A CPSS Approach for Emergency Evacuation in Building Fires
C YBER-PHYSIC AL-SOCIAL SYS TE MS
Editor: Liuqing Yang, Colorado State University, [email protected]
A CPSS Approach for
Emergency Evacuation
in Building Fires
Yuling Hu, Beijing Institute of Technology and Beijing University
of Civil Engineering and Architecture
Fei-Yue Wang and Xiwei Liu, Chinese Academy of Sciences
L
iving with high-rise buildings or skyscrapers
has raised important issues and huge chal-
lenges for occupant evacuation in fi res or other
emergencies. To alleviate casualties, it’s important
to have emergency evacuation plans in place before fi re outbreaks occur. However, there are too
many factors involving uncertainty, diversity, and
complexity in evacuations. How, then, can we
model human behaviors during evacuation processes? And how do we verify or evaluate those
prearranged emergency plans? Furthermore, how
can we guide evacuations timely and effectively
under varied, changing scenarios?
To address these problems, we propose a cyberphysical-social systems (CPSS) approach based
(ACP) methodology using artificial systems, computational experiments, and parallel execution.1–3
CPSS is the extension of cyber-physical systems
(CPS), which integrate with human and social
characteristics and bridge the physical world, cyberspace, and human society together.4 In a CPSS
approach, building structures, fi re scenarios, evacuees, and managers can be fully connected and interactive. To model the complex processes of evacuations, artificial evacuation systems are built and
used to determine evacuation options for the physical evacuation systems. Based on the artificial
evacuation systems, such prearranged emergency
plans can be tested and evaluated by computational experiments. Then, constructing datadriven parallel mechanisms between the artificial
evacuation systems and the physical evacuation
systems is the ultimate way to achieve evacuation
guidance in real time.
System Framework
As Figure 1 shows, the strategy database, artificial evacuation system, computational experiment
48
platform, and visualization output are the main
components. Various emergency evacuation strategies, static or dynamic, are included in strategy
databases, and are applied to artificial evacuation
systems. The computational experiment platform
can be considered as the laboratory for conducting
evacuation experiments.
As the counterparts of actual and physical evacuation systems, artificial evacuation systems are
developed by three interactional components:
• Building structures. According to the needs of
research, in the artificial evacuation systems we
can flexibly construct different building spatial
distributions and explore the influences of these
different structures on evacuation efficiencies.
• Fire scenarios. Obviously, different fi re scenarios will distinctly impact evacuees’ psychology
and physiology and change the evacuation efficiencies, even with the same strategy.
• Evacuees. It’s been verified that agent-based
modeling methods accurately imitate evacuees—
from physiological and psychological characteristics to movements and behaviors.5 To differentiate and distinguish among the evacuees, each
individual agent should have some basic physiological and psychological attributes. Additionally, the agent needs capabilities for external fi re
environment perception and decision making.
After the use of artificial evacuation systems for
modeling evacuation processes, we can design and
perform controllable evacuation experiments with
the computational experiment platform, which
enables us to evaluate and quantitatively analyze
various factors, especially human factors on evacuation efficiency, a difficult problem to address
thus far. According to the problems and objectives of investigation, many existing experimental
1541-1672/14/$31.00 © 2014 IEEE
Published by the IEEE Computer Society
IEEE INTELLIGENT SYSTEMS
Physical
evacuation system
Strategy
database
Building
structures
Fire
scenarios
Experiment
execution
Experiment
analysis
Evacuees
Evacuation
strategies
Experiment
design
Artificial
evacuation system
2D display
Strategies
evaluation
3D display
Computational
experiments
Visualization
Figure 1. System framework. The main components are the strategy database, artificial evacuation system, computational
experiment platform, and visualization output.
design theories and methods, such as
single-factor or multifactor experimental design, orthogonal or uniform experiment design, and so on,
can be utilized in evacuation computational experiments. 6
The visualization is also an important part of our system. With the
development of virtual reality technologies, visualization technologies
can help us easily identify evacuation
bottlenecks or congestion changes.
That’s helpful for improving or optimizing evacuation strategies.
Strategy Evaluation
By integrating the agent, grid computing, and fire simulation technologies,
we construct an artificial evacuation
system (see Figure 2). The building
structures, individual attributes, occupant distributions, and fire scenarios can be flexibly altered according
to investigative needs or in reference
to the actual evacuation systems. Figure 3 shows two representative fire
scenarios to illustrate the developments of fire environments over time.
MAY/JUNE 2014
(b)
(a)
(c)
Figure 2. An artificial evacuation system with 3D display. (a) The whole building and
evacuees, (b) the first floor, and (c) the second through tenth floors.
Note that this environment information is perceived by individual agents
and it affects their behaviors and
movement capabilities in real time.
Evacuation strategies can by evaluated by applying them to artificial
evacuation systems. Considering the
strategies listed in Table 1, uncertainties or random factors are introduced
by the Monte Carlo method and
www.computer.org/intelligent
some statistical knowledge in experiments and evaluation.7 With required
safe egress time (RSET) as the key
performance index, Figure 4 shows
computational experimental results
on two fire scenarios with different
strategies. Clearly, the results show
each strategy’s evacuation efficiency
while also revealing each fire scenario’s effects.
49
60.00
3.00
50.00
40.00
30.00
20.00
10.00
0.00
0.00
Smoke layer temp in right staircase
Smoke layer temp in ignition room
Smoke layer temp in corridor
2.50
2.00
1.50
1.00
0.50
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
60.00
Smoke layer height in right staircase
Smoke layer height in ignition room
Smoke layer height in corridor
Smoke layer CO2 concentration (ppm)
3.50
Smoke layer height (m)
Smoke layer temperature (C)
70.00
50.00
40.00
30.00
20.00
10.00
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
CO2 concentration in right staircase
CO2 concentration in ignition room
CO2 concentration in corridor
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
(a)
500.00
400.0
3.50
Smoke layer CO2 concentration (ppm)
3.00
400.00
Smoke layer height (m)
Smoke layer temperature (C)
450.00
350.00
300.00
250.00
200.00
150.00
100.00
2.50
2.00
1.50
1.00
0.50
50.00
0.00
0.00
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
Smoke layer temp in ignition room
Smoke layer temp in right staircase
Smoke layer temp in corridor
Smoke layer height in ignition room
Smoke layer height in right staircase
Smoke layer height in corridor
350.0
300.0
250.0
200.0
150.0
100.0
50.0
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
CO2 concentration in ignition room
CO2 concentration in right staircase
CO2 concentration in corridor
(b)
Figure 3. The evolution of smoke layer temperature, height, and carbon dioxide concentration in the ignition room, corridor,
and nearby staircase. (a) The ignition location is set in the middle room on the third floor by main fire burning. (b) The ignition
location is set on the left room near the staircase on the eighth floor by wooden panel burning. Here, ppm is the concentration
unit (parts per million).
Table 1. Evacuation strategy database.
Strategy
Notes
Strategy1
The shortest-path evacuation
The shortest-distance export evacuation strategy.
Strategy2
Exit equal-density evacuation
Evacuee is evenly distributed to stairs or exits.
Strategy3
Fire floor prior evacuation
It combines with the shortest-distance export evacuation.
Strategy4
Fire floor prior evacuation
It combines with the exit equal-density evacuation.
Obviously, improved or optimal
strategies can be obtained by such computational experiments and be used to
train managers or drill occupants to alleviate the effect of disasters.
Evacuation Guidance
In our approach, the most significant
function is to guide evacuation in real
50
time. Considering the dynamics and
uncertainties in evacuation processes,
the data-driven parallel mechanism8
between the actual evacuation systems and the artificial evacuation systems is important and useful.
In real-world applications, we can
use the artificial evacuation systems
to emulate the actual evacuation
www.computer.org/intelligent
system for learning and training.
Through analysis or evaluation of
evacuees’ behaviors on computers
with artificial systems in advance, we
can improve and optimize the actual
evacuation process’s performance. At
the same time, various information
and data collected by actual sensors
can be used to calibrate the artificial
IEEE INTELLIGENT SYSTEMS
Strategy1
Strategy2
Strategy3
Strategy4
0
100
200
300
400 500 600
Elasped time (s)
700
800
Elapsed people total
Elapsed people total
1,300
1,200
1,100
1,000
900
800
700
600
500
400
300
200
100
0
900 1,000
1,300
1,200
1,100
1,000
900
800
700
600
500
400
300
200
100
0
0
100
200
300
400
500
600
700
800
900 1,000 1,100
Elasped time (s)
(a)
(b)
1,100
1,000
1
900
2
3
4
1,000
800
900
700
800
600
RSET (s)
RSET (s)
Strategy1
Strategy2
Strategy3
Strategy4
500
400
1
3
2
4
700
600
500
400
300
300
200
200
100
100
0
1
2
3
0
4
1
Strategy number
(c)
2
3
Strategy number
4
(d)
Figure 4. The results of a computational experiment, using four strategies in each scenario (refer to Table 1 for the strategies).
(a) The relationship between elapsed time and the number of people who escaped in scenario 1. (b) The relationship between
elapsed time and the number of people who escaped in scenario 2. (c) Required safe egress time (RSET) for the strategies in
scenario 1. (d) RSET for the strategies in scenario 2.
evacuation systems and run them in
parallel with actual evacuation processes in real time. Figure 5 shows the
data-driven parallel mechanism.
The CPSS approach we proposed
here provides a tool to verify or evaluate emergency evacuation plans. It
enables operators to explore the possible consequences of decisions. In
addition, by building a data-driven
parallel mechanism between actual
and artificial evacuation systems, this
makes it possible to guide evacuation processes effectively in real time.
We hope further work using this approach will significantly improve fire
emergency management and public
safety.
MAY/JUNE 2014
Strategy database
Computational
experimental platform
Artificial evacuation
systems
Emergency
management platform
Smoke detector
Information
process
platform
...
Heat detector
Physical evacuation
systems
...
Video detector
Figure 5. Data-driven parallel mechanism between artificial evacuation systems and
physical evacuation systems.
www.computer.org/intelligent
51
Acknowledgments
This work was supported by Beijing Municipal Education Commission grant 051201515
and by National Natural Science Foundation
of China grants 61233001 and 91024030.
References
1. F.-Y. Wang, “Toward a Paradigm Shift
in Social Computing: The ACP Approach,” IEEE Intelligent Systems, vol.
22, no. 5, 2007, pp. 65–67.
2. F.-Y. Wang, “Parallel Control and Management for Intelligent Transportation
Systems: Concepts, Architectures, and
Applications,” IEEE Trans. Intelligent
Transportation Systems, vol. 11, no. 3,
2010, pp.1–9.
3. B. Ning et al., “ACP-Based Control and
Management of Urban Rail Transportation Systems,” IEEE Intelligent Systems,
vol. 26, no. 2, 2011, pp. 84–88.
NEW
STORE
4. F.-Y. Wang, “The Emergence of Intelligent Enterprises: From CPS to CPSS,”
IEEE Intelligent Systems, vol. 25, no.
4, 2010, pp. 85–88.
5. Y.-L. Hu et al., “ACP-Based Research
on Evacuation Strategies for HighRise Building Fire,” ACTA Automatica Sinica, vol. 40, no. 2, 2014,
pp. 185–196.
6. F.-H. Zhu et al., “A Case Study of Evaluating Traffic Signal Control Systems
Using Computational Experiments,”
IEEE Trans. Intelligent Systems, vol.
12, no. 4, 2011, pp. 1220–1226.
7. X. Zhang et al., “A Probabilistic Occupant Evacuation Model for Fire Emergencies Using Monte Carlo Methods,”
Fire Safety, vol. 58, no. 1, 2013,
pp. 15–24.
8. F.-Y. Wang, “Parallel Control: A Method for Data-Driven and Computational
Control,” ACTA Automatica Sinica,
vol. 39, no. 4, 2013, pp. 293–302.
Yuling Hu is a PhD candidate at the Beijing
Institute of Technology and an associate
professor at Beijing University of Civil Engineering and Architecture. Contact her at
[email protected].
Fei-Yue Wang is the vice president of the
Institute of Automation at the Chinese
Academy of Sciences. Contact him at feiyue.
[email protected].
Xiwei Liu is an associate professor of engi-
neering at the Chinese Academy of Sciences.
Contact him at [email protected].
Selected CS articles and columns
are also available for free at
http://ComputingNow.computer.org.
Save up to
40%
on selected articles, books,
and webinars.
Find the latest trends
and insights for your
• presentations
• research
• events
webstore.computer.org
52
www.computer.org/intelligent
IEEE INTELLIGENT SYSTEMS
Editor: Liuqing Yang, Colorado State University, [email protected]
A CPSS Approach for
Emergency Evacuation
in Building Fires
Yuling Hu, Beijing Institute of Technology and Beijing University
of Civil Engineering and Architecture
Fei-Yue Wang and Xiwei Liu, Chinese Academy of Sciences
L
iving with high-rise buildings or skyscrapers
has raised important issues and huge chal-
lenges for occupant evacuation in fi res or other
emergencies. To alleviate casualties, it’s important
to have emergency evacuation plans in place before fi re outbreaks occur. However, there are too
many factors involving uncertainty, diversity, and
complexity in evacuations. How, then, can we
model human behaviors during evacuation processes? And how do we verify or evaluate those
prearranged emergency plans? Furthermore, how
can we guide evacuations timely and effectively
under varied, changing scenarios?
To address these problems, we propose a cyberphysical-social systems (CPSS) approach based
(ACP) methodology using artificial systems, computational experiments, and parallel execution.1–3
CPSS is the extension of cyber-physical systems
(CPS), which integrate with human and social
characteristics and bridge the physical world, cyberspace, and human society together.4 In a CPSS
approach, building structures, fi re scenarios, evacuees, and managers can be fully connected and interactive. To model the complex processes of evacuations, artificial evacuation systems are built and
used to determine evacuation options for the physical evacuation systems. Based on the artificial
evacuation systems, such prearranged emergency
plans can be tested and evaluated by computational experiments. Then, constructing datadriven parallel mechanisms between the artificial
evacuation systems and the physical evacuation
systems is the ultimate way to achieve evacuation
guidance in real time.
System Framework
As Figure 1 shows, the strategy database, artificial evacuation system, computational experiment
48
platform, and visualization output are the main
components. Various emergency evacuation strategies, static or dynamic, are included in strategy
databases, and are applied to artificial evacuation
systems. The computational experiment platform
can be considered as the laboratory for conducting
evacuation experiments.
As the counterparts of actual and physical evacuation systems, artificial evacuation systems are
developed by three interactional components:
• Building structures. According to the needs of
research, in the artificial evacuation systems we
can flexibly construct different building spatial
distributions and explore the influences of these
different structures on evacuation efficiencies.
• Fire scenarios. Obviously, different fi re scenarios will distinctly impact evacuees’ psychology
and physiology and change the evacuation efficiencies, even with the same strategy.
• Evacuees. It’s been verified that agent-based
modeling methods accurately imitate evacuees—
from physiological and psychological characteristics to movements and behaviors.5 To differentiate and distinguish among the evacuees, each
individual agent should have some basic physiological and psychological attributes. Additionally, the agent needs capabilities for external fi re
environment perception and decision making.
After the use of artificial evacuation systems for
modeling evacuation processes, we can design and
perform controllable evacuation experiments with
the computational experiment platform, which
enables us to evaluate and quantitatively analyze
various factors, especially human factors on evacuation efficiency, a difficult problem to address
thus far. According to the problems and objectives of investigation, many existing experimental
1541-1672/14/$31.00 © 2014 IEEE
Published by the IEEE Computer Society
IEEE INTELLIGENT SYSTEMS
Physical
evacuation system
Strategy
database
Building
structures
Fire
scenarios
Experiment
execution
Experiment
analysis
Evacuees
Evacuation
strategies
Experiment
design
Artificial
evacuation system
2D display
Strategies
evaluation
3D display
Computational
experiments
Visualization
Figure 1. System framework. The main components are the strategy database, artificial evacuation system, computational
experiment platform, and visualization output.
design theories and methods, such as
single-factor or multifactor experimental design, orthogonal or uniform experiment design, and so on,
can be utilized in evacuation computational experiments. 6
The visualization is also an important part of our system. With the
development of virtual reality technologies, visualization technologies
can help us easily identify evacuation
bottlenecks or congestion changes.
That’s helpful for improving or optimizing evacuation strategies.
Strategy Evaluation
By integrating the agent, grid computing, and fire simulation technologies,
we construct an artificial evacuation
system (see Figure 2). The building
structures, individual attributes, occupant distributions, and fire scenarios can be flexibly altered according
to investigative needs or in reference
to the actual evacuation systems. Figure 3 shows two representative fire
scenarios to illustrate the developments of fire environments over time.
MAY/JUNE 2014
(b)
(a)
(c)
Figure 2. An artificial evacuation system with 3D display. (a) The whole building and
evacuees, (b) the first floor, and (c) the second through tenth floors.
Note that this environment information is perceived by individual agents
and it affects their behaviors and
movement capabilities in real time.
Evacuation strategies can by evaluated by applying them to artificial
evacuation systems. Considering the
strategies listed in Table 1, uncertainties or random factors are introduced
by the Monte Carlo method and
www.computer.org/intelligent
some statistical knowledge in experiments and evaluation.7 With required
safe egress time (RSET) as the key
performance index, Figure 4 shows
computational experimental results
on two fire scenarios with different
strategies. Clearly, the results show
each strategy’s evacuation efficiency
while also revealing each fire scenario’s effects.
49
60.00
3.00
50.00
40.00
30.00
20.00
10.00
0.00
0.00
Smoke layer temp in right staircase
Smoke layer temp in ignition room
Smoke layer temp in corridor
2.50
2.00
1.50
1.00
0.50
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
60.00
Smoke layer height in right staircase
Smoke layer height in ignition room
Smoke layer height in corridor
Smoke layer CO2 concentration (ppm)
3.50
Smoke layer height (m)
Smoke layer temperature (C)
70.00
50.00
40.00
30.00
20.00
10.00
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
CO2 concentration in right staircase
CO2 concentration in ignition room
CO2 concentration in corridor
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
(a)
500.00
400.0
3.50
Smoke layer CO2 concentration (ppm)
3.00
400.00
Smoke layer height (m)
Smoke layer temperature (C)
450.00
350.00
300.00
250.00
200.00
150.00
100.00
2.50
2.00
1.50
1.00
0.50
50.00
0.00
0.00
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
Smoke layer temp in ignition room
Smoke layer temp in right staircase
Smoke layer temp in corridor
Smoke layer height in ignition room
Smoke layer height in right staircase
Smoke layer height in corridor
350.0
300.0
250.0
200.0
150.0
100.0
50.0
0.00
0.00
200.00 400.00 600.00 800.00 1,000.00 1,200.00
Time (s)
CO2 concentration in ignition room
CO2 concentration in right staircase
CO2 concentration in corridor
(b)
Figure 3. The evolution of smoke layer temperature, height, and carbon dioxide concentration in the ignition room, corridor,
and nearby staircase. (a) The ignition location is set in the middle room on the third floor by main fire burning. (b) The ignition
location is set on the left room near the staircase on the eighth floor by wooden panel burning. Here, ppm is the concentration
unit (parts per million).
Table 1. Evacuation strategy database.
Strategy
Notes
Strategy1
The shortest-path evacuation
The shortest-distance export evacuation strategy.
Strategy2
Exit equal-density evacuation
Evacuee is evenly distributed to stairs or exits.
Strategy3
Fire floor prior evacuation
It combines with the shortest-distance export evacuation.
Strategy4
Fire floor prior evacuation
It combines with the exit equal-density evacuation.
Obviously, improved or optimal
strategies can be obtained by such computational experiments and be used to
train managers or drill occupants to alleviate the effect of disasters.
Evacuation Guidance
In our approach, the most significant
function is to guide evacuation in real
50
time. Considering the dynamics and
uncertainties in evacuation processes,
the data-driven parallel mechanism8
between the actual evacuation systems and the artificial evacuation systems is important and useful.
In real-world applications, we can
use the artificial evacuation systems
to emulate the actual evacuation
www.computer.org/intelligent
system for learning and training.
Through analysis or evaluation of
evacuees’ behaviors on computers
with artificial systems in advance, we
can improve and optimize the actual
evacuation process’s performance. At
the same time, various information
and data collected by actual sensors
can be used to calibrate the artificial
IEEE INTELLIGENT SYSTEMS
Strategy1
Strategy2
Strategy3
Strategy4
0
100
200
300
400 500 600
Elasped time (s)
700
800
Elapsed people total
Elapsed people total
1,300
1,200
1,100
1,000
900
800
700
600
500
400
300
200
100
0
900 1,000
1,300
1,200
1,100
1,000
900
800
700
600
500
400
300
200
100
0
0
100
200
300
400
500
600
700
800
900 1,000 1,100
Elasped time (s)
(a)
(b)
1,100
1,000
1
900
2
3
4
1,000
800
900
700
800
600
RSET (s)
RSET (s)
Strategy1
Strategy2
Strategy3
Strategy4
500
400
1
3
2
4
700
600
500
400
300
300
200
200
100
100
0
1
2
3
0
4
1
Strategy number
(c)
2
3
Strategy number
4
(d)
Figure 4. The results of a computational experiment, using four strategies in each scenario (refer to Table 1 for the strategies).
(a) The relationship between elapsed time and the number of people who escaped in scenario 1. (b) The relationship between
elapsed time and the number of people who escaped in scenario 2. (c) Required safe egress time (RSET) for the strategies in
scenario 1. (d) RSET for the strategies in scenario 2.
evacuation systems and run them in
parallel with actual evacuation processes in real time. Figure 5 shows the
data-driven parallel mechanism.
The CPSS approach we proposed
here provides a tool to verify or evaluate emergency evacuation plans. It
enables operators to explore the possible consequences of decisions. In
addition, by building a data-driven
parallel mechanism between actual
and artificial evacuation systems, this
makes it possible to guide evacuation processes effectively in real time.
We hope further work using this approach will significantly improve fire
emergency management and public
safety.
MAY/JUNE 2014
Strategy database
Computational
experimental platform
Artificial evacuation
systems
Emergency
management platform
Smoke detector
Information
process
platform
...
Heat detector
Physical evacuation
systems
...
Video detector
Figure 5. Data-driven parallel mechanism between artificial evacuation systems and
physical evacuation systems.
www.computer.org/intelligent
51
Acknowledgments
This work was supported by Beijing Municipal Education Commission grant 051201515
and by National Natural Science Foundation
of China grants 61233001 and 91024030.
References
1. F.-Y. Wang, “Toward a Paradigm Shift
in Social Computing: The ACP Approach,” IEEE Intelligent Systems, vol.
22, no. 5, 2007, pp. 65–67.
2. F.-Y. Wang, “Parallel Control and Management for Intelligent Transportation
Systems: Concepts, Architectures, and
Applications,” IEEE Trans. Intelligent
Transportation Systems, vol. 11, no. 3,
2010, pp.1–9.
3. B. Ning et al., “ACP-Based Control and
Management of Urban Rail Transportation Systems,” IEEE Intelligent Systems,
vol. 26, no. 2, 2011, pp. 84–88.
NEW
STORE
4. F.-Y. Wang, “The Emergence of Intelligent Enterprises: From CPS to CPSS,”
IEEE Intelligent Systems, vol. 25, no.
4, 2010, pp. 85–88.
5. Y.-L. Hu et al., “ACP-Based Research
on Evacuation Strategies for HighRise Building Fire,” ACTA Automatica Sinica, vol. 40, no. 2, 2014,
pp. 185–196.
6. F.-H. Zhu et al., “A Case Study of Evaluating Traffic Signal Control Systems
Using Computational Experiments,”
IEEE Trans. Intelligent Systems, vol.
12, no. 4, 2011, pp. 1220–1226.
7. X. Zhang et al., “A Probabilistic Occupant Evacuation Model for Fire Emergencies Using Monte Carlo Methods,”
Fire Safety, vol. 58, no. 1, 2013,
pp. 15–24.
8. F.-Y. Wang, “Parallel Control: A Method for Data-Driven and Computational
Control,” ACTA Automatica Sinica,
vol. 39, no. 4, 2013, pp. 293–302.
Yuling Hu is a PhD candidate at the Beijing
Institute of Technology and an associate
professor at Beijing University of Civil Engineering and Architecture. Contact her at
[email protected].
Fei-Yue Wang is the vice president of the
Institute of Automation at the Chinese
Academy of Sciences. Contact him at feiyue.
[email protected].
Xiwei Liu is an associate professor of engi-
neering at the Chinese Academy of Sciences.
Contact him at [email protected].
Selected CS articles and columns
are also available for free at
http://ComputingNow.computer.org.
Save up to
40%
on selected articles, books,
and webinars.
Find the latest trends
and insights for your
• presentations
• research
• events
webstore.computer.org
52
www.computer.org/intelligent
IEEE INTELLIGENT SYSTEMS