Fuzzy nerve network control method for automobile driving robot system

Abstract

The invention discloses a fuzzy nerve network control method for an automobile driving robot system. The method comprises: performing speed tracking on a test vehicle from the characteristic parameters of an automobile driving robot gearshift manipulator, a throttle mechanical leg, a brake mechanical leg and a clutch mechanical leg; utilizing a Sugeno fuzzy nerve network to realize the speed tracking and control of an automobile driving robot, inputting characteristic parameters, adopting a generalized bell-shaped function for a membership function in terms of type, using a mixing learning algorithm to train and optimize network parameters, and determining optimum network parameters; and according to the optimum network parameters and four characteristic parameters detected in real time of the automobile driving robot, using the Sugeno fuzzy nerve network to perform speed tracking and accurate control on the automobile driving robot. The method disclosed by the invention is provided with an on-line self-learning capability, can accurately track a given target speed, and has good robustness in various test conditions.

Description

technical field [0001] The invention relates to the technical field of automobile automatic driving, in particular to the control of an automobile driving robot, in particular to a fuzzy neural network control method for an automobile driving robot system. Background technique [0002] Automobile driving robot refers to an industrial robot that does not need to be modified on the vehicle, can be installed in the cab without damage, and is suitable for various models, replacing the driver to drive the vehicle in dangerous conditions and harsh environments. When the driving robot is used for car test, bionic disabled person driving car operation or applied in nuclear power plant, heavy chemical industry and other places with serious pollution, it can reduce the labor intensity of human beings, reduce the harm to human beings in dangerous and harsh environments, and improve the objectivity and accuracy of test results. Accuracy, the ability of the disabled to drive a car, and s...

AI Technical Summary

Problems solved by technology

For example, Chinese patent 200410065844.0 "Pneumatic-electric hybrid driving robot for automobile test" uses variable parameter PID control method to realize the vehicle speed tracking of the driving robot, but it has the disadvantage that the regulator parameters are difficult to adjust online, and the vehicle speed fluctuates greatly; 2) Feedforward PID control, such as the US patent US5372035 "robot for driving a car on a chassis dynamometer" uses feed-forward PID control to realize the control of the driving robot system, but there are controller parameters that need to be manually set in advance, and the anti-interference ability for different test conditions poor

Images