Abstract:
A method of sensing the air/fuel ratio in a combustion chamber of an internal combustion engine that may be easily implemented by a respective low-cost device is disclosed. The device includes a pressure sensor and a learning machine that generates a sensing signal representing the air/fuel ratio by processing the waveform of the pressure in at least a cylinder of the engine. In practice, the learning machine extracts characteristic parameters of the waveform of the pressure and in function of a certain number of them generates the sensing signal.
Abstract:
An efficient method of determining the three duty-cycles of the three PWM control signals of an inverter that drives a three-phase load in SVM mode and that requires a reduced number of steps advantageously exploits the fact that the three duty-cycles are modulated, in function of the angle γ of the phasor that represents the triplet of voltages to be applied, according to modulating signals that may be reconstructed from only two waveforms using reflection and reversal operations. By storing these two waveforms in two look-up tables, it is possible to reconstruct the modulating signals for each value of the angle γ, with a single memory access and eventually an addition.
Abstract:
A method for estimating data of the state of a system, as well as the EKF filters, allows to use nonlinear mathematical models that describe the systems, but ensures a more accurate precision. This result is due to the fact that the method is based on the EKF technique, but it differs from the EKF technique for adapting it to work with a filter that implements a first degree Stirling approximation formula. This method may be used for estimating position and speed of a brushless motor and may be implemented in a relative device. Such a device may be used for controlling a brushless motor, or it may be introduced in the control loop of a brushless motor of a power steering system, in order to make the driver feel an antagonist torque on the steering wheel determined, according to a pre-established waveform, in function of the speed of the vehicle and of the steering angle. The disclosed methods may even be implemented via software by a program executed by a computer or a microprocessor.
Abstract:
A probabilistic neural network, comprising a hidden layer of neurons, each computing respective membership matrix elements for an input vector of the neural network according to a respective radial basis function defined by a respective spread factor and according to the distance of the input vector from a respective constant vector, wherein said hidden layer comprises at least two neurons having different spread factors (S). A method of training the novel probabilistic neural network is also disclosed.
Abstract:
Variations of position of the rotor of a motor having a plurality of uniformly spaced sensors that generate a position pulse at every rotation by a pre-established angle of the rotor, are estimated by carrying out the steps of: generating a first clock signal at a first frequency; generating a second clock signal at a second frequency multiple than the first frequency by a first factor; in each interval between two consecutive position pulses, - counting the number of pulses of the first clock signal comprised in the interval; - counting the number of pulses of the second clock signal comprised in the interval; - each time the number of counted pulses of the second clock reaches the number of pulses of the first clock counted in an interval that precedes the currently considered interval, signaling that the position of the rotor has varied of an angle equal to the ratio between the pre-established angle and the factor.
This method is implemented in a respective fully digital device that is particular simple to realize.