Adaptive Human Model In The Presence of Plant Uncertainty 
In this research, an adaptive human model is proposed which mimics the crossover model despite input bandwidth deviations and plant uncertainties. The proposed human pilot model structure is based on the model reference adaptive control, and the adaptive laws are obtained using the Lyapunov-Krasovskii stability criteria applied to the overall closed loop system including the human pilot, considering time delay, and the plant. The proposed model can be employed for human-in-the-loop stability and performance analyses with different controllers and plant types. Model validation is done by comparing the adaptive human model and participants’ data. A statistical analysis, consists of confidence interval calculation, hypothesis test and power analysis, is conducted to measure the predictive power of the proposed model. You can access the participants’ data by clicking here.

Adaptive Control Allocation

In this research, an adaptive control allocation approach for over-actuated systems is developed. The methodology does not utilize the control input matrix estimation to tolerate actuator faults and, therefore, the proposed control allocation method does not require persistence of excitation or additional sensors to determine actuator effectiveness. The proposed adaptive control allocation method has a modular design, allowing the flexibility to develop the outer loop controller and the control allocation strategy separately. 

Modeling Cyber-Physical Human Systems Using Reinforcement Learning and Game Theory

Predicting the outcomes of cyber-physical systems with multiple human interactions is a challenging problem. We address this problem by exploiting a modeling framework where reinforcement learning (RL) and game theory (GT) is used together. In this framework, GT is used to model strategic decision making of humans and RL is used to model time-extended (multi-move) decisions. The most attractive feature of the method is proposing a computationally feasible approach to simultaneously model multiple humans as decision-makers, instead of determining the decision dynamics of the intelligent agent of interest and forcing the others to obey certain kinematic and dynamic constraints imposed by the environment. 

Human-Controller Interactions: Stability and Performance Analysis

Collaborators: Northeastern University, University of South Florida

In this research, we study human-in-the-loop control systems from the points of stability and performance. While providing rigorous mathematical analysis, our aim is to develop a framework for human-controller interaction analysis capable of providing us with complete description of the human-in-the-loop control architectures, regardless of the incorporated human model. This framework has certain applications in numerous fields including, but not limited to, telerobotics, robotics, and flight control.

Pressure Control of Gas Generator for Throttleable Ducted Rockets using Adaptive Control Theory

Sponsor: ROKETSAN A.S.

Throttleability provides a major superiority to rockets against their non-throttleable counterparts. This feature can be utilized at solid rocket motors by the help of changing the throat area of the nozzle at the outlet of the gas generator which generates fuel in gaseous form. 
In order to have a desired variable fuel mass flow rate and thus a desired variable thrust, the gas pressure in the gas generator chamber needs to be precisely controlled. In this research, the goal is to obtain precise control of pressure using the state of the art advanced control system technologies. In the existing literature, there are mostly constant gain or gain-scheduled controllers utilized for the purpose of pressure control. These approaches are not very successful in dealing with the nonlinear terms of the plant, or it is very time consuming to arrange look-up tables for gain scheduling. In this study, we eliminate the need for a precise system model for classical approaches. This is achieved by developing and utilizing a delay resistant adaptive controller which includes a unique combination of the elements of the Adaptive Posicast Controller (APC) with the ability to accommodate large delays and closed loop reference model (CRM) adaptive controller with improved transient response of closed loop adaptive system. Simulation and experimental results verifying the performance improvement of the delay resistant controller over the alternatives are obtained.

Figure taken from:

Developing A Game Theoretical Modeling and Simulation Framework For the Integration of Unmanned Air Vehicles into the National Airspace

Sponsor: The Scientific and Technological Research Council of Turkey (TÜBİTAK): CAREER Award (2014)

The focus of this research is to develop a game theoretical modeling and simulation framework for the
integration of Unmanned Aircraft Systems (UAS) into the National Air Space (NAS). The problem of predicting the outcome of complex scenarios, where UAS and manned air vehicles co-exist in future civil applications of UAS such as agriculture, journalism, commercial and so on, is the research problem of this study. The most challenging part of this research problem is predicting the reactions of human pilots during interactions with UAS and automation. The foundations of the proposed modeling method in this paper is formed by game theory, which analyzes strategic decision making between intelligent agents, bounded rationality concept, which is based on the fact that humans cannot always make perfect decisions, and reinforcement learning, which is shown to be effective in human behavior in psychology literature. These concepts are used to develop a simulator which can be used to obtain the outcomes of scenarios consisting of UAS, manned vehicles, automation and their interactions. 

Predicting the Evolution of Complex Scenarios in Next Generation Airspace

Sponsor: NASA Ames Research Center/UCSC contract

Effective automation is critical in achieving the capacity and safety goals of the Next Generation Air Traffic System. Unfortunately, the interactions between automation and their human counterparts is complex and unpredictable. In this research, we showed how outcomes of complex scenarios that involve human-human interactions in the presence of advanced Next Generation technologies can be predicted by leveraging a game theory based framework. In this framework, human users are not modeled explicitly. Instead, their goals are modeled and through reinforcement learning their actions are predicted. Such a framework allows for efficient trade studies and feasibility testing on a wide range of automation scenarios. We tested this framework on a scenarios where up to 50 aircraft need to self-navigate using Automatic Dependent Surveillance-Broadcast information. In these scenarios, we showed how the framework can be used to predict the ability of pilots to adequately balance aircraft separation and fly efficient paths. We analyzed the scenarios with several levels of complexity and airspace congestion.

Pilot Induced Oscillation Recovery Using Optimal Control Allocation

Sponsor: NASA Ames Research Center/UCSC contract

In this research, we developed a control allocation technique that help pilots recover from pilot induced oscillations (PIO). When actuators are rate-saturated due to aggressive pilot commands, high gain flight control systems or some anomaly in the system, the effective delay in the control loop may increase depending on the nature of the cause. This effective delay increase manifests itself as a phase shift between the commanded and actual system signals and can instigate PIOs. The proposed control allocator reduces the effective time delay by minimizing the phase shift between the commanded and the actual attitude accelerations. This allocator, which we named as CAPIO (Control Allocator to recover from Pilot Induced Oscillations) distinguishes itself as being one of the first PIO recovery system designed for Multi Input Multi Output Systems without the need for ganging the actuators. After simulation studies, CAPIO was successfully tested with real pilots at Vertical Motion Simulator at NASA Ames Research Center. 

Stable Adaptive Control of Automotive Powertrains - Air Fuel Ratio Control

Sponsor: Ford-MIT Alliance

In this research, we developed control solutions for the problem of precisely adjusting the fuel-to-air (FAR) ratio of a spark ignition internal combustion engine, using an adaptive control method of time-delay systems. The objective in FAR control is to maintain the in-cylinder FAR at a prescribed set point, determined primarily by the state of the three-way catalyst, so that the pollutants in the exhaust are removed with the highest efficiency. The FAR controller must also reject disturbances due to canister vapor purge and inaccuracies in air charge estimation and wall-wetting compensation. Two adaptive controller designs were considered. The first design was based on feed forward adaptation while the second design was based on both feedback and feed forward adaptation incorporating the recently developed Adaptive Posicast Controller (APC). Following successful simulation studies, experimental tests were conducted using vehicles provided by Ford Motor Company. Experimental results showed dramatic performance improvements with the proposed control solution over the development controller. 


Stable Adaptive Control of Automotive Powertrains - Idle Speed Control

Sponsor: Ford-MIT Alliance

In this research, we developed control solutions for the problem of precisely adjusting the idle speed of a spark ignition internal combustion engine, using a recently developed Adaptive Posicast Controller (APC) for time-delay systems. The objective was to regulate the engine speed to a prescribed set-point in the presence of accessory load torque disturbances such as those due to air conditioning and power steering. The adaptive controller, integrated with the existing proportional spark controller, was used to drive the electronic throttle actuator. Following successful simulation studies, we conducted experiments using vehicles provided by Ford Motor Company. Experimental results showed dramatic performance improvements with the proposed control solution over the development controller.

Adaptive Control of Time Delay Systems

Sponsor: Ford-MIT Alliance

In this research, we developed an Adaptive Posicast Controller that deals with parametric uncertainties in linear systems with delays. It was assumed that the plant has no right half plane zeros and the delay is known. The adaptive controller was based on the Smith Predictor and Finite Spectrum Assignment with time-varying parameters adjusted online. A novel Lyapunov–Krasovskii functional was used to show semi-global stability of the closed-loop error equations. The controller was applied to automotive control problems. The implementation results showed  that the Adaptive Posicast Controller significantly improves the closed-loop performance when compared to the case with the existing  baseline controller.