Current research areas in Mechanical Systems
The major research foci and participating faculty in mechanical systems
are:
1. Robotics (Professors Kumar and Ostrowski)
2. Dynamics and control (Professors Kumar and Ostrowski)
3. Design and manufacturing (Professors Ananthasuresh, Kumar,
Ostrowski and Ulrich)
4. Micro-electromechanical systems (Professors Ananthasuresh, Bau and
Ostrowski)
There are active collaborations with other departments including
bioengineering (Professor Bogen), computer and information science
(Professors Bajcsy, Metaxas, Mintz and Paul), electrical engineering
(Professor Santiago and Zemel), and operations and information
management in the Wharton School (Professors Harker and Ulrich).
ROBOTICS
Manipulation and grasping
Professor Kumar and his co-workers are actively investigating the
control of robot arms, particularly in tasks that require two or more
arms to cooperatively manipulate objects. The advantages of using
multiple arms include a larger payload capacity andthe ability to
manipulate large objects without using special purpose fixtures [YK91].
However the coupling between the arms and the requirements on stable
grasping [HK 96] introduce complexity into the control issues. A
concurrent study of human manipulation and grasping [ZKDH 95] has
resulted in control algorithms that yield optimal load sharing and
cooperation between the two arms. A variation on this work is the
development of control algorithms for coordination between autonomous
cooperating mobile manipulators [DWZK 96] and the application to
teleoperated material-handling in hazardous environments.
Locomotion systems
An important goal of the work in robot dynamics and control is to
develop a better understanding of the biological systems that have
inspired recent advances in robotics. Professor Ostrowski and his
co-workers have used mathematical tools from differential geometry to
study the locomotion of snakes, inchworms and paramecia in order to
develop robotic locomotion systems [OB 96a, OB 96b ODK 96]. Professor
Kumar and his students have studied the locomotion of quadrupeds (e.g.,
goats and horses) and hexapods (insects) [PKBW 88]. This has resulted in
the design of walking robots and the development of terrain adaptive
control algorithms for such robots [KW 89a, KW 89b]. The example of one
such design is seen in the wheelchair with legs in Figure 1, where the
conventional wheel-based design has been used in conjunction with
insect-like legs to give wheelchair users the ability to navigate
unstructured environments [WKKH 95, KK 96].
Motion planning
Motion planning is the problem of determining optimal trajectories and
actuator forces for robotic systems. Professor Kumar and his coworkers
have formulated the motion planning problem as a problem in the calculus
of variations on a Riemannian manifold [ZDK 96]. In particular, there
is a class of systems in which the dynamics can be approximated by the
rigid body dynamics of a single body. In such cases, the problem of
generating optimal smooth motions can be formulated using the methods of
differential geometry and Lie groups. In fact, the optimal motion turns
out to be a geodesic on the Riemannian manifold that is the set of all
possible positions and orientations [ZKC 96]. The resulting motion
depends on the choice of the metric which must make sense for physical
applications and be invariant with respect to transformations. The
choice of metrics, the resulting connections and geodesics are an
active area of research.
CONTROL
Locomotion
Professor Ostrowski and fellow researchers are using mathematical tools
from differential geometry to establish a general framework of the
control of a class of locomotion systems exemplified by snakes,
inchworms, paramecia, eel and fish with practical applications to
attitude control of satellites, control of underwater submersibles and
flight control. An important direction of this research is the
development of robotic mechanisms capable of swimming. Such systems
offer many potential advantages for use in underwater activities. The
streamlined bodies of eels and fish are very efficient at generating
locomotion. Also, the flexibility and agility of eel- and fish-like
systems can be used in underwater search and rescue to enter areas that
are not easily accessible by traditional types of vehicles.
Smart Structures and Actuators
An important component of this work involves studying methods of
actuation for undulatory systems. Traditional actuators, such as DC
motors, are awkward and bulky compared to those found in biological
systems. For this reason, Professor Ostrowski has begun a program to
investigate the mechanics and control of smart materials and actuators,
e.g., using shape memory alloy wires (SMA). These thermally activated
wires often have a higher strength-to-weight ratio than motors and can
be directly integrated into the material in which they are housed.
Current research involves modeling composite ³smart² structures using
homogenization (averaging) techniques.
Another avenue of current locomotion research involves studying low
Reynolds' number swimming (this is the swimming regime for
micro-organisms, such as paramecia). This research builds on previous
results in order to develop robotic systems. A medium-range goal is to
design and fabricate micro-locomotion devices on the cellular biological
scale using MEMS technology. It is envisioned that one day these types
of robots could swim through the blood stream to deliver drugs or even
fight diseases.
Optimal control
Professors Kumar and Ostrowski and their students are working on the
optimal control of multiple degree-of-freedom electromechanical systems
for robotics and manufacturing applications. Typically such systems are
underconstrained and there is an opportunity to exploit the surplus
degrees of freedom to optimize the trajectory and/or the actuator
forces. Previous work has resulted in the development of efficient
algorithms for determing the optimal open-loop control inputs for
affine, controllable systems with state and/or input inequality and/or
equality constraints. These algorithms have been shown to have
applications to robotic manipulation systems [ZDK 96], locomotion
systems [ODK 96], and biological systems [ZDHK 95]. An important
direction for future research is the modeling of state and modeling
uncertainty while making minimal assumptions on the nature of the
uncertainty.
MICRO-ELECTRO-MECHANICAL SYSTEMS (MEMS)
Professor Ananthasuresh's interests in MEMS are directed towards both
design and fabrication of micromechanical devices. The design effort is
focused on applyting the easy-to-fabricate monolithic compliant
mechanism concept in developing several devices for micro mechanical
manipulation. These devices have applications in microsensors,
microactuators and microsurgery. His work also deals with modeling
coupled elastomechanics-electrostatics. Interesting hysteresis
phenomenon in contact electromechanics is already demonstrated through
3-D modeling. Several compliant MEMS devices are fabricated using bulk
and surface micromachining processes. Professor Ananthasuresh is also
interested in designing "material microstructure" for tailored effective
properties by optimizing the basic bulding block which is similar to a
compliant mechanism. As an illustrative example, different "material
designs" with negative Poisson's ratio have been fabricated using
wafer-bonding micromachining.
DESIGN & MANUFACTURING
Compliant Mechanisms
Professor Ananthasuresh is studying a novel concept in mechanical design
which minimizes assembly and sometimes avoids it altogether. The key to
this concept is designing mechanical systems to be "compliant and
strong" rather than "rigid and strong", and effectively using elastic
deformation, rather than jointed rigid members only, for motion and
force transmission. By using continuum mechanics based formulation for
optimally synthesizing the shape of these deforming structures, or
compliant mechanisms, a firm foundation is being laid for systematic
design. Applications of this in the manufacture-efficient and
cost-effective product design are being explored. He is also applying
this "single-piece design" idea to the fast-growing interdisciplinary
area of Micro-Electro-Mechanical Systems (MEMS) and has fabricated
compliant micro devices that have applications in a wide variety of
micro devices and micro surgery. Thus, this research is an amalgam of
three important areas viz. compliant mechanisms, topology optimization,
and MEMS.
Design Optimization
The current focus in the area of design optimization is on structural
optimization of supporting structures (or suspensions) for MEMS devices
and optimal design of kinematic mechanisms. Professor Ananthasureshıs
work addresses the issues of topology and shape optimization of
supporting structures for micromechanical (MEMS) structures to achieve
required stiffness and dynamical characteristics. Professors
Ananthasuresh and Kumar and their students are developing optimization
based techniques for customized design of mechanisms. This is an
integral part of a larger effort in virtual prototyping. Optimal design
of a new configuration of single-actuator, open-chain linkages are being
investigated. These linkages are amenable for customized design and are
capable of generating a wide variety of complex tasks. They lie in the
middle ground between overly expensive robots and functionally
resctrictive single-task closed-loop mechanisms.
Rapid prototyping of customized assistive devices
Professors Ananthasuresh, Harker, Kumar and Ostrowski are investigating
methods to rapidly design and prototype customized products that are
worn by humans. An important focus of this work is the design and
manufacture of one-of-a-kind assistive aids for people with disabilities
[KBHH 96]. One of the technical challenges is virtual prototyping, the
process of detail-designing and prototyping a product and testing and
evaluating it on a computer before committing to the expensive process
of manufacturing. This process is particularly beneficial for low volume
customized parts. As an example, a feeding device for quadriplegics was
designed using such methods.
Click on the following topics for additional information.
(The descriptions will be available at a later time)
Compliant Mechanisms
Computer Graphics
Controls
Customized Design for Rapid Prototyping
Dynamics
Micro-Electro-Mechanical Systems (MEMS)
Optimal Design
Product Design and Management
Robotics
Professor Ananthasuresh is studying a novel concept in mechanical design
which minimizes assembly and sometimes avoids it altogether. The key to
this concept is designing mechanical systems to be "compliant and
strong" rather than "rigid and strong", and effectively using elastic
deformation, rather than jointed rigid members only, for motion and
force transmission. By using continuum mechanics based formulation for
optimally synthesizing the shape of these deforming structures, or
compliant mechanisms, a firm foundation is being laid for systematic
design. Applications of this in the manufacture-efficient and
cost-effective product design are being explored. He is also applying
this "single-piece design" idea to the fast-growing interdisciplinary
area of Micro-Electro-Mechanical Systems (MEMS) and has fabricated
compliant micro devices that have applications in a wide variety of
micro devices and micro surgery. Thus, this research is an amalgam of
three important areas viz. compliant mechanisms, topology optimization,
and MEMS.
The current focus in the area of design optimization is on structural
optimization of supporting structures (or suspensions) for MEMS devices
and optimal design of kinematic mechanisms. Professor Ananthasureshıs
work addresses the issues of topology and shape optimization of
supporting structures for micromechanical (MEMS) structures to achieve
required stiffness and dynamical characteristics. Professors
Ananthasuresh and Kumar and their students are developing optimization
based techniques for customized design of mechanisms. This is an
integral part of a larger effort in virtual prototyping. Optimal design
of a new configuration of single-actuator, open-chain linkages are being
investigated. These linkages are amenable for customized design and are
capable of generating a wide variety of complex tasks. They lie in the
middle ground between overly expensive robots and functionally
resctrictive single-task closed-loop mechanisms.