Knocking detection system and knocking detection method of internal combustion engine
AssigneeToyota Jidosha Kabushiki Kaisha
InventorsMasahiro Mohri, Akinori Kitazume, Daisuke Kobayashi, Kenji Senda
Life of this patent
Estimating the term…
Abstract
A knocking sensor detecting vibration of an engine body and a pressure sensor detecting a pressure of a combustion chamber are provided. A value representing the knocking strength is acquired from output values of the pressure sensor. Weights of a neural network are learned using a value representing the vibration of the engine body detected by the knocking sensor as an input value of the neural network and using the acquired value representing the knocking strength as training data. The value representing the knocking strength is estimated from the output values of the knocking sensor by using the learned neural network.
Claims12
A knocking detection system of an internal combustion engine comprising:
- a knocking sensor detecting vibration of an engine body,
- a pressure sensor detecting a pressure inside a combustion chamber of the internal combustion engine,
- a learning device for using a neural network to estimate a value representing a knocking strength of the internal combustion engine, and
- a controller configured to control ignition of the internal combustion engine based on the value of the knocking strength of the internal combustion engine, wherein
- the value representing the knocking strength is acquired from output values of the pressure sensor,
- weights of the neural network are learned by the learning device by using the value representing vibration of the engine body detected by the knocking sensor as an input value of the neural network and using the acquired value representing the knocking strength as training data, and
- the value representing the knocking strength is estimated from the output values of the knocking sensor by using the learned neural network.
A knocking detection method of an internal combustion engine using a neural network to estimate a knocking strength of the internal combustion engine based output values of a knocking sensor detecting vibration of an engine body and output values of a pressure sensor detecting a pressure of a combustion chamber of the internal combustion engine, wherein said knocking detection method comprises:
- finding a value representing the knocking strength from the output values of the pressure sensor,
- preparing a training data set showing a relationship between the value representing vibration of the engine body detected by the knocking sensor and the value representing the knocking strength,
- learning weights of the neural network using the value representing vibration of the engine body detected by the knocking sensor as an input value of the neural network and using the acquired value representing the knocking strength as training data,
- using the learned neural network to estimate the value representing the knocking strength from the output values of the knocking sensor, and
- controlling ignition of the internal combustion engine based on the knocking strength.
A non-transitory computer readable storage medium sorting a computer program, that when executed by a processor, implements a knocking detection method of an internal combustion engine using a neural network to estimate a knocking strength of the internal combustion engine based output values of a knocking sensor detecting vibration of an engine body and output values of a pressure sensor detecting a pressure of a combustion chamber of the internal combustion engine comprising:
- finding a value representing the knocking strength from the output values of the pressure sensor,
- preparing a training data set showing a relationship between the value representing vibration of the engine body detected by the knocking sensor and the value representing the knocking strength,
- learning weights of the neural network using the value representing vibration of the engine body detected by the knocking sensor as an input value of the neural network and using the acquired value representing the knocking strength as training data,
- using the learned neural network to estimate the value representing the knocking strength from the output values of the knocking sensor, and
- controlling ignition of the internal combustion engine based on the knocking strength.
Dependencies are inferred from the claim text.