James F. Whidborne

A three-term backpropagation algorithm

Abstract The standard backpropagation algorithm for training artificial neural networks utilizes two terms, a learning rate and a momentum factor. The major limitations of this algorithm are the existence of temporary, local minima resulting from the saturation behaviour of the activation function, and the slow rates of convergence. In this paper, the addition of an extra term, a proportional factor, is proposed in order to speed-up the weight adjusting process. This new three-term backpropagation algorithm is tested on three example problems and the convergence behaviour of the three-term and the standard two-term backpropagation algorithm are compared. The results show that the proposed algorithm generally out-performs the conventional algorithm in terms of convergence speed and the ability to escape from local minima.

Electronic Tuning of Misaligned Coils in Wireless Power Transfer Systems

The misalignment and displacement of inductively coupled coils in a wireless power transfer system (WPT) can degrade the power efficiency and limit the amount of power that can be transferred. Coil misalignment leads the primary coil driver to operate in an untuned state which causes nonoptimum switching operation and results in an increase in switching losses. This paper presents a novel method to electronically tune a Class-E inverter used as a primary coil driver in an inductive WPT system to minimize the detrimental effects of misalignment between the inductively coupled coils which may occur during operation. The tuning method uses current-controlled inductors (saturable reactors) and a variable switching frequency to achieve optimum switching conditions regardless of the misalignment. Mathematical analysis is performed on a Class-E inverter based on an improved model of a resonant inductive link. Experimental results are presented to confirm the analysis approach and the suitability of the proposed tuning method.

Tuning Class E Inverters Applied in Inductive Links Using Saturable Reactors

This paper investigates the performance of Class E inverters used in wireless power transfer applications based on resonant inductive coupling. The variations in the load and the distance between the coils cause Class E inverters to operate under nonoptimal switching conditions, which result in inefficient operation and can lead to permanent damage to its switching transistor. Therefore, a novel approach to tune Class E inverters electronically is proposed. The tuning method relies on saturable reactors to ensure that the inverter operates under optimal switching conditions regardless of variations in the load and the distance between the coils. In addition, a more accurate model of inductive links is presented in order to provide a better understanding of the major power losses in resonant inductive links. Experimental results are presented to confirm the improved accuracy of the inductive link model and the validity of the tuning method.

Direct Method Based Control System for an Autonomous Quadrotor

This paper proposes a real time control algorithm for autonomous operation of a quadrotor unmanned air vehicle. The quadrotor is a small agile vehicle, which as well as being a excellent test bed for advanced control techniques could also be suitable for internal surveillance, search and rescue and remote inspection. The proposed control scheme incorporates two key aspects of autonomy; trajectory planning and trajectory following. Using the differentially-flat dynamics property of the system, the trajectory optimization is posed as a non-linear constrained optimization within the output space in the virtual domain, not explicitly related to the time domain. A suitable parameterization using a virtual argument as opposed to time is applied, which ensures initial and terminal constraint satisfaction. The speed profile is optimized independently, followed by the mapping to the time domain achieved using a speed factor. Trajectory following is achieved with a standard multi-variable control technique and a digital switch is used to re-optimize the reference trajectory in the event of infeasibility or mission change. The paper includes simulations using a full dynamic model of the quadrotor demonstrating the suitability of the proposed control scheme.

Robust control of an unknown plant—the IFAC 93 benchmark

Many control design specifications, such as the IFAC 1993 benchmark, include stringent closed-loop performance requirements, which are required to be met in the face of uncertainties. Such problems can be effectively designed by a combination of the method of inequalities, which designs for explicit closed-loop performance, and analytical optimization techniques, from which robustness is obtained. This paper describes how to combine the method of inequalities with a variety of analytical optimization methods. An introduction to μ synthesis is also included. The design of robust controllers for the benchmark problem using these methods is described, and a simple adaptive gain scheme to improve the performance is suggested.

Reduction of controller fragility by pole sensitivity minimization

Presents a method for the reduction of controller fragility. The method is based on the sensitivity of closed-loop poles to perturbations in the controller parameters. By means of a state space parameterization of the controller, the closed-loop pole sensitivity can be reduced. A controller fragility measure based on the closed-loop pole sensitivity is proposed. Conditions for the optimal state-space realization of the controller are presented, along with a numerical method for obtaining the solution.

Instantaneous friction components model for transient engine operation

Abstract An engine friction model is developed in order to determine the instantaneous friction components at any crank angle during the overall engine response. In order that it can be meaningfully incorporated in overall engine control models, the engine friction model should represent the relevant trends but be relatively simple to compute. The main friction components are the piston assembly, the bearing, the valve train and the auxiliaries. The model includes new analytically derived equations for the friction components of the ring assembly, the bearing with mixed lubrication and the valve train. Many factors contribute to the successful starting of diesel engines and one of them is the effect of engine friction due to high oil viscosity (cold start) on engine startability. The model takes into consideration the effect of temperature variations on the viscosity of the oil. The friction equations are based on theoretical calculations for hydrodynamic and mixed lubrication (where the oil film has collapsed). They rely on Reynolds equation and dynamic analysis. Simulation results are presented, which compared with experimental data indicate an accuracy of more than 97 per cent.

Wireless Power Transfer Using Class E Inverter With Saturable DC-Feed Inductor

Resonant converters used as coil drivers in inductive links generally operate efficiently at optimum switching conditions for constant load values and ranges. Changes in load and range can shift the operation of the coil driver to a nonoptimum switching state which results in higher switching losses and reduced output power levels. This paper presents a method to adapt to variations in range for a Class E inverter used as a coil driver in a wireless power transfer (WPT) system based on resonant inductive coupling. It is shown that by controlling the duty cycle of the inverters switch and the value of the DC-feed inductance, the Class E inverter can be tuned to operate at optimum switching conditions as the distance between the resonant coils of the WPT system changes. Mathematical analysis is presented based on a linear piecewise state-space representation of the inverter and the resonant inductive link. Extensive experimental results are presented to verify the performed analysis and validity of the proposed tuning procedure.

A Unifying Framework for Finite Wordlength Realizations

A general framework for the analysis of the finite wordlength (FWL) effects of linear time-invariant digital filter implementations is proposed. By means of a special implicit system description, all realization forms can be described. An algebraic characterization of the equivalent classes is provided, which enables a search for realizations that minimize the FWL effects to be made. Two suitable FWL coefficient sensitivity measures are proposed for use within the framework, these being a transfer function sensitivity measure and a pole sensitivity measure. An illustrative example is presented.

Detailed analytical model of a single-cylinder diesel engine in the crank angle domain

Abstract A detailed analytical non-linear dynamic model for single-cylinder diesel engines is developed. The model describes the dynamic behaviour between fuelling and engine speed and includes models of the non-linear engine and dynamometer dynamics, the instantaneous friction terms and the engine thermodynamics. The model operates in the crank angle domain. The dynamometer model enables the study of the engine behaviour under loading. The instantaneous friction model takes into consideration the viscosity variations with temperature. Inertia variations with piston pin offset are presented. In-cycle calculations are performed at each crank angle, and the correct crank angles of ignition, speed variations, fuel supply and air as well as fuel burning rate are predicted. The model treats the cylinder strokes and the manifolds as thermodynamic control volumes by using the filling and emptying method. The model is validated using experimentally measured cylinder pressure and engine instantaneous speeds, under transient operating conditions, and gives good agreement. The model can be used as an engine simulator to aid diesel engines control system design and fault diagnostics.