Rafsha Mazhar

Academic

Career

Verbatim text

Pilot Buddy 
PilotBuddy is a fault-injection and detection framework for real-time anomaly detection in the 
course of a UAV flight. The framework is a web-app linked externally with FlightGear Simulator, 
through socket programming, for better visualization. The framework allows users to connect to 
the simulator and inject aileron, rudder, elevator or engine faults into the flight simulation and 
study its effects on the UAV’s flight. Furthermore, PilotBuddy comes integrated with a model 
trained on the changes in the different parameters of a UAV, before and after fault injection. This 
helps our framework recognize faulty behavior and generate an alert when any anomaly is detected 
in 
a 
real-time 
flight, so 
recovery 
actions 
may 
be 
taken. 
Features include: 
- Real-time connection with FlightGear simulator to directly visualise the effects  
- A Fault Injection Framework to enable users to inject faults in-real-time and study their effects 
- Visualization of parameter changes during the course of a flight in graphical form, for better 
understanding 
- Detection of actuator and engine failures 
- Display fault diagnosis information against injected failures 
 
 
 
 
 
 
 
 
 
 
Technology Used: 
Python, Flask framework,  
Socket Programming, Google Colab, Github 
Supervisor Name: 
Dr. Irum Inayat 
Mr. Shoaib Ashraf (external) 
Group Members:   
Ayesha Ateeq (i17 - 0010) 
Namrah Rasool (i17 - 0018) 
Rafsha Mazhar (i17 - 0028)