Current projects
Predicting Human Musculoskeletal Health on the MoonNational Aeronautics and Space Administration, PI, with co-PI Steven Elmer (St. Catherine University). This project will investigate human locomotion on lunar surfaces by developing a musculoskeletal model and reinforcement learning algorithm, and designing and fabricating a hypogravity simulator to validate the theoretical findings. As the lunar environment can also affect human musculoskeletal health, this project will also assess such impact and design a suitable exercise countermeasure to optimize performance and reduce risk for injuries. It will help ensure the well-being and functionality of astronauts during space exploration missions and guide the design of rehabilitation programs for astronauts after returning to the earth. |
Using Symmetric Neural Networks to Calculate Fractional-Order DerivativesOak Ridge Associated Universities (ORAU), PI The application of fractional order dynamics spans several fields, including control systems (fractional-order PID), biomedical engineering (drug absorption modeling, tissue deformation), and fluid mechanics (modeling of non-Newtonian fluid behaviours). While the concept is useful and straightforward, for an engineer first dealing with a fractional-order system, the learning curve is a bit steep and may prevent immediate access to the utility of fractional-order modeling. This research is inspired by acknowledging the connection between symmetric networks and derivative calculation to train a complicated fractional-order neural network model FractionalNet. The training data is not based on fractional-order derivatives but on integer-order derivatives, which are much easier to obtain. |
Towards Robust and Efficient Bipedal Robot Locomotion on the Moon through Reinforcement LearningNational Science Foundation, PI Beyond Earth, planetary exploration is still heavily reliant on wheeled rovers, which show limited mobility when it comes to difficult terrain and inclined regions, for instance, featuring rocks, craters, and narrow crevices. Bipedal robots hold promise for expanding the scope of tasks on the Moon due to their unique locomotion capabilities. However, the challenges of lunar gravity and the presence of fine surface’s dust pose stability challenges for these robots. This research seeks to gain a fundamental understanding of lunar bipedal locomotion mechanics and design robust and energy-efficient controllers for bipedal robot locomotion on the Moon by using reinforcement learning. In addition to advancing novel knowledge and technology, this research will contribute to building and improving the robotics curriculum at Michigan Technological University while engaging undergraduate and graduate students, as well as high school students, in robotics education and research. |
Completed projects
Biped Locomotion on Slippery SurfacesResearch Excellence Fund - Research Seed Grant, Michigan Technological University, PI Recent advances in robotics have enabled bipedal robots to walk on challenging terrains such as stairs and uneven surfaces. However, locomotion on slippery terrain, such as snow and ice, remains a significant challenge, limiting the real-world deployment of bipedal robots, particularly in snowy regions. This seed grant aims to establish preliminary results for bipedal locomotion on slippery surfaces. The project will focus on the development of a bipedal robotic platform, the use of machine learning to better understand robotic foot–surface interaction dynamics on slippery terrain, and the design of state-of-the-art controllers for legged locomotion under such conditions. Output
|