Automatic parking technology will help to solve the problem of parking difficulties caused by narrow parking spaces or poor driver skills and prevent accidents. The core of this technology is fuzzy controller. Therefore, this paper focuses on the fuzzy controller control method of automatic parking system. Based on the introduction of fuzzy control principle, this paper analyzes the types and the specific implementation methods of obtaining fuzzy rules, and concludes several feasible fuzzy controller methods.
With the increasing number of cars in our country, the road is crowded, the parking space is tense, and the parking situation becomes more and more complex. When the drivers parking in narrow parking spaces, they are often unable to park easily because of the lack of practical driving skills. Especially when parking in parallel, the drivers tend to be “forward-looking” but carelessly, they will lead to the rear of the car and obstacles. Raw collision. Therefore, it provides a good solution for parking difficulties in narrow parking spaces. Therefore, it is of great significance to actively explore and study the control method of fuzzy controller for automatic parking system. The principle of fuzzy automatic control is to simulate the control activities described by natural language by computer, so as to achieve effective control of industrial processes. As mentioned above, in order to control an industrial object effectively, a skilled operator must complete the process of fuzzification of precise quantity, fuzzy decision-making and precision of fuzzy quantity. When we use computer to simulate human control, because the computer itself can not think, we must sum up the operator’s control strategy as a set of qualitative and inaccurate control rules expressed in language, after certain mathematical processing, stored in the computer. In addition, we should imitate the human reasoning process and determine the approximate reasoning rules. In this way, the computer can make the fuzzy decision according to the input fuzzy information, the control rules and the approximate reasoning rules, and finally realize the automatic control. Fuzzy control rules are also called fuzzy control algorithm. In fact, the operator’s experience is summarized and transformed into a fuzzy control language that can be recognized by the fuzzy control system, that is, the fuzzy control rules [4]. This rule is the core of the fuzzy control system. In the process of designing a specific fuzzy controller, people often choose the input variables of the fuzzy control system as error E (E=Y-R, R as reference input) or the sum of error E and error S (S=E dt) or error E and error change rate EC (EC=dE/dt), while the output variables of the fuzzy control system are chosen as the change of control quantity U or control quantity U. The fuzzy controller has only two inputs and one output. The input variables are set to error E (E = Y-R, R is reference input) and error change EC (EC = dE/dt), and the output variables are set to de-ambiguity U or de-ambiguity change rate U. In this paper, the following fuzzy conditional statements are used to describe the control: if E is a and EC is B then U is C or if E is a and EC is B then U is c., a, B and C are subsets of fuzzy control [5]. This kind of fuzzy controller only has many inputs and many outputs. In this paper, the following fuzzy conditional statements are used to describe the control: If X1 is A1 and X2 is A2 and… Then U is c1, U is c2… In the formula, a1, a2,…
C1, C2… It is a subset of fuzzy control. Several specific implementation methods for obtaining fuzzy rules. Based on the actual operation process. Many skilled technicians use fuzzy control rules to operate in the complex industrial system controlled by human. But most of the time they don’t know what fuzzy control is, because all their operations are based on the experience of experts or the experience of predecessors. They do not know what the principle is, nor can they accurately describe it. If we know the principle of fuzziness, we can use this phenomenon to extract fuzzy rules. By observing and recording the actual operation process of technical personnel, for example, when controlling a device, we can summarize the fuzzy rules by input and output records. A method of extracting fuzzy rules is similar to that based on actual operation process.
This method can be called artificial experience and control based approach. It can also be said to directly extract expert experience. The first way is to indirectly extract expert experience. In essence, there is no difference between experts’ experience. Autonomous learning based on fuzzy controller. This kind of fuzzy controller has the function of autonomous learning, and is the process of autonomous learning of fuzzy control. It is similar to human autonomous learning, but it is different. This kind of control is very novel. Its structure is hierarchical. We can divide his rule base into two parts: the first is general rule base; the second is macro rule base. This rule base composed of macro rules has the ability of autonomous learning, which is similar to human autonomous learning.
According to the grasp of the overall performance of the fuzzy system, the fuzzy controller can learn autonomously, so that the system tends to be perfect. The fuzzy control rules obtained by using the fuzzy control system for reasoning are all fuzzy variables, but in the actual control process, the fuzzy variables can not meet the actual control requirements. It is hoped that the desired results can be obtained only when the precise operation quantity is obtained, and the executing mechanism can be accurate. Usually, it is necessary to transform the ambiguity. Usually, this process is also called deblurring output. Maximum membership method. This method usually needs the element with the greatest degree of membership in the output universe. In practical fuzzy control systems, it is necessary to select the greatest degree of membership in the fuzzy set of output variables.
If there are many domain elements of the maximum degree of membership, the domain elements can be obtained by calculating the arithmetic mean. The control function of the element with the largest membership degree is simple and easy, and the real-time performance is good. But this method can use less information because the smaller element is not selected. Median method. This method usually needs mathematical method to make full use of the set of output fuzzy variables. This method also uses the corresponding universe elements, and needs to be selected in the fuzzy set of output variables. Usually, the corresponding universe elements are obtained by means of the mean points of the area enclosed by the membership function curve of the set of output fuzzy variables and their corresponding abscissa coordinates. For the fuzzy control system, if you want to complete the behavior of the fuzzy control, as long as you input the corresponding numerical value to the fuzzy controller, after several steps of the fuzzy control, you can get an accurate control quantity, which acts on the controlled object through the fuzzy controller.
Usually, the off-line calculation method is used to generate the corresponding fuzzy control table. The content of the fuzzy control table is the observation value and the corresponding control value, so when the measured value is obtained, the appropriate transformation is made. The corresponding control value can be obtained. This method can effectively reduce the amount of on-line calculation in practical application. The expected goal of parking assistant system is to find the optimal parking path, and the ultimate goal is to make the car parking in place, but the driving position of the car is affected by many uncertainties [8]. For example, car body deflection angle, body direction and vehicle speed, so there may be errors in the parking process. In addition, the problem of velocity fluctuation at inflection point is also difficult to deal with, and the planning of the initial parking position is the result of the ideal model calculation, which will have an error with the actual situation [9]. Driver driving technology is also a consideration for whether a car can park at an ideal initial parking position. Moreover, the signal acquisition system can not meet the accuracy requirements, the collected data will also have errors, and the control system will also have errors in data processing. In summary, we can see that the car can not be parked in a completely regular trajectory, thermostatic element but can be roughly driven in accordance with the path, so the fuzzy control system is adopted as the control system. But no matter what kind of control method, there will be advantages and disadvantages. In order to be truly applied in practice, it is necessary to do real vehicle experiments and error analysis in order to improve the practical application value.