Real-Time In-Flight Aircraft State Estimation

OUSD (R&E) CRITICAL TECHNOLOGY AREA(S): Advanced Computing and Software; Integrated Network Systems-of-Systems; Trusted AI and Autonomy OBJECTIVE: Develop a method that utilizes existing aircraft sensors to estimate an aircraft's weight, center of gravity location, airspeed, wind speed, and/or other flight critical aircraft state. DESCRIPTION: Aircraft are often heavily dependent on key state information, which require redundant sensors to meet flight safety standards or mission requirements. In the case of a failure with a dual system, it is often difficult to determine which sensor is the faulty one. In addition, aircraft weight is often required to be entered by the aircrew, which limits its usability in the vehicle management system (VMS) due to potential inaccuracy. The Navy requires the ability to utilize additional existing sensors to estimate aircraft states in real time while in-flight, which could lower the number of redundant sensors, lower the likelihood of mission abort, and/or increase pilot situational awareness. The proposer should validate the estimation methodology using simulation or flight test data; and determine the level of accuracy of the estimations. Some of the parameters that would be targeted to estimate are (but are not limited to) on-ground/in-air transitions, airspeed/ground speed, center of gravity location, aircraft gross weight, aircraft with or without an external load, and engine status/performance/approaching a failure. PHASE I: Determine the technical feasibility of using sensor fusion to create a real-time, in-flight estimation of key aircraft states for an aircraft. Determine the methodology and which existing aircraft sensors are best suited for providing the estimations. The Phase I effort will include prototype plans to be developed under Phase II. PHASE II: Validate the estimation methodology using simulation or flight test data. Determine the level of accuracy of the estimations. Some of the parameters that would be targeted to estimate are (but are not limited to) on-ground/in-air transitions, airspeed/ground speed, center of gravity location, aircraft gross weight, aircraft with or without an external load, engine status/performance/approaching a failure. PHASE III DUAL USE APPLICATIONS: Final testing would be incorporating the state estimation into a flight control algorithm as a sensor monitor and introducing sensor failures to test if the state estimation methodology is able to correctly identify the failed sensor and provide the proper aircraft state to the flight control algorithm. If successful, the estimation methodology would be implemented into new aircraft sensor voting algorithms and reduce the number of needed sources of data. The ability to utilize existing sensors and reduce the number of additional required sensors to provide accurate, reliable aircraft state information would benefit commercial and military platforms as they share common redundancy requirements. The benefit would be a reduction in system complexity, cost, and weight. With the projected rapid expansion of the electric vertical take-off and landing (eVTOL) and urban air mobility (UAM) market, the current levels of probability of loss of aircraft (PLOA), even for airliners, may not be sufficient when considering the predicted orders-of-magnitude increase in flight hours and the operation near highly populated, urban areas. We will need to find new technologies (like this) to increase safety without the burden of extra layers of redundancy. These small, weight-sensitive aircraft will not be able to handle the weight and space burden associated with operation in highly populated areas. REFERENCES: Bi, N.; Haas, D. and McCool, K. Investigation of in-flight gross weight and cg estimation. 60th Annual Forum Proceedings-American Helicopter Society, 7-10 June 2004, Baltimore, Maryland, pp. 2244-2254. https://vtol.org/store/product/investigation-of-inflight-gross-weight-and-cg-estimation-for-the-v22-aircraft-4091.cfm Jarrell, J. A. Employ sensor fusion techniques for determining aircraft attitude and position information. Thesis, Dissertation, West Virginia University, 2016. https://search.worldcat.org/title/1158304207 Henderson, I. Physics informed neural networks (PINNs): An intuitive guide. Towards Data Science, October 24, 2022. https://towardsdatascience.com/physics-informed-neural-networks-pinns-an-intuitive-guide-fff138069563 Yasar, K. Neural network. TechTarget. https://www.techtarget.com/searchenterpriseai/definition/neural-network KEYWORDS: Artificial Intelligence/Machine Learning; Sensor Fusion; Flight Critical; State Estimation; Redundancy; Vehicle Management Systems; Model-Based Systems Engineering