MEAM Seminar Series Summer 2015

For Spring 2015 Seminars, click here.

Seminars are held on Tuesday mornings, with coffee at 10:30 am in the Towne Building and the seminar beginning at 10:45 am in Towne Building, room 337 (unless otherwise noted).

To be added to the MEAM Events mailing list (which sends notifications regarding all departmental seminars and events) please email us at meam-events@lists.seas.upenn.edu.


June 1

PhD Seminar

Philip Dames, PhD Candidate, University of Pennsylvania
Advisor: Vijay Kumar

3:45 pm, Towne 337

"Multi-Robot Active Information Gathering Using Random Finite Sets"

Abstract:

Information gathering tasks are becoming increasingly important in the modern world, ranging from infrastructure inspection to environmental monitoring to search and rescue. Teams of mobile sensor platforms have the ability to automate these information gathering tasks, which are often too dull, dirty, or dangerous for humans to perform. Additionally, robots are able to gather information that humans simply cannot gather, either by using mobility (e.g., flying overhead or crawling through rubble) or advanced sensors (e.g., multi-spectral cameras). In all of these tasks, the robot team must detect and localize an unknown number of objects of interest within the environment.

This talk will describe a unified estimation, control, and communication framework for active information gathering for multi-robot teams. There are many sources of uncertainty in these scenarios: in the number of objects of interest within the environment, in the locations of the objects of interest, in the sensor readings (e.g., false positive and false negative detections and noisy measurements), etc. We use the Probability Hypothesis Density (PHD) filter to simultaneously estimate the number of objects in the environment and their locations while taking into account these uncertainties. Using sets of potential actions generated at multiple length scales for each robot, the team selects the joint action that maximizes the expected information gain over a finite time horizon. We demonstrate the real-world applicability of the proposed autonomous exploration strategy through hardware experiments, exploring an office environment with a team of ground robots. We also conduct a series of simulated experiments, varying the planning method, target cardinality, environment, and sensor modality.

 

June 2

PhD Seminar

John Martin, PhD Candidate, University of Pennsylvania

Advisor: Robert Mauck

"Nanofibrous Disc-like Angle Ply Structures (DAPS) for Total Disc Replacement in a Small Animal Model"

Abstract:

The intervertebral discs are cartilaginous structures that impart multidirectional flexibility to the spine in support of routine physiologic loading. The discs of the lumbar spine are the largest avascular tissues in the body and, as a consequence, these discs predictably degenerate with age. Chronic low back pain is directly linked to intervertebral disc degeneration and is a significant socioeconomic burden, encompassing a yearly cost upwards of $100-$200 billion in the US.

My global hypothesis for this work is that cell-seeded synthetic discs can restore normal mechanical function to the spine after end-stage disc degeneration. To that end, our lab has developed what we call nanofibrous disc-like angle ply structures (DAPS) that replicate the natural structure, composition, and cellularity of the native intervertebral disc. In this talk, I will describe the optimization of the in vitro culture and fabrication processes of DAPS, the development of a small animal model of total disc replacement, and the evaluation of DAPS maturation and mechanical function in vivo.

 

 

June 8

PhD Defense

Philip Dames, PhD Candidate, University of Pennsylvania

Advisor: Vijay Kumar

2:30 pm, Levine 307

"Multi-Robot Active Information Gathering Using Random Finite Sets"

Abstract:

Many tasks in the modern world involve collecting information, such as infrastructure inspection, security and surveillance, environmental monitoring, and search and rescue. All of these tasks involve searching an environment to detect, localize, and track objects of interest, such as damage to roadways, suspicious packages, plant species, or victims of a natural disaster. In any of these tasks the number of objects of interest is often not known at the onset of exploration. Teams of robots can automate these often dull, dirty, or dangerous tasks to decrease costs and improve speed and safety.

 
This dissertation addresses the problem of automating data collection processes, so that a team of mobile sensor platforms is able to explore an environment to determine the number of objects of interest and their locations. In real-world scenarios, robots may fail to detect objects within the field of view, receive false positive measurements to clutter objects, and be unable to disambiguate true objects. This makes data association, \ie matching individual measurements to targets, difficult. To account for this, we utilize filtering algorithms based on random finite sets to simultaneously estimate the number of objects and their locations within the environment without the need to explicitly consider data association. Using the resulting estimates they receive, robots choose actions that maximize the mutual information between the set of targets and the binary events of receiving no detections. This effectively hedges against uninformative actions and leads to a closed form equation to compute mutual information, allowing the robot team to plan over a long time horizon. The robots either communicate with a central agent, which performs the estimation and control computations, or act in a decentralized manner. Our extensive hardware and simulated experiments validate the unified estimation and control framework, using robots with a wide variety of mobility and sensing capabilities to showcase the broad applicability of the framework.


June 16

PhD Defense

Heather Culbertson, PhD Candidate, University of Pennsylvania

Advisor: Katherine Kuchenbecker

2:00 pm, Towne 337

"Data-Driven Haptic Modeling and Rendering of Realistic Virtual Textured Surfaces"

Abstract:

The haptic sensations one feels when interacting with physical objects create a rich and varied impression of the objects, allowing one to gather information about the objects' physical characteristics such as their texture, shape, and compressibility. The human sense of touch excels at sensing and interpreting these haptic cues, even when the object is felt through an intermediary tool instead of directly with a bare finger. Dragging, pressing, and tapping a tool on the object allow you to sense the object's roughness, slipperiness, and hardness as a combination of vibrations and forces. Unfortunately, the richness of these interaction cues is missing from many virtual environments, leading to a less satisfying and less immersive experience than one encounters in the physical world. However, we can create the perceptual illusion of touching a real object by displaying the appropriate haptic signals during virtual interactions.

This thesis presents methods for creating haptic models of textured surfaces from acceleration, force, and speed data recorded during physical interactions. The models are then used to synthesize haptic signals that are displayed to the user during rendering through vibrotactile and/or kinesthetic feedback. The haptic signals, which are a function of the interaction conditions and motions used during rendering, must respond realistically to the user's motions in the virtual environment. We conducted human subject studies to test how well our virtual surfaces capture the psychophysical dimensions humans perceive when exploring textured surfaces with a tool.


Three haptic rendering systems were created for displaying virtual surfaces using these surface models. An initial system displayed virtual versions of textured surfaces on a tablet computer using models of the texture vibrations induced when dragging a tool across the real surfaces. An evaluation of the system showed that displaying the texture vibrations accurately captured the surface's roughness, but additional modeling and rendering considerations were needed to capture the full feel of the surface. Using these results, a second system was created for rendering a more complete three-dimensional version of the haptic surfaces including surface friction and event-based tapping transients in addition to the texture vibrations. An evaluation of this system showed that we have created the most realistic haptic surfaces to date. The force-feedback haptic device used in this system, however, was not without its limitations, including low surface stiffness and undesired inertia and friction. We developed an ungrounded haptic augmented reality system to overcome these limitations. This system allowed us to change the perceived texture and friction of a physical three-dimensional object using the previously-developed haptic surface models.

 

July 7

PhD Seminar

Alison Koser Patteson, PhD Candidate, University of Pennsylvania
Advisor: Paulo Arratia

10:45 am, Towne 321

"Living fluids and their interactions with particles and polymers"

Abstract:

Living fluids, or fluids that contain living matter such as swimming microorganisms, are of great technological and scientific interest. For instance, microorganisms colonize in the mucus of human stomachs and lungs, which contain particles and/or polymers. These fluid constituents impart nonlinear complex material properties, such as shear-thinning viscosity and elasticity. Although swimming microorganisms have been well studied in simple water-like fluids, the motility of bacteria in complex fluids is not fully understood. Many bacteria utilize a run and tumble swimming gait to seek food and flee toxins. This combination of straight swimming segments (runs) and sudden erratic rotations (tumbles) controls their spread and diffusion. We aim to provide a systematic experimental study to explore the interaction between the run and tumble dynamics of the model bacteria E. coli and complex polymeric solutions. By directly tracking the position and orientations of cells, we find that bacterial transport dramatically depends on the fluid properties. We show that the swimming velocity increases with elasticity and tumble frequency decreases with increasing fluid viscosity. Simultaneously, we observe that the swimming of E. coli can drive fluids out of equilibrium. This leads to fascinating features not possible in passive fluids. In particular, we find the anomalous size-dependent diffusion of hard spheres in suspensions of bacteria. Our results suggest novel ways in which the transport and dynamics of swimming microorganisms, flexible polymers, and hard spheres can be controlled.


July 14

David Gagnon
PhD Seminar

10:45 am, Towne 321


Abstract: TBA

July 16

Michael Norton
PhD Defense

Abstract: TBA

July 17

Yi Yang
PhD Defense

Abstract: TBA

July 28

James Paulos
PhD Seminar

10:45 am, Towne 321


Abstract: TBA

August 4

Sheng Mao
PhD Seminar

Abstract: TBA

August 11

Nathan Ip
PhD Seminar

Abstract: TBA

August 18

Helen Minsky
PhD Seminar

Abstract: TBA