My Graduate Research in Urban Air Mobility
 |
|---|
| Image Credits : NASA |
The growth of UAM and the role of autonomy in it#
The world as a collective is transitioning to a more autonomous future with the advent of intelligent transportation vehicles running on renewable energy. The technological advancements run in parallel with the rapid increase in urban population density and the need for quick and reliable point-to-point travel. The UAM concept highlights the use of Urban Air Taxis, that can take off and land vertically from densely populated areas, and as such need to be highly reliable as the UAM market is projected to grow to roughly 28.3 billion USD upto 2030 along with an infrastructure growth of 58.7% (World UAM Market - 2022 : ResearchAndMarkets.com)
With this rapid growth, comes numerous infrastructure scaling challenges, with air traffic control being a primary concern. Introducing verifiable autonomy into this infrastructure is what I focus on.
Current Work#
Our team is focused on the Systems integration of the flight control algorithms and the collision aware path planning, into a Multi-Agent System of Systems level simulation. In this simulation, an MBSE decision making framework is integrated to handle off-nominal scenarios pertaining to both a single aircraft, or a fleet of aircrafts keeping in mind the different stakeholders and actors in the system.
I am currently working with using ROS2 and C++ to build this modeling and simulation pipeline with a SysML model acting as the foundation for decision making. In parallel, I also proposed and am investigating the use of Multi Agent Reinforcement Learning to devise key system metrics for incentivizing safety and reliability for all aircrafts. The use of Multi Agent RL would also allow us to simulate and understand emergent behavior that cannot be otherwise predicted through conventional methods.
The Verification and Validation of this framework would be done by incorporating an Explainable AI module into the entire system to better understand the root cause of erroneous decisions.