Data processing: vehicles – navigation – and relative location – Vehicle control – guidance – operation – or indication – With indicator or control of power plant
Reexamination Certificate
2000-05-10
2001-09-11
Dolinar, Andrew M. (Department: 3747)
Data processing: vehicles, navigation, and relative location
Vehicle control, guidance, operation, or indication
With indicator or control of power plant
C701S106000, C706S023000
Reexamination Certificate
active
06289275
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to a methodology of transient fuel control for an internal combustion engine in a motor vehicle and, more specifically, to a neural network based transient fuel control method for an internal combustion engine in a motor vehicle.
2. Description of the Related Art
A motor vehicle is typically powered by an internal combustion engine which generally operates in two modes, steady state and transient. A steady state mode could be characterized by maintaining the vehicle at a constant engine speed and load, whereby the power demands on the engine do not fluctuate. In contrast, a transient mode could be characterized by driving the vehicle under varying conditions, such as accelerating, decelerating, or climbing a grade, whereby the power demands are continuously changing.
Precise metering of the amount of fuel and air delivered to the engine is necessary to achieve the desired combustion as well as acceptable vehicle emissions. It is more difficult to maintain this mixture, known as the air/fuel ratio, when the vehicle is operating in a transient mode. For example, when the vehicle is accelerating, fuel demands increase to meet increased power needs. Conversely, when the vehicle is decelerating, fuel demands decrease to meet decreased power requirements. At steady state, fuel demands remain constant.
Traditionally, an engine controller receives inputs from a variety of sensors providing information regarding vehicle operating conditions such as engine speed, throttle position, and spark advance. The controller processes the information, determines the requisite amount of air and fuel, and relays this information to a fuel actuator (e.g., a fuel injector), which adjusts the fuel flow rate to provide the precise amount of fuel to the engine. One method of determining the amount of fuel is through the use of look-up tables within the controller containing information regarding the amount of fuel required during certain vehicle operating conditions, including transient and steady state modes. The information contained within the tables is experimentally derived from a combination of vehicle testing and engineering experience. This process can be time consuming and costly.
SUMMARY OF THE INVENTION
It is, therefore, one object of the present invention to provide a neural network based system and method for controlling fuel delivered to an internal combustion engine during transient driving conditions.
It is another object of the present invention to provide a fuel control system utilizing neural network technology to establish a fuel multiplier during transient modes.
It is yet another object of the present invention to provide a more precise method of determining fuel delivery to an internal combustion engine during transient modes.
It is a further object of the present invention to provide an improvised method for fuel control for a new motor vehicle.
To achieve the foregoing objects, the present invention is a neural network based system and method for controlling fuel flow in a vehicle fuel control system operating in a transient mode. The vehicle fuel control system comprises a plurality of inputs for detecting vehicle conditions, including engine speed, manifold absolute pressure, throttle position, AIS motor position, oxygen, and spark advance position, and a neural network based controller for calculating a fuel multiplier based on the inputs, such that the fuel multiplier controls fuel flow in the vehicle fuel control system. This neural network based controller comprises an input node for each of the input sensors, for normalizing each of the inputs and for multiplying each of the inputs by an input weight. Two hidden nodes receive and sum the output from each of the input nodes into an intermediate output, and then a hyperbolic tangent function for each hidden node receives this intermediate output and transmits a hidden layer output. Next, an output node receives the hidden layer output from each of the hidden nodes, multiplies each of these hidden layer outputs by an output weight resulting in a weighted output, and sums the weighted output into a summed output, before an output hyperbolic tangent function receives this summed output, transmits an output layer output and denormalizes the output layer output into the fuel multiplier.
One advantage of the present invention is that a neural network based transient fuel control system and method is provided for an internal combustion engine. Another advantage of the present invention is that the design method optimized the inputs and network topology used to implement the neural network fuel control system. Yet another advantage of the present invention is that the method improves the development cycle for a new vehicle since testing time is decreased and there is increased commonality between different engine types.
Other objects, features and advantages of the present invention will be readily appreciated as the same becomes better understood after reading the subsequent description taken in conjunction with the accompanying drawings.
REFERENCES:
patent: 5410477 (1995-04-01), Ishi et al.
patent: 5479571 (1995-12-01), Parlos et al.
patent: 5542403 (1996-08-01), Borland
patent: 5584277 (1996-12-01), Chen et al.
patent: 5781700 (2000-08-01), Puskorius et al.
patent: WO-96/05421 (1996-02-01), None
Joseph R. Asik, Jennifer M, peters, Garth M. Meyer, and Donald X. Tang, Ford Motor Co.; Transient AF Estimation and Control Using a Neural Network; International Congress & Exposition, Detroit, Michigan, Feb. 24-27, 1997, SAE Technical Paper Series 9070619.
Liu Sam K.
Robinson William O.
Stander Douglas M.
Sway-Tin Min
Calcaterra Mark P.
Chrysler Corporation
Dolinar Andrew M.
LandOfFree
Neural network based transient fuel control method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Neural network based transient fuel control method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Neural network based transient fuel control method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2520845