Shashikanth Suryanarayanan

Design of simultaneously stabilizing controllers and its application to fault-tolerant lane-keeping controller design for automated vehicles

Simultaneous stabilization deals with the following question: given a finite number of LTI plants P/sub 1/,P/sub 2/,...P/sub k/ does there exist a single LTI controller C such that each of the feedback interconnections (P/sub i/,C) (i=1,2,...,k) is internally stable? This paper presents a new methodology for the design of simultaneously stabilizing controllers for two or more plants that satisfy a sufficient condition. A classic result from simultaneous-stability theory is invoked to cast the sufficient condition as a linear matrix inequality (LMI). It is shown that in this setting, the problem of design of simultaneously stabilizing controllers can be reduced to that of a standard H/sub /spl infin// control problem. The technique developed is applied to the design of a fault-tolerant controller for lane-keeping control of automated vehicles. The controller makes the system insensitive to a failure in either one of two lateral error measuring sensors used for lane-keeping control. Experimental results confirm the efficacy of the design and reinforce analytical predictions of performance.

Appropriate Sensor Placement for Fault-Tolerant Lane-Keeping Control of Automated Vehicles

This paper deals with the sensor placement problem in the realization of failure-tolerant lane-keeping control of front-wheel-steered automated vehicles. The scenario considered is one in which lane-keeping action is performed using two sensors that independently measure the lateral deviation of their locations of installation from a reference. The problem of interest is to determine appropriate locations for their installation so that the vehicle can be steered safely even in the event of failure of one of the two sensors. It is shown that for safe lane-keeping action at low as well as high speeds, both sensors should be placed ahead of the rear axle of the vehicle. In addition, the paper discusses a pedagogical problem - namely, the lane-keeping control problem with lateral error information from a sensor placed behind the rear axle. It is shown that, contrary to intuition, it is easier to steer the vehicle at higher speeds. Results based on experiments conducted on vehicles used in the Partners for Advanced Transit on Highways (PATH) program demonstrate the validity of analytical predictions.

Design of Luenberger state observers using fixed-structure H/sub /spl infin// optimization and its application to fault detection in lane-keeping control of automated vehicles

Lane-keeping control forms an integral part of fully automated intelligent vehicle highway systems (IVHS) and its reliable operation is critical to the operation of an automated highway. We present the design of a fault detection filter for the lane-keeping control systems onboard vehicles used by California-PATH, USA in its automated highways program. We use a Luenberger structure for the fault detection filters and tune the observer gains based on an H/sub /spl infin//-based cost. Such a choice of cost was motivated by the need to explicitly incorporate frequency-domain-based performance objectives. The linear matrix inequality (LMI)-based formulation of an H/sub /spl infin// optimization problem of Luenberger state observers does not allow for the augmentation with dynamic performance weightings in the optimization objective, since it makes the problem a nonconvex optimization problem. We present an algorithm to locally solve the problem of the design of Luenberger state observers using H/sub /spl infin// optimization by transforming the problem into an H/sub /spl infin// static output feedback controller problem. Experimental results demonstrate the efficacy of the tuning methodology by comparing the fault detection performance of filters that use H/sub /spl infin// Luenberger observers versus those that use Kalman filters. Implementation issues of the observers are also discussed.

Towards Pitch-Scheduled Drive Train Damping in Variable-Speed, Horizontal-Axis Large Wind Turbines

Given the prohibitive costs of replacement of damaged gearboxes, damping drive train osillations is of immense significance in large wind turbine control design. The Drive Train Damper (DTD) is an important constituent of the power production control routine in variable-speed, pitch-regulated large wind turbines. The DTD is used to mitigate fatigue loading of drive train components. This paper motivates the need for scheduling the parameters of the damper based on blade pitch angle. It is shown that due to the strong coupling between blade edgewise motion and the drive train torsion, the fundamental frequency associated with the drive trains torsional motion can vary significantly across the range of blade pitch angles observed in practice. If left unaccommodated, this variation in fundamental frequency is shown to adversely affect the performance of the DTD.

On the dynamics of the pitch control loop in horizontal-axis large wind turbines

Most commercial large wind turbines use blade pitch action to mitigate structural loads in high wind velocity conditions. In this paper, we study the linearized dynamics of the map from blade pitch to tower top fore-aft deflection in horizontal-axis wind turbines. We show that the mass and stiffness distribution of the blades at certain operating conditions determine the presence (or absence) of a right half-plane zero in the transfer function of interest. We conclude that blade design can impose constraints on the achievable tower fore-aft oscillation mitigation through collective blade pitch control. Consequences of this result to wind turbine design are discussed.

Fault Tolerant Lateral Control of Automated Vehicles Based on Simultaneous Stabilization

Abstract This paper proposes the design of a fault tolerant controller for lateral control of automated vehicles based on simultaneous stabilization. The scenario of loss of information from either one of two sensors critical to the lane-keeping operation of the vehicle is considered. Then, the problem of simultaneously stabilizing both the non-faulty and faulty plants is analyzed. The class of simultaneously stabilizing controllers is introduced and the condition for simultaneous stability is formulated as an LMI constraint for a particular chosen representation of the system which lends itself naturally to the design of the fault tolerant controller. Simulation results illustrating satisfactory performance are included.

A Procedure for the Development of Control-Oriented Linear Models for Horizontal-Axis Large Wind Turbines

In this work we describe a methodology to construct control-oriented, multi-input, multi-output linear models representing the dynamics of variable-speed, pitch-controlled horizontal-axis wind turbines (HAWT). The turbine is treated as an interconnection of mechanical elements with distributed mass, damping, and stiffness characteristics. The behavior of the structural components of the turbine is approximated as that of their dominant modes and the wind-blade aerodynamic interaction is modeled using the Blade Element Momentum (BEM) theory. The modeling procedure explicitly exploits the horizontal-axis configuration and constraints imposed thereof. The models developed using the outlined procedure are parametrized based on a handful of parameters that are often used to specify mass/stiffness distributions and geometry. The predictions of the linear models so constructed are validated against that of an established nonlinear model. The use of the modeling procedure in addressing problems of immediate interest to the wind turbine industry is presented.

System dynamics and control of bicycles at high speeds

This paper presents the system dynamics and automated roll-rate control of front and rear-wheel steered bicycles. Automated steering control of bicycles gains importance in the context of a recent effort, initiated by bicycle designer Matt Weaver, to develop controllers to steer bicycles at very high speeds (70-100 mph). This paper extends earlier work on rear-wheel steered bikes, importantly Kleins unridable bicycle. Controllers for both front and rear-wheel steered bicycles are designed based on classical control techniques. Simulation results demonstrate good robustness and disturbance rejection properties. Implementation is currently underway.

H/sub /spl infin// optimization of Luenberger state observers and its application to fault detection filter design

This paper considers H/sub /spl infin// optimization of Luenberger state observers. The conventional formulation of H/sub /spl infin//-optimal state observers does not allow the augmentation of dynamic performance weightings in the optimization objective, since it makes the problem a nonconvex optimization problem. We propose an algorithm to locally solve an H/sub /spl infin// optimization problem of Luenberger state observers by transforming the problem into an H/sub /spl infin// optimization problem of a static output feedback controller. The proposed approach offers an intuitive and efficient way to explicitly design the estimation error dynamics of the observer in the frequency domain. As an application example, the proposed approach is applied to the tuning of fault detection filters for lateral control of automated passenger vehicles. Numerical simulations are conducted to show the effectiveness of the proposed tuning method.