Powered Knee & Ankle Prosthesis

Individuals with transfemoral amputation are subject to higher levels of exertion when using energetically passive prostheses. Powered prostheses have been proven to reduce this increased metabolic expenditure as well as provide additional benefits of reducing increased loading seen on the intact limb post-amputation. The EPIC lab has developed a powered prosthesis that utilizes two motors to provide assistance independently at both the knee and ankle joints. There a variety of embedded sensors such as encoders, inertial measurement units (IMUs), and a 6-DOF loadcell to help extract user information as one ambulates across a variety of walking modes (i.e. overground, ramps, and stairs). We are interested in developing intelligent intent recognition systems using machine learning and sensor fusion techniques that restores and improves overall independence and mobility.

Graduate Student Lab Members: Krishan Bhakta, Jonathan Camargo, Jairo Maldonado

Collaborators: Dr. Aaron Ames, Dr. Lee Childers, Kinsey Herrin, Robert Kistenberg, Eric Ambrose, Rachel Gehlhar

Biomechanics and Sensor Fusion

This team focuses on instrumentation with wearable sensors, including IMU, goniometers, pressure sensors and a novel epidermal flexible EMG. We are interested in analyzing the information carried by these sensors to develop intent recognition algorithms and gait state estimation using machine learning techniques. We implemented a full data collection system including motion capture and force plates in a configurable experimental area that includes ramps, stairs and ground level walking. With this study, we can evaluate the biomechanics of ambulation at different conditions and get a better background for the development of controllers for assistive devices.

Graduate student: Jonathan Camargo.

Undergraduate collaborators: Noel Csomay-Shanklin, Will Flanagan, Bharat Kanwar, Aditya Ramanathan, Chris Mao.

Powered Hip Exoskeleton

Our autonomous powered hip exoskeleton augments human by providing a hip assistance across wide ranges of locomotion tasks. The device integrates human neural commands by measuring the surface electromyography (EMG) signals to control the device providing a more natural assistance to the user. We utilize these biological signals along with mechanical sensors from the device to better understand what the user intent is during different gait modes with different machine learning techniques (i.e. level-ground, stairs, and ramps). We primarily focus on translating this technology to more realistic settings such as outdoor terrain as well as to the clinical populations such as stroke survivors who experience limited mobility in community ambulation.

Graduate Student Lab Members: Inseung Kang, Dean Molinaro, Pratik Kunapuli, Julian Park, Henry Luk

Collaborators: Dr. Geza Kogler, Dr. Stephen Sprigle

Pediatric Knee Exoskeleton

The primary aim of the project is to help assist the children with limited mobility mainly caused by the knee joint in walking with the help of powered assistance. The specific targeted groups are the children who present hyper-extended or excessively flexed knee during walking. He is hoping that the understanding of the interaction between an exoskeleton device and the targeted users through this research help exoskeleton technology become and promote an effective robotic rehabilitation for children in their growth in the future.

Graduate Student Lab Members: Dawit Lee, Inseung Kang

Collaborators: Dr. Geza Kogler, Dr. Benjamin Rogozinski, Dr. Erin Eggebrecht

Rapid Operator Awareness via Mobile Robotics (ROAMR)

Human-robot interaction in structured environments has been made incredibly safe over the past few years. In these cases, the environment is controlled to ensure human safety. However, many spaces such as disaster zones, construction sites, and more are unstructured. There is a need to develop scientific principles that ensure human safety in these environments as well. The ROAMR project [NSF link] (PI, Dr. Anirban Mazumdar [DART Lab website], and Co-PI, Dr. Aaron Young), proposes the use of wearable robotics to quickly alert the user to threats, recognize the user’s intent, and assist in agile movement to ensure the human user is safe.

Graduate Student Lab Members: Aakash Bajpai, Inseung Kang

Optimization of Hip Exoskeleton Control and Physiological Outcomes

To realize the potential of exoskeletons, it is necessary to understand the two-way relationship between assistive robotics and the human biomechanical/metabolic system. Using an exoskeleton emulator (implemented with a hip exoskeleton end-effector a.k.a HipEE) actuated by powerful off-board motors, we are able to mimic passive devices (spring-based) and rapidly optimize variety of controllers (including proportional myoelectric, impedance, neuromuscular model based, hybrids, etc.) for their effects on muscles, tendons, and the energetic cost of walking. We understand these effects by collecting and analyzing motion capture data (kinematics and kinetics of motion), EMG signals (muscular activation), ultrasound images (muscular shape change/stiffness), and respiratory data (energetic cost). We will apply this collected information to optimize controllers to drive specific changes to the musculoskeletal system as well as assist/enhance locomotion in specific populations (elderly, stroke patients, military, etc.).

Graduate Student Lab Members: Ben Shafer

Collaborators: Dr. Greg Sawicki, Pawel Golyski, Stefan Klein

Biomechanical Risk Factors of Low Back Pain

Work-related musculoskeletal disorder is one of the most prevalent causes of occupational injuries. For this reason, extensive research has been conducted in investigating human biomechanics during lifting tasks; however, very little research has investigated the effects of highly dynamic tasks, such as throwing heavy or bulky objects. To better understand the risk factors associated with various dynamic materials handling techniques, a biomechanical investigation has been conducted using sensory data from EMG, motion capture and ground reaction forces. This data is analyzed using biomechanical modeling tools such as Vicon and OpenSim. Using these findings, we plan to develop novel wearable devices and controllers to mitigate these risk factors.




Image result for nsf research traineeship