Contributions to Navigation Under Unknown Input and Cyber-Physical Security

Overview & Synopsis

This thesis contributes to state estimation on the special orthogonal group SO(3) under unknown input and to cyber-physical systems (CPS) security. Two challenges related to unknown input are addressed. The first is attitude estimation without relying on the gyroscope, due to its drawbacks such as high-power consumption. The second is the separation of Earth gravity and external acceleration from accelerometer measurements, as the former is needed for attitude estimation and the latter for position and velocity. Additionally, two security problems related to navigation applications are addressed: secure attitude estimation on SO(3) under randomly occurring false data injection attacks, and security analysis of the ground vehicle lateral model under zero dynamics attacks. Finally, the thesis proposes an active defense strategy called “misleading unauthorized observers”, which targets attackers attempting unauthorized remote state estimation. This strategy introduces a new direction in CPS security based on active rather than traditional passive defenses.