Shawn Keshmiri

Robust and Adaptive Nonlinear Model Predictive Controller for Unsteady and Highly Nonlinear Unmanned Aircraft

The nonlinear and unsteady nature of aircraft aerodynamics in the presence of adverse conditions and external disturbances, together with a limited range of flight variables makes the use of the linear control theory inadequate in such conditions. To address these constraints and significantly enhance aircraft control capabilities, this brief presents an adaptive framework for a robust nonlinear model predictive control (NMPC). Control algorithms are tested on a 1100-pound unmanned aerial system, with nonlinear, coupled, and unstable open-loop dynamics, subjected to environmental disturbances and measurement noise. Given the usual frequency content exclusion between disturbance and noise, this solution addresses the lack of robustness in model predictive control by the inclusion of frequency-dependent weighting matrices and a nonlinear version of the mixed sensitivity approach. Furthermore, real-time aerodynamic parameter estimation and predictive model updating is carried out by online adaptive artificial neural networks. Through assessment and validity of control algorithms, it is demonstrated that two originally competing control concepts, robustness and performance, are integrally attained in real time. This is usually unreachable in the classical NMPC framework for complex systems.

H-Infinity gain scheduling design for the meridian UAS for a broader range of operation and for fault tolerant applications

This paper describes the design of a MIMO H-infinity gain schedule for the Meridian UAS and illustrates its application for an extended operational range and for a moderate fault tolerant capability. The gain scheduling concept is based on smooth interpolation between two H-infinity controllers, covering a broad range of trim conditions. This interpolation is provided by the homogeneous design of the controllers by using the same augmented model structure within the H-infinity algorithm. The scheduling variable is defined as the commanded airspeed, acting as an exogenous variable, driving a smooth transition between the current and the next controller. The utilization of an exogenous signal avoids the occurrence of hidden coupling effects, typical in gain scheduling applications. Also, a new procedure is defined to adapt the controller when a failure or a significant dynamic change in the UAS takes place. Using the best knowledge of the failure the H-infinity controller is calculated. This controller is computed with a predefined gamma value, avoiding iterations, generating the new controller instantaneously.

A Simplex Architecture for Intelligent and Safe Unmanned Aerial Vehicles

Unmanned Aerial Vehicles (UAVs) are increasingly demanded in civil, military and research purposes. However, they also possess serious threats to the society because faults in UAVs can lead to physical damage or even loss of life. While increasing their intelligence, for example, adding vision-based sense-and-avoid capability, has a potential to reduce the safety threats, increased software complexity and the need for higher computing performance create additional challenges -- software bugs and transient hardware faults -- that must be addressed to realize intelligent and safe UAV systems. In this paper, we present a fault tolerant system design for UAVs. Our proposal is to use two heterogeneous hardware and software platforms with distinct reliability and performance characteristics: High-Assurance (HA) and High-Performance (HP) platforms. The HA platform focuses on simplicity and verifiability in software and uses a simple and transient fault tolerant processor, while the HP platform focuses on intelligence and functionality in software and uses a complex and highperformance processor. During the normal operation, the HP platform is responsible for controlling the UAV. However, if it fails due to transient hardware faults or software bugs, the HA platform will take over until the HP platform recovers. We have implemented the proposed design on an actual UAV using a low-cost Arduino and a high-performance Tegra TK1 multicore platform. Our case-studies show that our design can improve safety without compromising performance and intelligence of the UAV.

Nonlinear aerodynamics of an unmanned aircraft in wind shear

Purpose The purpose of this paper is to evaluate the accuracy of linear and quasi-steady aerodynamic models of aircraft aerodynamic models when a small unmanned aerial system flies in the presence of strong wind and gust at a high angle of attack and a high sideslip angle. Design/methodology/approach Compatibility analysis were done to improve the quality of recorded flight test data. A robust method called fuzzy logic modeling is used to set up the aerodynamic models. The reduced frequency is used to represent the unsteadiness of the flow field according to Theodorsen’s theory. The work done by the aerodynamic moments on the motions is used as the criteria of stability. Findings In portions of flight, aircraft’s stability and control derivatives were unstable and nonlinear functions of airflow angles and angular rates. The roll angle had an important effect on unsteadiness of directional oscillatory damping derivatives. The pilot-induced oscillation and wing rock possibilities were investigated and dismissed so that the lateral directional oscillatory motion was classified as a nonlinear Dutch roll oscillation. Major modeling enhancements or real-time parameter identification are required for the control of a small unmanned aerial system in off-nominal conditions. The robustness tests of all-weather autopilot systems must be done with consideration of sign change. Originality/value Oscillatory damping derivatives were reconstructed using flight test data and the inadequacy of engineering level software in predicting this type of instability observed and demonstrated for a flight in the presence of wind shear and external disturbances.

Guidance of Multi-Agent Fixed-Wing Aircraft Using a Moving Mesh Method

This paper presents a novel guidance logic for multi-agent fixed-wing unmanned aerial systems using a moving mesh method. The moving mesh method is originally designed for use in the adaptive numerical solution of partial differential equations, where a high proportion of mesh points are placed in the regions of large solution variations and few points in the rest of the domain. In this work, the positions of the aircraft are considered as mesh nodes connected to form a triangular mesh in two spatial dimensions. The outer aircraft positions are planned with the reference point algorithm. This logic provides the outer agents moving point positions that are relative to a virtual point position with the desired heading angle and velocity. The inner agents, or interior mesh nodes, are moved with a moving mesh technique to keep the whole mesh as uniform as possible. The moving mesh technique has built-in mechanisms to keep the mesh as uniform as possible and prevent nodes from crossing over or tangling. This property can be seen as an automatic internal collision avoidance mechanism. It also has explicit formulas for nodal velocities, making the technique easy to implement on computer. The mesh nodes are replaced by unmanned aerial systems with nonlinear six degrees of freedom dynamics. The centralized moving mesh guidance is complimented by a decentralized nonlinear predictive controller to control each aircraft. To validate flexibility and coherency of agents and formation, the moving point concept is used in the simulation to follow an arbitrary, linear, sinewave-like, or curvature shaped flight segments. Robustness of the algorithm is also verified where agents were affected by external wind.

Guidance of multi-agent fixed-wing aircraft using a moving mesh method

This paper presents implementation of a novel guidance logic for multi-agent fixed-wing unmanned aerial systems using a moving mesh method. The moving mesh method is originally designed for use in the adaptive numerical solution of partial differential equations where a high proportion of mesh points are placed in the regions of large solution variation and few points in the rest of the domain. In this work, the positions of the aircraft are considered as mesh nodes and connected to form a triangular mesh in two spatial dimensions. The outer aircraft, or the boundary nodes of the mesh, are moved with the velocities which are specified using the the relative distance from and the heading angle of a reference point while keeping the formation of the aircraft in a desired shape. The inner agents or the interior mesh nodes, on the other hand, are moved with a moving mesh technique to keep the whole mesh as uniform as possible. The moving mesh technique has built-in mechanisms to keep the mesh as uniform as possible and prevent nodes from crossing over or tangling. This property can be seen as an automatic collision avoidance mechanism. It also has explicit formulas for nodal velocities, which makes the technique easy to implement on computer. The centralized moving mesh guidance was complimented by a decentralized nonlinear predictive controller to control each aircraft. The mesh nodes are replaced by unmanned aerial systems with nonlinear six degrees of freedom dynamics. To increase flexibility of aircraft in the formation, the moving point concept is used to follow arbitrary linear or curvature shaped flight segments.

Nonlinear Parameter Estimation of Unmanned Aerial Vehicle in Wind Shear Using Artificial Neural Networks

Operation of unmanned aerial vehicles (UAVs) for Earth science missions often ensures subjection to hostile environments where external disturbances induce highly nonlinear and unsteady aerodynamic behavior. For autonomous control of an aircraft to be feasible and guarantee stability in such conditions, the physics based model of the vehicle must account for unsteadiness and nonlinearities in the flight envelope where stability and control derivatives are nonlinear and time-varying. Modeling unsteady and nonlinear characteristics of system parameters is a challenging task requiring high fidelity parameter estimation algorithms, such as fuzzy logic modeling, which are computationally expensive and often take weeks or months to converge to a trustworthy output. This work presents a novel system identification method of parameter estimation from real flight test data using artificial neural networks (ANNs) in which the ANNs have the capability of mimicking input-output relationships of existing high fidelity methods. The method is successfully validated using flight test data in high wind shear to estimate select lateral-directional stability derivatives of a small UAV. Other novelties of this work include the use of reduced frequencies to account for unsteadiness and validation and testing on two sets of flight test data, in considerably dissimilar flight conditions, with both radio controlled and autonomous piloting modes considered.

Control of Multi-Agent Collaborative Fixed-Wing UASs in Unstructured Environment

Swarms of unmanned aircraft are the inevitable future of the aerospace industry. In recent years, swarming robots and aircraft have been a subject of much interest; however, many research projects make impractical assumptions such as point mass dynamics with no aerodynamic effects for aircraft models, and most works stop short of fully validating their methods via flight testing. This work presents a proximity based guidance, navigation, and control of multi-agent fixed-wing unmanned aerial systems in an unstructured environment. A scalable swarm navigation method is developed using adaptive moving mesh partial differential equations controlled by the free energy heat flow equation. To emulate the physics-based dynamic characteristics of fixed-wing UASs, mesh nodes are constrained by aircraft six degrees of freedom equations of motion. An optimal control based path planning using virtual points was developed and constrained with aircraft dynamical limitations. Lateral acceleration is used for lateral guidance of aircraft and the aircraft pitch attitude error is used for longitudinal guidance of multi-agent unmanned aerial systems. A decentralized optimal automatic controller was developed to control each system. Using an advanced in-house autopilot system, validation and verification flight tests were successfully conducted using two large unmanned aerial systems with four meter wingspans flying at 35 knots.

Multi-spectral radar measurements of ice and snow using manned and unmanned aircraft

We present an overview of a set of radar instruments developed at the University of Kansas for multi-spectral measurements of ice and snow properties. The systems operate at different frequency bands ranging from 14 MHz to 38 GHz, onboard manned and unmanned aircraft. The data collected with these systems are used to estimate parameters such as ice thickness, ice surface and bedrock topography, snow cover thickness on sea ice, and annual snow accumulation. We give a summary of recent field programs (including operations out of Punta Arenas, Chile) and discuss current collaborations with Chilean institutions.

Radar ECHO sounding of russell glacier at 35 MHz using compact radar systems on small unmanned aerial vehicles

We have developed an unmanned aerial system consisting of a compact sounding radar operating in the frequency bands of 14 and 35 MHz integrated into a fixed-wing UAV for remote surveys of glaciers and ice-sheets. The system is capable of collecting coherent sounding measurements along multiple parallel tracks. With the use of differential GPS for precise trajectory determination, we demonstrate multipass SAR array processing. The system was recently deployed by CReSIS personnel in the spring of 2016 to survey the Russell glacier in Greenland. This paper reports on the instrumentation including the integration of the radar, antennas, and aircraft; the survey flights in Greenland; and results from measurements collected at 35 MHz.