nt years to offer data acquisition capabilities on par
with those of traditional tethered data acquisition systems. Entire structural monitoring systems
assembled from wireless sensors have proven to be low-cost, easy to install and accurate. However, the
functionality of wireless sensors can be further extended to include actuation capabilities. Wireless
sensors capable of actuating a structure could serve as building blocks of future generations of structural
control systems. In this study, a wireless sensor prototype capable of data acquisition, computational
analysis and actuation is proposed for use in a real-time structural control system. The performance of a
wireless control system is illustrated using a full-scale structure controlled by a semi-active
magnetorheological (MR) damper and a network of wireless sensors. One wireless sensor is designated
as a controller that automates the task of collecting state data, calculating control forces, and issuing
commands to the MR damper, all in real-time. Additional wireless sensors are installed to measure the
acceleration and velocity response of each system degree-of-freedom. Base motion is applied to the
structure to simulate seismic excitations while the wireless control system mitigates the inter-story drift
and relative acceleration response of the structure. An optimal linear quadratic regulation (LQR) solution
is formulated for embedment within the computational cores of the wireless sensors.

KEY WORDS: Structural control; Wireless sensors; Embedded computing; Magnetorheological dampers


1. INTRODUCTION

Recent natural catastrophes have revealed the vulnerabilities of critical civil infrastructure systems
(bridges, buildings, tunnels, dams) exposed to earthquakes, hurricanes, and typhoons. To mitigate
structural responses resulting from dynamic loads, feedback control systems have been proposed by Yao

*
Corresponding author.
E-mail address:
jerlynch@umich.edu
(Assistant Professor Jerome P. Lynch) [1] for installation in civil structures. Since that time, feedback control systems have been widely adopted
with over 50 buildings and 20 long-span bridges in Asia currently employing feedback control [2]. Early
structural control systems proposed for civil structures employed large actuators for the direct application
of control forces. While active control systems were successful at mitigating structural responses to wind
loads, force capacities of actuators often saturate during large seismic events thereby limiting their
effectiveness. In response to this limitation, the concept of semi-active structural control was proposed.
Unlike the actuators employed in active control systems, semi-active control devices are designed to
develop internal structural forces by changes to the damping and stiffness properties of the structure.
Examples of semi-active devices include, but are not limited to: active variable stiffness (AVS) devices
[3], semi-active hydraulic dampers (SHD) [4], electrorheological (ER) dampers [5], and
magnetorheological (MR) dampers [6]. A benefit of developing control forces in a structure indirectly is
that semi-active control devices consume an order of magnitude less power than actuators associated with
active control systems [7]. In addition to inherent energy efficiencies, semi-active control devices are
also compact and low-cost. These attractive attributes encourage the use of large numbers of semi-active
control devices in a single structure; examples include 88 SHD devices in the Shiodome Tower, Tokyo
and over 350 SHD devices installed in the Mori Tower, Tokyo [2].

As recent installations suggest, future semi-active control systems will continue to be defined by ever
greater nodal densities. As structural control systems grow in size, the design and installation
complexities of the systems increase in tandem. For example, structural control systems currently employ
extensive lengths of coaxial wire to accommodate communication between sensors, actuators and a
centralized controller. As nodal densities increase, more coaxial wire is needed for communication. In
2002, the installation of coaxial wire between sensors and a central data repository has been cited to cost
as high as a few thousand dollars per sensor channel [8]. As a result, the benefit derived from additional
control devices are eroded by the high installation costs associated with increasing lengths of coaxial
wire. To eradicate the high cost of a wired control system, the use of wireless communications is
proposed for systems defined by high nodal densities.

Other researchers have previously explored wireless communications for adoption in feedback control
systems. Unlike current control systems that have dedicated coaxial wires between sensors, actuators and
the centralized controller, a control system adopting wireless communications requires sensors and
controllers to share a common wireless medium for communication. When a closed-loop control system
is implemented using a common communication medium (wired or wireless), network quality strongly
influences the performance of the control solution. Specifically, time delays governed by deterministic
and stochastic processes are often introduced by the network. Lian, et al. [9] proposes the use of network
protocols that guarantee deterministic transmission times between transmitting and receiving nodes so
that delays can be accounted for by the control solution. However, stochastic delays sometimes can not
be avoided and are difficult to account for a priori. Specific to wireless networks, multiple researchers
have begun to explore real-time closed-loop control using wireless sensors. Eker, et al. [10] explores the
implementation of a linear quadratic regulation (LQR) control solution using a wireless controller that
communicates using the Bluetooth wireless communication protocol. Randomly varying delays within the
wireless communication channel are compensated for in the design of the LQR controller using a
compensation technique proposed by Nilsson, et al. [11]. Ploplys, et al. [12] implements a closed-loop
control solution for an inverted pendulum system using a wireless sensor network communicating upon
the IEEE 802.11b communication standard. To ensure timely delivery of data packets, the User
Datagram Protocol (UDP) is adopted to provide fast sample rates and to reduce network congestion.

In this study, a real-time structural control system for civil structures is proposed using wireless
sensor networks. As a fundamental building block of the control system, a wireless sensor prototype is designed to provide the functionality required for real-time control including data collection, computation
and actuation. The hardware design of the wireless sensor described herein is largely based upon a
wireless sensor previously proposed for infrastructure monitoring [13]. The actuation interface of the
modified wireless sensor is designed to output an analog voltage signal to command semi-active control
devices in real-time. One challenge associated with wireless communications is to ensure the reliable
delivery of data in the network. To address this challenge, a reliable wireless communication protocol is
proposed based upon a time division multiple access (TDMA) communication scheme. The feasibility of
a wireless control system is validated using a full-scale three-story steel structure excited by seismic
ground motions. A 20 kN MR damper is installed at the base of the structure for mitigation of structural
responses (inter-story drifts and floor accelerations). The bounded input-bounded output (BIBO) stability
property of semi-active dampers protect the test structure from becoming unstable should the wireless
control system perform poorly. Two control system architectures are implement; one architecture adopts
velocity transducers while the second adopts accelerometers. The performance of the wireless control
system will be quantified by comparing the closed-loop control performance to utilizing the MR damper
in a passive configuration. In addition, the wireless control system performance is also compared to that
of a control system implemented using a wired laboratory data acquisition system. As a result of
computational and communication overhead, the wireless control system is operated at a 12.5 sample rate
while the baseline tethered control system operates at 200 Hz. The wireless control system is shown to be
both effective in reducing structural responses and reliable in the wireless delivery of state data at each
time step.


2. PROTOTYPE WIRELESS SENSOR FOR MONITORING AND CONTROL

Numerous commercial and academic wireless sensors have been proposed for monitoring civil structures
[14]. The hardware design of all of these wireless sensors can be divided into four functional
components: sensing interface, computational core, wireless communication channel and actuation
interface. The sensing interface allows a variety of sensing transducers to be interfaced to a wireless
sensor. A core element of the sensing interface is an analog-to-digital converter (ADC) which converts
analog sensor signals to binary representations. After data is collected and digitized by the sensing
interface, it is passed to the computational core where data is stored, analyzed and readied for
communication. The third functional component is the wireless communic