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:
The amount of fuel to be injected in each cylinder of a multi-cylinder spark ignition internal combustion engine may be determined with enhanced precision if the fuel injection durations are determined in function of the sensed mass air flow in all the cylinders of the engine, instead of considering only the air flow in the same cylinder. This finding has led to the realization of a more efficient method of controlling a multi-cylinder spark ignition internal combustion engine and an innovative feedforward control system.
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.