Jacob Nixon ’22 has been interested in biomechanics since he was in middle school – namely the coordinated motion and physical interaction of different parts of the body, such as muscles and bones... Read more about Jacob Nixon '22 - studying the relationship between the gut microbiome and bone strength
Rachel Zheng '21 - Building Search-and-Rescue Robotic Mission Simulator
Last year, Rachel Zheng ’21, mechanical engineering major in the Sibley School of Mechanical and Aerospace Engineering, worked with the Autonomous Systems Lab to build a search-and-rescue robotic mission simulator.
Led by Mark Campbell, John A. Mellowes '60 Professor in Mechanical Engineering, the research was recently published in IEEE Robotics and Autonomation Letters.
Zheng used Gazebo, a robot simulation tool, to simulate actual robot dynamics, which adds to the possibility of mission failure due to navigation. A well-designed simulator makes it possible for researchers to rapidly test algorithms, design robots, perform regression testing, and train AI system using realistic scenarios.
The simulator includes two ground vehicles and one drone and has the ability to use simultaneous localization and mapping (SLAM) to map any unknown environment. SLAM is a method used for autonomous vehicles that allows researchers to build a map and localize the vehicle in that map at the same time. This information allows for tasks such as autonomous navigation.
Using a search path outputted from a path planner, the robot can then autonomously navigate to the search waypoints while actively avoiding obstacles and high-risk areas within the environment. Zheng and the Autonomous Systems Lab then used the simulator to run experiments with a risk-aware multi-robot planner, which takes into consideration environment danger levels to protect the search and rescue robots.
“Our results suggest that a risk-aware planner can reduce losses on one set of agents, without a significant loss in performance, even when dealing with practical navigation challenges and imperfect scene knowledge,” says Zheng.
This semester, she is working on implementing landmark-based SLAM for mapping large outdoor environments. Large environments are often hard to map accurately because they are sparse with few features to localize from. By using tree detections from a stereo camera as landmarks, Zheng hopes to be able to build more accurate maps than other SLAM methods such as loop closure.
The mentorship from graduate students in the Autonomous Systems Lab proved to be immensely helpful for Zheng, even inspiring her to pursue graduate school after graduation.
''Working with Rachel has been a pleasure. She is diligent and reliable in her work, and she has become a valued member of our team,” says Beatriz Asfora, Ph.D. candidate in the Autonomous Systems Lab. “Understanding how things work rather than just ‘making them work’ is crucial in research, and Rachel is fortunately very curious.''
In addition to conducting research, Zheng has had many opportunities for collaboration during her time in the Sibley School. She was previously the subteam lead on Cornell Hyperloop. The team is working to revolutionize transportation by developing a Hyperloop pod from conceptualization and are aiming to participate in SpaceX’s annual Hyperloop competition.
The summer after her junior year, Zheng interned at the NASA Ames Research Center where she created a Matlab tool to depict the physical response of an aeroelastic fixed-wing aircraft with a novel Variable Camber Continuous Trailing Edge Flap (VCCTEF) system. “The tool helped research engineers better visualize aircraft’s response to flutter, maneuver loads, and gusts, improving the design and tuning of control strategies,” says Zheng. Zheng has interned at Uber ATG where she worked on thermal camera calibration for the next-generation autonomous vehicle.
After graduation, Zheng will be pursuing an MS in Robotic Systems at Carnegie Mellon University and then plans to work in the robotics industry. Before starting her MS, Zheng will spend the summer interning at Amazon Robotics.