Towards intelligent middleware for more resilient robots
Integrated modular avionics is the primary system architecture used in modern aircraft. It is designed upon a strict principle of fixed and static allocation of redundant tasks on redundant hardware for safety assurance. However, in the event of hardware or communication failures, the safety can be improved if the reallocation of tasks to another hardware is allowed, leading to a concept of reconfigurable integrated modular avionics. One of the challenges for the reconfiguration problem is the way to ensure the feasibility of the new configuration and the continuity of service during the in-flight reconfiguration. This presentation proposes a method to tackle this issue based on formal mathematical optimization that maximizes the number of safety-critical tasks executing on the platform. To illustrate the proposed method, a number of examples have been done on different platforms, including a hardware emulator of multi-core avionics, a software-in-the-loop simulation, and hardware implementation of a modular flying machine.
Thanakorn Khamvilai received his Ph.D. from Georgia Institute of Technology in 2021. He is currently an assistant research professor at the Department of Aerospace Engineering, Pennsylvania State University. His research interests include unmanned aerial vehicle, avionics, control, and optimization.