme case in which
248 the maximum braking force of 10 000 N is applied. This is
249 for the worst-case analysis. However, the fast braking case is
250 representative for understanding safety issues. In this paper, the
251 minimum vehicle distance dmin = 15 m is used to distinguish
252 “acceptable” and “unsafe” conditions. When a distance is re253 duced to 0, a collision occurs.
⎧
α
v̇0 = − m
⎪
⎪
g
(d )
⎪
⎪
⎨ v̇1 = − 1m 1
(d2 )
v̇2 = − g1m
⎪
⎪
⎪
⎪ d˙1 = v0 − v1
⎩
d˙2 = v1 − v2
226
227
228
254
IE
E
Pr E
oo
f
225
IV. S AFETY A NALYSIS
We conduct safety analysis under the scenario specified in
Section III-B. Some simplifications will be made so that explicit
expressions can be derived to clarify the main underlying safety
issues.
259
We observe that, under this braking force, the influence
260 of the tire/road resistance and aerodynamic drag force bv 2 is
261 relatively small. a is proportional to the tire deformation and
255
256
257
258
(5)
with the initial conditions v0 (0) = v1 (0) = v2 (0) = 25 m/s and 275
d1 (0) = d2 (0) = dref = 40 m.
276
A. Safety Regions
277
In a platoon, usually, vehicle 2 acts later than vehicle 1 due
to information cascading structures (vehicle 1 sees the slowing
down of the leading vehicle before vehicle 2). Suppose that,
after vehicle 1 applied the maximum braking force at an earlier
time, vehicle 2 starts to apply the maximum braking force at t0 .
Theorem 1: Assume that v1 (t0 ) < v2 (t0 ). Denote η =
v22 (t0 ) − v12 (t0 ) and δ = d2 (t0 ). The final distance is
278
dfinal
=δ−
2
279
280
281
282
283
284
ηm
.
2Fmax
Proof: For t ≥ t0 , the two vehicles have the dynamics
v̇1 = −Fmax /m and v̇2 = −(Fmax /m), which implies v1 (t) =
v1 (t0 ) − (Fmax /m)(t − t0 ) and v2 (t) = v2 (t0 ) − (Fmax /m)
(t − t0 ).
Vehicle 1 stops after traveling the total stopping time
v1 (0)m/Fmax and the total length Δ1 = v12 (t0 )m/(2Fmax ).
Similarly, the total length traveled by vehicle 2 to a complete
stop is Δ2 = v22 (t0 )m/(2Fmax ). Thus, the final distance is
285
286
287
288
289
290
291
292
dfinal
=δ−
2
v22 (t0 ) − v21 (t0 ) m
ηm
=δ−
.
2Fmax
2Fmax
For any given final distance dfinal
=
C,
the
function
η
=
2
(2Fmax /m)(δ − C) defines the iso-final-distance line on the
δ − η space, shown in Fig. 8, in which the acceptable region
and collision avoidance region are also marked.
293
294
295
296
297
XU et al.: ENHANCED SAFETY OF HIGHWAY VEHICLE PLATOONS
5
Now, the final distance dfinal
is dfinal
= dref − (Lj − Lj−1 ), 324
j
j
which implies that
325
min dfinal
= dref − max (Lj − Lj−1 ).
j
j=1,...,n
j=1,...,n
326
This completes the proof.
C. Delay–Distance Relationship
Fig. 8. δ − ηv lines under a given final distance. Acceptable safety regions
and collision avoidance regions can be derived from such curves.
298
B. Platoon Distance Progression
A platoon consists of many vehicles. Under typical information structures, there is intervehicle distance progression that
301 must be considered in platoon management.
302
Assumption 1: 1) At t = 0, the platoon of n following
303 vehicles is at the cruising condition with equal distance dref
304 and speed v(0). 2) The information on the braking action
305 F0 = −Fmax of the leading vehicle at t = 0 is passed to
306 the following vehicles in a progressive manner: For t > 0,
307 F1 (t) < F2 (t) < · · · < Fn (t), except when the braking forces
308 reach the saturation values −10 000 N , in which case, one
309 may have F1 (t) = F2 (t), etc. 3) Suppose that vehicle j starts
310 to apply the maximum braking force at tj . We assume that
311 t1 < t2 < · · · < tn .
312
Theorem 2: Under Assumption 1, the total travel length Lj
313 of vehicle j before a complete stop satisfies
L0 =
314
v(0)m
< L1 < L2 < · · · Ln .
2Fmax
dfinal
= dref − v(0) τ +
1
j=1,...,n
Proof: Since the braking force for the leading vehicle is 349
−Fmax , its speed profile is
350
j=1,...,n
Proof: The expression L0 = (v(0)m)/(2Fmax ) is proven
316 in Theorem 1.
317
Let the braking force for vehicle j be −fj (t) with fj > 0.
t
318 The speed profile is vj (t) = v(0) − 0 (fj (τ )/m)dτ . The total
Tj
319 travel time Tj satisfies 0 (fj (τ )/m)dτ = v(0). The total
320 length traveled by vehicle j until a complete stop is
315
Tj
Tj t
vj (t) dt = v(0)Tj −
Lj =
0
fj (τ ) dτ dt.
0
0
Under Assumption 1, we have the inequalities
v1 (t) < v2 (t) < · · · < vn (t),
322
t>0
(6)
which implies that
T1 < T2 < · · · Tn .
323
Fmax 2
τ .
2m
The minimum final distance is
min dfinal
= dref − max (Lj − Lj−1 ).
j
321
This paper concentrates on communication latency and its 328
impact on vehicle safety. Here, a relationship between the com- 329
munication delay time and its detrimental effect on intervehicle 330
distance is derived. To single out the delay effect, we impose 331
the following assumption.
332
1) Direct Transmission of Braking Action: Suppose that the 333
leading vehicle transmits its braking action directly to the 334
vehicle behind it. This is the fastest way to inform the following 335
vehicle to take action. If no time delay is involved, then the 336
following vehicle will brake immediately, and the intervehicle 337
distance will be kept contact until both vehicles come to the 338
complete stop. However, communication delays will postpone 339
the following vehicle’s action. The main question follows: How 340
much delay can be tolerated?
341
Assumption 2: 1) The leading vehicle and the following 342
vehicle travel at the cruising condition with distance dref and 343
speed v(0). 2) The information on the braking action F0 = 344
−Fmax of the leading vehicle at t = 0 is immediately trans- 345
mitted to vehicle 1 with a communication delay τ . 3) No other 346
information is available to vehicle 1.
347
Theorem 3: Under Assumption 2, the final distance dfinal
is 348
1
IE
E
Pr E
oo
f
299
300
327
These imply L1 < L2 < · · · Ln .
(7)
v0 (t) = v(0) −
Fmax
t.
m
At time τ , its speed is v0 (τ ) = v(0) − (Fmax /m)τ . Vehicle 1 351
receives the braking information at τ and immediately applies 352
the maximum braking force −Fmax with the initial speed v(0). 353
As a result, η = v 2 (0) − v02 (τ ).
354
By Theorem 1, the final distance is
355
dfinal
= dref − ηm/(2Fmax )
1
= dref − v 2 (0) − v02 (τ ) m/(2Fmax )
2 Fmax
2
τ
m/(2Fmax )
= dref − v (0) − v(0) −
m
2
Fmax
Fmax
2
τ−
τ m/(2Fmax )
= dref − 2v(0)
m
m2
= dref − v(0)τ +
Fmax 2
τ .
2m
356
6
357
358
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
Corollary 1: For a given required minimum distance dmin ,
the maximum tolerable communication delay is
v(0) − v 2 (0) − 2 Fmax
m (dref − dmin )
.
τmax =
2
Proof: By Theorem 3, to satisfy dfinal
≥ dmin , the max1
360 imum tolerable τ is solved from dmin = dref − v(0)τ +
361 (Fmax /2m)τ 2 or
359
Fmax 2
τ − v(0)τ + (dref − dmin ) = 0
2m
363
364
whose smaller solution is
v(0) − v 2 (0) − 2 Fmax
m (dref − dmin )
.
τmax =
2
In particular, for collision avoidance, dmin = 0, and we have
v(0) − v 2 (0) − 2 Fmax
m dref
.
τmax =
2
For the evaluation scenario in Section III-B, v(0) = 25,
Fmax = 10 000, m = 1500, dref = 40, and dmin = 15. The
corresponding maximum tolerable delay is τmax = 3.9609 s.
For collision avoidance, dmin = 0 and τmax = 7.7129 s.
369
However, if the vehicle weight is increased to m = 1800 kg
370 and the platoon cruising distance is reduced to dref = 30, the
371 tolerable delay is reduced to τmax = 1.7956 s.
372
Typical vehicle braking control must balance safety and
373 drivability. Consequently, intervehicle distances may reduce
374 more significantly than the scenario here. As a result, the
375 maximum tolerable delay may be significantly less. These will
376 be evaluated in the subsequent case studies.
377
2) Broadcasting Schemes and Consequence: The leading
378 vehicle’s braking action can be broadcast to the platoon. The
379 average communication latency depends on the distance be380 tween the sending (leading vehicle) node and the receiving
381 node. Using the basic square relationship, if the first following
382 vehicle experiences a delay τ1 = τ , then the second vehicle will
383 have a delay around τ2 = 4τ , the third vehicle with τ3 = 9τ ,
384 and so on.
385
For example, if dref = 40 m and τ1 = 100 ms, then τ2 =
386 400 ms (at 80 m), . . . , τ7 = 4.9 s (at 280 m), which implies
387 that d7 will fall below 15 m, violating the minimum distance
388 requirement.
389
This analysis indicates that communication schemes need to
390 be carefully designed when a platoon has many vehicles.
365
366
367
368
391
392
V. C OMMUNICATION S YSTEMS
A. Communication Standards and Latency
To study more realistically how communication systems
and control interact, we use a generic communication scheme
shown in Fig. 9. In this scheme, a data packet is generated and
396 enters the queue for transmission. The queuing time depends
393
394
395
Data transmission schemes.
on network traffic and data priorities. The packet contains both 397
data bits and error checking bits. We assume that the error 398
checking mechanism is sufficient to detect any faulty packet. 399
If the packet transmission is successful, the receiver returns 400
an acknowledgment message to the sender, which completes 401
the transmission. If the packet is received with error, it will be 402
discarded, and a request is sent back to the sender to retransmit 403
the same packet. The permitted total time for transmission of 404
a packet is predetermined by the control updating times. If 405
a packet was not successfully transmitted when the control 406
updating time is up, the packet will be considered lost.
407
Intervehicle communications (IVC) can be realized by using 408
infrared, radio, or microwaves waves. For instance, in IEEE 409
802.11p, a bandwidth of 75 MHz is allotted in the 5.9-GHz 410
band for dedicated short-range communication (DSRC) [19], 411
[20]. Alternatively, ultrawideband (UWB) technologies have 412
been used for IVC. IEEE 802.11x, where x ∈ {a, b, g, p, . . .}, 413
have been studied for intervehicle use. Currently, many appli- 414
cations use DSRC with IEEE 802.11p (a modified version of 415
IEEE 802.11 (Wi-Fi) standard) at the physical and medium 416
access control (MAC) layers. IEEE 802.11g and IEEE 802.11p 417
are used for experimental studies in this paper.
418
In the middle of protocol stack, DSRC employs IEEE 1609.4 419
for channel switching, 1609.3 for network service, and 1609.2 420
for security service. In the network service, users have a choice 421
between the wireless access for vehicle environments short 422
message protocol (WSMP) or the Internet protocol version 6 423
(IPv6) and user datagram protocol (UDP)/transmission con- 424
trol protocol (TCP). Single-hop messages typically use the 425
bandwidth-efficient WSMP, whereas multihop packets use the 426
IPv6 + UPD/TCP for its routing capability.
427
IVC uses wireless networks that are subject to severe un- 428
certainties. For example, the signal-to-interference-plus-noise 429
ratio (SINR) [21] attenuates with distance (it decreases inverse 430
proportionally to the cubic of the distance between the two 431
vehicles). It is also affected by obstructions such as buildings, 432
bridges, other vehicles, etc. Other factors include queue delays, 433
network data traffic conditions, routes, signal fading, signal 434
interference from other vehicles, Doppler shifts, and traffic and 435
weather conditions. These uncertainties depend significantly on 436
channel coding schemes and communication networks. These 437
factors collectively determine packet delivery delays, packet 438
loss rates, etc. This paper will focus on delay effects. To be 439
concrete in treating communication systems, we will employ 440
IEEE 802.11 standards as our benchmark systems and the 441
related latency data [19].
442
IE
E
Pr E
oo
f
362
Fig. 9.
XU et al.: ENHANCED SAFETY OF HIGHWAY VEHICLE PLATOONS
7
Fig. 10. Round trip delay.
Bandwidth–delay product is often used to characterize the
ability of a network pathway in carrying data flows [22], [23].
445 When the TCP is used in data communications, packet-carrying
446 capacity of a path between two vehicles will be limited by
447 this product’s upper bound (for more detailed discussions on
448 capacity/delay tradeoffs, see [19] and the references therein).
449 Note that latency is further caused by delays in each hub’s
450 queues, routes (multihub), packet delivery round-trip time
451 (RTT), channel reliability, retransmission, and scheduling poli452 cies in interference avoidance strategies. Although typical
453 transmission delays can be as low as several milliseconds, ve454 hicular traffic scenarios introduce combined latency of several
455 hundreds of milliseconds even several seconds. In this paper,
456 we will show that delays of such scales will have a significant
457 impact on vehicle safety.
443
444
IE
E
Pr E
oo
f
458
B. Single-Hop Experimental Study
We assume the three-vehicle scenario in Fig. 2. Communication channels between v0 and v2 use the WSMP. This protocol
can carry messages on both the control channel (CCH) and the
462 service channel (SCH). The WSMP allows direct control of the
463 lower layer parameters, such as transmission power, data rates,
464 channel numbers, and receiver MAC addresses. The WSMP
465 over the CCH can skip the steps of forming a WAVE basic
466 service set that delivers IP and WAVE short message data on
467 the SCH. Those methods can potentially reduce communication
468 latency.
469
The RTT under this protocol includes measurement time for
470 the variables (vehicle distance, speed, etc.), source data creation
471 time (creating packets, adding verification codes, scheduling,
472 etc), communicating the packet to v2 , receiver verification, and
473 travel time for sending back acknowledgment from v2 . Fig. 10
474 shows some of the time delays from these steps.
475
In an ideal case that v0 can capture the CCH during each
476 CCH time slot, v0 can send its beacon and update its status to v2
477 at the rate of 10 Hz. If a package is successfully transmitted and
478 verified during the first round, the package delivery rate (PDR)
479 is 1, and the RTT τ 0 ≤ 100 ms since IEEE 1609.4 specifies the
480 reoccurrence of the CCH at the rate of every 100 ms.
481
The physical limitations on wireless channels (bandwidth
482 and power constraints, multipath fading, noise and interfer483 ence) present a fundamental technical challenge to reliable
484 high-speed communications. One or several retransmissions
485 are often necessary to meet a PDR requirement. In this case,
486 delay is τ = nτ 0 , where n is the number of average rounds
487 for a successful transmission. In the following examples, we
459
460
461
Fig. 11. PDR versus separation distance under different data rates in the rural
road environment (with 95% confidence interval). Here, the data rates are 6 and
18 Mb/s. The transmission power is 20 dBm.
show how modulation rates and channel interference affect the 488
number of retransmission and delay τ . Due to the network 489
system heterogeneity and highway environments, we are using 490
the ground-truth data, rather than ns-3 simulations.
491
Example 1: In [25], experimental data of IEEE 802.11p 492
DSRC from a team of vehicles driving on certain Michigan 493
highways are reported. PDRs are measured under different 494
driving conditions, traffics, and surroundings. A typical curve 495
from [25] is regenerated in Fig. 11. When the modulation rate 496
is 6 Mb/s, the PDR is about 75% at a distance of 85 m. The first 497
round trip takes about 100 ms. Each subsequent round trip must 498
catch the next CCH, and it takes on average more than three 499
retransmissions to achieve a PDR over 98.5%. Consequently, 500
the average delay is τ ≈ 0.3 s. When the modulation rate is 501
increased to 18 Mb/s, the PDR is reduced to 36% at 85 m. To 502
meet the same PDR of 98.5%, the delay is more than 1 s.
503
Example 2: In this experimental study, we use the IEEE 504
802.11g standard to analyze the affects of multipath interfer- 505
ence. The communicating nodes reside on laptop computers 506
and are moved from a short distance of 20–95 m. In the first 507
experimental setting, the transmission pathway does not have 508
obvious obstacles, except low grass on the open field (OF). 509
Communication latency is recorded by the synchronized clocks 510
on these computers. Fig. 12 provides the experiment data on 511
recorded latency for different internode distances. A simplified 512
curve can be obtained by data fitting, which is also shown in the 513
same figure. It is noted that latency between 100 and 600 ms is 514
typical in this case study.
515
Example 3: Extending on the experiment in Example 2, we 516
now evaluate impact of obstacles on transmission pathways. 517
Under the same experimental protocols as in Example 2, we 518
select a field with many trees but one that is not overly dense. 519
Consequently, depending on distances, the transmission path- 520
ways are obstructed by several trees. Fig. 13 demonstrates the 521
experimental data on communication latency under different 522
transmission distances. It is seen clearly that, with obstacles, 523
communication latency increases significantly to a range of 3.4 s. 524
8
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
Fig. 12. Dependence of latency on distance without obstacles on the transmission pathway.
Fig. 14. Average delay of high-priority message dissemination for five hops
of communication as functions of the transmission range.
IE
E
Pr E
oo
f
VI. P LATOON I NFORMATION S TRUCTURE
A. Safety Under Front Sensor Information
539
540
We start with the basic information structure of using front 541
distance sensors only. For the three-car platoon in Fig. 2 and 542
the control law F = g1 (d) in Fig. 4, the closed-loop system 543
becomes
544
⎧
v̇0 = m10 F0 − a0 v0 + b0 v02
⎪
⎪
⎪
⎪
⎨ v̇1 = m11 g1 (d1 ) − a1 v1 + b1 v12 (8)
v̇2 = m12 g1 (d2 ) − a2 v2 + b2 v22
⎪
⎪
⎪
˙
⎪
⎩ d1 = v 0 − v 1
d˙2 = v1 − v2 .
Fig. 13. Dependence of latency on distance with trees on the transmission
pathway.
525
C. Multihop Communication Data
IVC may involve multihops, which create further delays.
Typically, the IPv6 + UDP/TCP can be used in such systems. Unlike the WSMPs, which use 11-B overhead, the IPv6
protocol requires a minimum overhead of 52 B. Although
530 this is more complicated in coding and less efficient in using
531 the data resource, this protocol provides more flexible routing
532 schemes. There are many experimental studies of IEEE 802.11p
533 under multihop and highway environment. Since we are only
534 concerned with latency data, here, we quote the studies in [19]
535 that contain extensive experimental results. A typical curve
536 from [19] is regenerated in Fig. 14. It is noted that, although
537 IEEE 802.11p uses higher power and faster speed, a latency of
538 hundreds of milliseconds is typical under highway conditions.
526
527
528
529
Example 4: We consider the scenario defined in Section III-B. 545
Suppose that the platoon uses only front sensors to measure 546
intervehicle distances, namely, the information structure in 547
Fig. 2(a) is in effect. The feedback control function F = g1 (d) 548
is shown in Fig. 6. We will use the following fast braking 549
condition for comparison.
550
Under the fast braking scenario in Section III-B, suppose that 551
the leading vehicle uses a braking force of 5000 N, which brings 552
it to a stop from 25 m/s in 7.5 s. The distance trajectories of 553
d1 and d2 are shown in Fig. 15. In this case, the minimum 554
distances are 20.6 m for d1 , which is acceptable, but 0 m for 555
d2 . This means that vehicle 2 will collide with vehicle 1 during 556
the transient time.
557
To explain this scenario, we note in the top plot of Fig. 15 558
that, since vehicle 2 relies on d2 to exercise its braking control 559
function, there is a dynamic delay in initiating its braking. d2 is 560
reduced to about 20 m when vehicle 2 starts to act. For a large 561
platoon, this dynamic delay from vehicle to vehicle is a serious 562
safety concern.
563
B. Adding Distance Information by Communications
564
We next expand on the information structures beyond front 565
sensors by adding distance information by communications. 566
XU et al.: ENHANCED SAFETY OF HIGHWAY VEHICLE PLATOONS
Fig. 17. Distance trajectories when the distance information d1 is made
available to vehicle 2. It shows improvement over Fig. 15.
IE
E
Pr E
oo
f
Fig. 15. Distance trajectories under fast braking.
9
Fig. 16. Enhanced information structure by sending d1 to vehicle 2 by
communication links in Example 5.
Example 5: Continuing from Example 4, we consider the
same three-car platoon under the same initial conditions: The
nominal intervehicle distances are 40 m, the cruising vehicle
570 speeds are 25 m/s, and the maximum braking force is 10 000 N.
571
Under the fast braking scenario as in Example 4, suppose
572 now that vehicle 1 sends d1 information to vehicle 2 by com573 munication. As a result, vehicle 2 can use both d1 and d2 in its
574 control function (see Fig. 16).
575
Suppose that vehicle 2 modifies its braking control function
576 from the previous F2 = g1 (d2 ) to the weighted sum F2 =
577 0.5g1 (d2 ) + 0.5g1 (d1 ) that uses both distances. The resulting
578 speed and distance trajectories are shown in Fig. 17. Now, the
579 minimum distances are 20.6 m for d1 and 15.9 m for d2 , both
580 are within the safety region.
581
To compare Figs. 15 and 17, we note that with information
582 feeding of d1 into vehicle 2, vehicle 2 can slow down when d1
583 is reducing before d2 changes. Consequently, it is able to act
584 earlier, resulting in a reduced distance swing for d2 during the
585 transient.
567
568
569
586
587
VII. P LATOON I NFORMATION C ONTENTS
A. Adding Speed Information by Communications
We now add the speed information of the leading vehicle to
both vehicles 1 and 2 by communications.
Example 6: For the same three-car platoon under the same
initial conditions as Example 5, we add the leading vehicle’s
592 speed v0 into the information structure. This information is
593 transmitted (or broadcast) to both vehicles 1 and 2. Under the
594 fast braking scenario like in Example 5, suppose that vehicles 1
588
589
590
591
Fig. 18. Enhanced information structure by sending d1 to vehicle 2 and v0 to
both vehicles 1 and 2.
and 2 receive the additional speed information v0 , resulting in 595
a new information structure shown in Fig. 18.
596
From the control functions of Example 5, additional control 597
actions g2 (v0 , v1 ) and g2 (v0 , v2 ) are inserted. The resulting 598
speed and distance trajectories are shown in Fig. 19. Now, the 599
minimum distances are 28.3 m for d1 and 27.1 m for d2 , which 600
is a much improved safety performance.
601
B. Adding Braking Event Information by Communications
602
Intuitively, if the leading vehicle’s braking action can be 603
also communicated, the following vehicles can act much ear- 604
lier than their measurement data on vehicle movements. To 605
evaluate benefits of sending the driver’s action, we add the 606
braking event information of the leading vehicle to vehicle 2 by 607
communications.
608
Example 7: For the same three-car platoon under the same 609
initial conditions as Example 6, we now further add the leading 610
vehicle’s braking event information F0 into the information 611
structure. To understand the impact, we purposely assume that 612
vehicle 1 does not receive this information. In other words, this 613
information will be transmitted only to vehicle 2 by communi- 614
cations. Under the fast braking scenario in Example 6, suppose 615
that vehicle 2 receives the additional braking event information 616
F0 , resulting in a new information structure shown in Fig. 20. 617
10
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
Fig. 21.
Distance trajectories with added braking event information.
IE
E
Pr E
oo
f
Fig. 19. Distance trajectories when both distance and speed information is
made available.
Fig. 20. Enhanced information structure by sending the braking event F0 to
vehicle 2.
From the control functions of Example 6, an alternative control action F0 is inserted when d2 < dref = 40 m. The resulting
620 speed and distance trajectories are shown in Fig. 21. Now,
621 the minimum distances are 28.3 m for d1 and 30.6 m for d2 ,
622 which is a much improved safety over the case in Example 6.
623 It is interesting to note that by knowing the leading vehicle’s
624 action, vehicle 2 can react faster than even vehicle 1, which
625 does not receive the braking action data.
618
619
626
627
628
VIII. I MPACT OF R ADAR AND
C OMMUNICATION U NCERTAINTIES
Distance trajectories under a radar of low resolution (1 m).
Fig. 23. Distribution of minimum distances d2 under a radar of low resolution (1 m).
A. Impact of Radar Resolution and Missed Detection
Radar sensors provide a stream of measurement data, typically using 24-, 35-, 76.5-, and 79-GHz radars. In general, radar
sensor measurements are influenced by many factors that limit
632 their accuracy and reliability. These include signal attenuation
633 by the medium, beam dispersion, noises, interference, multiob634 ject echo (clutter), jamming, etc.
635
We first consider the impact of radar’s resolution on a platoon
636 system. Within the same setup as Example 5, vehicle 2 receives
637 the distance information of d1 and d2 in which d2 is measured
629
630
631
Fig. 22.
by radar. Taking into consideration radar resolution, the mea- 638
sured distance is d˜2 = d2 + γδ, where γ is a resolution level, 639
and δ is a standard Gaussian noise N (0, 1).
640
Fig. 22 shows a simulation result under a radar with a reso- 641
lution of 1 m. The distribution of the minimum distances after 642
repeated runs to account for randomness is shown in Fig. 23. 643
Although the expectation is 8.01 m, the minimum distance has 644
a high probability of having values close to zero. Consequently, 645
this low-resolution radar is not suitable for this application.
646
XU et al.: ENHANCED SAFETY OF HIGHWAY VEHICLE PLATOONS
11
Fig. 24. Distance trajectories with high-resolution radar.
Fig. 26. Distance trajectories when communication delays are considered.
IE
E
Pr E
oo
f
TABLE I
I MPACT OF C OMMUNICATION D ELAYS
Fig. 25. Distance distribution of d2 with high-resolution radar.
Next, we upgrade the radar to a higher resolution of 0.1 m.
A corresponding simulation is shown in Fig. 24. The minimum
649 distances for both d1 and d2 are much improved. The distri650 bution of minimum distances of d2 is shown in Fig. 25. The
651 random minimum distances have expectation of 15.92 m and
652 variance σ 2 = 0.31. This is an acceptable resolution for this
653 application.
654
It is noted further that uncertainties of radar signals include
655 also random false alarms or missed detection. In this scenario,
656 the sensor does not provide information at the sampling time,
657 and the control/brake action must rely on its previous measure658 ments and other available information from different resources.
659 This situation is similar to Example 12 when communica660 tion information is unavailable, which will be detailed in the
661 following.
647
648
662
B. Impact of Communication Delay
Communications introduce a variety of uncertainties. Most
common types are communication latency and packet loss.
665 These can be caused by many factors, as listed in Section I.
663
664
This paper focuses on communication latency. Depending on 666
environment and communication protocols, communication la- 667
tency can be near a constant, distance dependent, or random. 668
We cover these cases in the following.
669
1) Fixed Delays: We first consider fixed delays.
670
Example 8: Under the same system and operating condition 671
as Example 5, we assume that the communication channel for 672
the distance information has a delay of τ s. The impact of the 673
communication delay is shown in Fig. 26. Without the delay, the 674
minimum distance for d2 is 15.9 m. When a delay of τ = 0.6 s 675
is introduced, the minimum distance for d2 is reduced to 11 m. 676
Table I lists the relationship between the delay time and the 677
minimum distance for d2 .
678
Next, we use experimental delay data in our simulation 679
studies.
680
Example 9: Under the same system and operating condition 681
as Example 5, we assume that communication systems use 682
the single-hop scenario in Section V-B. Under a scenario of 683
latency τ = 0.1 s (CCH delay only), the minimum distance 684
for d2 is 15.1 m. It remains as an acceptable safe distance. 685
Many factors affect such delays. One essential consideration 686
is channel capacity. Shannon’s channel capacity claims that, 687
if the channel is too noisy, which reduces channel capacity, 688
information cannot be effectively transmitted. This is translated 689
into very large channel latency under a required PDR. In this 690
sense, impact analysis of channel latency is, in fact, a study 691
on communication resources. Here, we use platoon safety as a 692
performance criterion in this paper.
693
2) Distance-Dependent Delays: In vehicle platoon environ- 694
ment, communication latency depends directly on intervehicle 695
distances. These are reflected clearly in Figs. 11–13. It is ob- 696
served that, during platoon formation and braking, intervehicle 697
distances change substantially. Here, we consider delays as a 698
function of distance.
699
12
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
Fig. 28. Distance trajectories when communication pathways are obstructed,
as shown at Fig. 13.
Example 10: Under the same system and operating condi701 tion as Example 9, we now use more realistic experimental data
702 in Fig. 12 for latency, which is a function of distance. Based
703 on the relationship of distance and latency, the simulation in
704 Fig. 27 shows that the minimum distance for d2 is now 12.7 m.
705 Furthermore, if signal interference, obstructions, and fading are
706 considered, the latency is increased to these in Fig. 13. The
707 simulation results in a minimum distance for d2 as 5.6 m. This
708 is shown in Fig. 28, which causes safety concerns.
709
Example 11: Continuing the study of Example 9, we con710 sider the multihop scenario in Section V-C. In that scenario,
711 transmission from v0 to v2 is over five hops. Suppose that
712 each hop has the same priority and that each loses the CCH
713 once followed by one successful retransmission. Based on
714 the distances between the vehicles in the example, the total
715 communication delay τ > 1.5 s. The simulation shows that the
716 minimum distance for d2 approaches to 0, leading to a collision.
717
3) Random Delays: Typically, communication delays are
718 random variables with certain distributions. Depending on la700
Fig. 29. Distance trajectories under communication latency, which is
Gaussian distributed.
IE
E
Pr E
oo
f
Fig. 27. Distance trajectories when communication delays are dependent on
vehicle distances, whose function form is given in Fig. 12 for the “no obstacle”
scenario.
Fig. 30.
Distance distribution of d2 under random communication latency.
tency control mechanisms of transmission protocols, the la- 719
tency can have different distributions. We use the common 720
Gaussian distribution for our study here.
721
Example 12: Assume that communication latency is a ran- 722
dom variable, due to the random features of wireless transmis- 723
sions. In this example, we model τ as a random variable that 724
is Gaussian distributed with E(τ ) = 1.2 (second) and variance 725
σ 2 = 0.09. Continuing the study of Example 11, the simulation 726
in Fig. 29 shows that the minimum distance d2 approaches 727
5.09 m.
728
Simulation results of minimum distance distribution are 729
shown in Fig. 30. The variance of d2 is σ 2 = 0.142.
730
C. Impact of Doppler Frequency Shift and Signal Spreading
731
Mobility-induced Doppler spread is one of the main factors 732
that degrade the performance of orthogonal frequency-division 733
multiplexing (OFDM) schemes. It introduces intersymbol inter- 734
ference (ISI) and intercarrier interference (ICI) by destroying 735
the orthogonality between adjacent subcarriers.
736
XU et al.: ENHANCED SAFETY OF HIGHWAY VEHICLE PLATOONS
13
Fig. 32. System integration of a platoon with a VANET framework.
Fig. 31. Impact of relative velocities on the PDR (with the 95% confidence
interval). A bin of 20 packets is used to calculate PDR values as well as relative
velocities.
770
IE
E
Pr E
oo
f
In most cases, DSRC is adequate to restore both zero ISI
and zero ICI in highly mobile severe-fading vehicular environ739 ments, as discussed with great detail in [24]. In the physical
740 layer of IEEE 802.11p, the bandwidth of each DSRC channel
741 is 10 MHz, which entails less ISI and ICI than IEEE 802.11a,
742 which uses a channel bandwidth of 20 MHz. This brings
743 better wireless channel propagation with respect to multipath
744 delay spreads and Doppler effects caused by high mobility and
745 roadway environments. Moreover, DSRC expands the guard
746 band (GB) to 156 KHz and has 1.6-μs guard interval for OFDM
747 schemes. The GB between subcarriers can ensure that mobility748 induced Doppler spreads do not cause two adjacent subcarriers
749 to overlap.
750
On the other hand, with high operation frequency at 5.9 GHz,
751 IEEE 802.11p is subject to higher Doppler frequency shifts.
752 When vehicle speeds are extremely high (such as 250 km/h
753 on German highways), the issue of Doppler frequency shifts
754 become more pronounced. Currently, fast network topology
755 switching and complicated road environments are still chal756 lenges with respect of ISI and ICI and remain to be resolved
757 by new technologies.
758
Fig. 31, reproduced from [25], compares the impacts of
759 OF and rural freeway (RRF) on the PDR. The PDR remains
760 nearly unchanged in the OF environment when relative vehicle
761 velocities vary from 0 to 25 m/s. In contrast, the PDR drops
762 dramatically in the RRF environment. For example, when the
763 relative velocity is 12.5 m/s, the PDR of the communication
764 link in the RRF environment is reduced to one third of that with
765 the OF environment. This implies that, in the RRF environment,
766 much more communication resources are needed to ensure the
767 same level of safety. As a result, it is advisable that these
768 DSRC characteristics be incorporated into the platoon design
769 by vehicular ad hoc network (VANET) designers.
737
738
D. System Integration With VANET Framework
The generic platoon model of this paper is an important
772 component of a VANET framework, as shown in Fig. 32. In our
771
exploration, the actual communication routes are not specified. 773
Within a VANET, the links among vehicles can be realized by 774
vehicle-to-vehicle communications or vehicle-to-infrastructure 775
pathways involving access points, wireless towers, and other 776
infrastructures. Our model provides a fundamental framework 777
to study the impact of communications on vehicle safety and 778
can be specified to different communication configurations. The 779
findings of this paper can be used as guidelines in selecting 780
VANET parameters. For example, transmission power, mod- 781
ulation rate, and coding scheme can be selected so that they 782
meet the requirements of an acceptable minimum intervehicle 783
distance. In addition, a platoon can potentially enhance VANET 784
data access performance. By using vehicles as transmission 785
hubs, data can be replicated and relayed to more vehicles 786
in the group. This structure improves VANET resources in a 787
distributed manner and, if used properly, can improve overall 788
performance.
789
While this paper is focused on one platoon formation, 790
a platoon experiences many dynamic variations in real im- 791
plementations. These include lane change, vehicle departure 792
and addition, platoon reformation, etc. At the network level, 793
such changes amount to network topology variations. At the 794
communication/physical level, some uncertainties will be intro- 795
duced, such as echo among vehicles and road infrastructures. 796
A VANET can easily accommodate such topology changes by 797
using vehicle IDs and their links. Furthermore, by seamless 798
integration into a VANET, a platoon can have access to VANET 799
resources, including GPS, Internet, distributed live databases, 800
VANET-enabled applications, etc. Consequently, a platoon can 801
potentially utilize additional information in its safety consider- 802
ations via IVC and emission reduction via traffic information. 803
These topics are, however, beyond the scope of this paper (see 804
[11] and [12] for some related studies).
805
IX. D ISCUSSIONS AND C ONCLUSION
806
This paper has investigated the interaction between control
and communications, in the framework of highway platoon
safety. Information structure, information content, and information reliability have been taken into consideration in this paper.
807
808
809
810
14
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
It is well perceived that communication systems introduce
uncertainties that are of many types and values. To be concrete,
we have selected communication latency as a key uncertainty
814 in this paper.
815
The main results of this paper demonstrate that commu816 nications provide critical information that can enhance vehi817 cle safety effectively beyond distance sensors. In fact, from
818 our simulation studies, platoon control may mandate com819 munications for additional information. Although traditionally,
820 distance and vehicle speed are immediate candidates for trans821 mission, our results have that drivers’ braking events contain
822 very effective information for platoon management. Our sim823 ulations suggest that platoon communications place event data
824 under more prominent considerations.
825
This paper have shown that communication latency is a
826 critical factor in information exchange. Large latency can di827 minish values of data communication in platoon control. It is a
828 common framework in multivehicle communication scenarios
829 that vehicles within an interference radius do not transmit
830 simultaneously. A direct consequence is that latency becomes
831 larger. For instance, under the IEEE 802.11p standard, trans832 mission radius can reach 1 km. If 50 vehicles are in this region
833 and each transmission (or broadcasting) takes 30 ms, a delay of
834 1.5 s will occur between consecutive transmissions of a given
835 vehicle. This paper shows that such a delay has an alarmingly
836 high impact on vehicle safety. This issue deserves further
837 study.
838
What is reported in this paper is a first step in this di839 rection. There are many unresolved issues. We are currently
840 investigating the impact of communications on platoon safety
841 under packet erasure channels. Furthermore, we have only
842 considered basic driving conditions: Straight lanes, dry surface
843 conditions, good weather conditions, and no lane changes or
844 platoon reformation after vehicle departure or addition. System
845 integration with VANET framework is a worthy topic to pursue.
846 All these issues are worth further studies.
811
[9] G. Guo and W. Yue, “Autonomous platoon control allowing range-limited
sensors,” IEEE Trans. Veh. Technol., vol. 61, no. 7, pp. 2901–2912,
Sep. 2012.
[10] C. Y. Liang and H. Peng, “String stability analysis of adaptive cruise
controlled vehicles,” JSME Int. J. Series C, vol. 43, no. 3, pp. 671–677,
2000.
[11] L. Y. Wang, A. Syed, G. Yin, A. Pandya, and H. W. Zhang, “Coordinated
vehicle platoon control: Weighted and constrained consensus and communication network topologies,” in Proc. CDC, Maui, HI, USA, Dec. 2012,
pp. 4057–4062.
[12] L. Y. Wang, A. Syed, G. Yin, A. Pandya, and H. W. Zhang, “Control of
vehicle platoons for highway safety and efficient utility: Consensus with
communications and vehicle dynamics,” J. Syst. Sci. Complexity, to be
published.
[13] J. S. Freudenberg and R. H. Middleton, “Feedback control performance
over a noisy communication channel,” in Proc. Inf. Theory Workshop,
Porto, Portugal, May 2008, pp. 232–236.
[14] R. Luck and A. Ray, “Experimental verification of a delay compensation algorithm for integrated communication and control system,” Int. J.
Control, vol. 59, no. 6, pp. 1357–1372, 1994.
[15] A. J. Rojas, J. H. Braslavsky, and R. H. Middleton, “Fundamental limitations in control over a communication channel,” Automatica, vol. 44,
no. 12, pp. 3147–3151, Dec. 2008.
[16] D. H. McMahon, J. K. Hedrick, and S. E. Shladover, “Vehicle modelling
and control for automated highway systems,” in Proc. Amer. Control
Conf., San Diego, CA, USA, May 23–25, 1990, pp. 297–303.
[17] R. T. O’Brien, “Vehicle lateral control for automated highway systems,”
IEEE Trans. Control Syst. Technol., vol. 4, no. 3, pp. 266–273, May 1996.
[18] S. Sheikholeslam and C. A. Desoer, “Longitudinal control of a platoon of
vehicles with no communication of lead vehicle information: A system
level study,” IEEE Trans. Veh. Technol., vol. 42, no. 4, pp. 546–554,
Nov. 1993.
[19] M. J. Neely and E. Modiano, “Capacity and delay tradeoffs for Ad-Hoc
mobile networks,” IEEE Trans. Inf. Theory, vol. 51, no. 6, pp. 1917–1936,
Jun. 2005.
[20] K. A. Hafeez, L. Zhao, B. Ma, and J. W. Mark, “Performance analysis
and enhancement of the DSRC for VANET’s safety applications,” IEEE
Trans. Veh. Technol., vol. 62, no. 7, pp. 3069–3083, Sep. 2013.
[21] A. Kumar, D. Manjunath, and J. Kuri, Wireless Networking. Amsterdam,
The Netherlands: Elsevier, 2008.
[22] L. L. Peterson and B. S. Davie, Computer Networks, 2nd ed.
San Francisco, CA, USA: Morgan Kaufmann, 2000.
[23] W. Stevens, TCP/IP Illustrated. Reading, MA, USA: Addison-Wesley,
1994.
[24] L. Cheng, B. E. Henry, D. D. Stancil, F. Bai, and P. Mudalige, “Mobile
vehicle-to-vehicle narrow-band channel measurement and characterization of 5.9 GHz dedicated short range communication (DSRC) frequency
band,” IEEE J. Sel. Areas Commun., vol. 25, no. 8, pp. 1501–1516,
Oct. 2007.
[25] F. Bai, D. D. Stancil, and H. Krishnan, “Toward understanding characteristics of dedicated short range communications (DSRC) from a perspective of vehicular network engineers,” in Proc. MobiCom, Chicago, IL,
USA, Sep. 20–24, 2010, pp. 329–340.
873
874
875
876
877
878
879
880
881
882 AQ1
883
884
885 AQ2
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
Lijian Xu (S’12) received the B.S. degree from
the University of Science and Technology Beijing,
Beijing, China, in 1998 and the M.S. degree from
the University of Central Florida, Orlando, FL, USA,
in 2001. He is currently working toward the Ph.D.
degree with the Department of Electrical and Computer Engineering, Wayne State University, Detroit,
MI, USA.
From 2001 to 2007, he worked for AT&T, FL, and
Telus Communication, Canada, as an Engineer and
as an Engineering Manager. His research interests
include network control systems, digital wireless communications, consensus
control, and vehicle platoon and safety.
Dr. Xu received a Best Paper Award at the 2012 IEEE International Conference on Electro/Information Technology.
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
IE
E
Pr E
oo
f
812
813
847
R EFERENCES
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
[1] J. K. Hedrick, D. McMahon, and D. Swaroop, “Vehicle modeling
and control for automated highway systems,” Univ. Calif., Berkeley,
Berkeley, CA, USA, PATH Res. Rep. UCB-ITS-PRR-93-24, 1993.
[2] P. Ioannou and C. Chien, “Autonomous intelligent cruise control,” IEEE
Trans. Veh. Technol., vol. 42, no. 4, pp. 657–672, Nov. 1993.
[3] R. Rajamani, H. S. Tan, B. Law, and W. B. Zhang, “Demonstration of
integrated lateral and longitudinal control for the operation of automated
vehicles in platoons,” IEEE Trans. Control Syst. Technol., vol. 8, no. 4,
pp. 695–708, Jul. 2000.
[4] F. Knorr, D. Baselt, M. Schreckenberg, and M. Mauve, “Reducing
traffic jams via VANETs,” IEEE Trans. Veh. Technol., vol. 61, no. 8,
pp. 3490–3498, Oct. 2012.
[5] K. S. Chang, W. Li, P. Devlin, A. Shaikhbahai, P. Varaiya, J. K. Hedrick,
D. McMahon, V. Narendran, D. Swaroop, and J. Olds, “Experimentation
with a vehicle platoon control system,” in Proc. Veh. Navigat. Inf. Syst.
Conf., 1991, pp. 1117–1124.
[6] D. H. Narendran, V. K. Swaroop, D. Hedrick, J. K. Chang, and
K. S. Devlin, “Longitudinal vehicle controllers for IVHS: Theory
and Experiment McMahon,” in Proc. Amer. Control Conf., 1992,
pp. 1753–1757.
[7] Y. F. Zhao and H. Ogai, “Development of a platooning control algorithm
based on RoboCar,” in Proc. SICE, 2011, pp. 352–355.
[8] J. Bom, B. Thuilot, F. Marmoiton, and P. Martinet, “A global control
strategy for urban vehicles platooning relying on nonlinear decoupling
laws,” in Proc. Intell. Robots Syst., 2005, pp. 2875–2880.
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
Le Yi Wang (S’85–M’89–SM’01–F’12) received
the Ph.D. degree in electrical engineering from
McGill University, Montreal, QC, Canada, in 1990.
Since 1990, he has been with Wayne State University, Detroit, MI, USA, where he is currently
a Professor with the Department of Electrical and
Computer Engineering. His research interests include complexity and information, system identification, robust control, H ∞ optimization, time-varying
systems, adaptive systems, hybrid and nonlinear
systems, information processing and learning, and
medical, automotive, communications, power systems, and computer applications of control methodologies.
Dr. Wang was a Keynote Speaker at several international conferences. He
served as an Associate Editor for the IEEE T RANSACTIONS ON AUTOMATIC
C ONTROL and several other journals. He currently serves as an Associate
Editor for the Journal of System Sciences and Complexity and the Journal of
Control Theory and Applications.
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
George G. Yin (S’87–M’87–SM’96–F’02) received
the B.S. degree in mathematics from the University
of Delaware, Newark, DE, USA, in 1983; the M.S.
degree in electrical engineering; and the Ph.D. degree in applied mathematics from Brown University,
Providence, RI, USA, in 1987.
Since 1987, he has been with Wayne State University, Detroit, MI, USA, where he became a
Professor in 1996. His research interests include
stochastic systems and applied stochastic processes
and applications.
Dr. Yin has served on several technical committees. He was the Cochair of a
couple of American Mathematical Society–Irish Mathematical Society–Society
for Industrial Applied Mathematics (AMS–IMS–SIAM) Summer Conferences
and the Cochair of the 2011 SIAM Control Conference. He has served as
an Associate Editor for several journals, including the SIAM Journal on
Control and Optimization, Automatica, and the IEEE T RANSACTIONS ON
AUTOMATIC C ONTROL.
15
Hongwei Zhang (S’01–M’07–SM’13) received the
B.S. and M.S. degrees in computer engineering from
Chongqing University, Chongqing, China, and the
Ph.D. degree in computer science and engineering
from The Ohio State University, Columbus, OH, USA.
He is currently an Associate Professor of computer science with Wayne State University, Detroit,
MI, USA. His research has been an integral part
of several U.S. National Science Foundation (NSF)
and Defense Advanced Research Projects Agency
projects, such as the GENI WiMAX and the ExScal
projects. His main research interests include modeling, algorithmic, and systems issues in wireless, vehicular, embedded, and sensor networks.
Dr. Zhang received the NSF CAREER Award.
IE
E
Pr E
oo
f
AQ3
XU et al.: ENHANCED SAFETY OF HIGHWAY VEHICLE PLATOONS
977
978
979
980
981
982
983
984
985
986
987
988
989
990