Automatic Fault Diagnosis/Repair

Autonomous Fault Diagnosis/Repair

There has been a recent surge in the development of autonomous vehicles that traverse dangerous or difficult terrains. For these systems to be fully autonomous they need to be fault resilient, especially when an autonomous system is likely will be used for long periods of time with little human supervision. One way to deal with these faults is to detect and repair them automatically while the vehicle is running. Many methods have been developed to try and address this problem.

Model-based methods is one group or popular approaches to this problem. However, these methods can be difficult to implement because the model that it relies upon can be complex and usually requires expert knowledge to create. We are interested in simplifying the process of creating these models to make these methods more usable. This will both increase the resiliency of autonomous systems and enable design teams to troubleshoot errors that occur when designing and building a vehicle.