Fuzzy Logic based Autonomous Parking Systems -- Part III: A Fuzzy Decision Tree System

Authors Yu Wang and Xiaoxi Zhu present a paper on the design of a Hybrid Fuzzy Controller for speed and steering angle control in an Intelligent Autonomous Parking System (IAPS). The proposed controller combines a Base Fuzzy Controller (BFC) with a Supervisory Fuzzy Decision Tree Controller (SFDTC) to ensure optimal parking trajectory with minimal computational cost. Building upon previous work in fuzzy logic control for unmanned parking, the BFC provides efficient trajectory planning while the SFDTC enhances system robustness in noisy environments by integrating Decision Tree theory and fuzzy inference mechanisms. A data training process is also implemented to improve control performance. Through simulations conducted using MATLAB, the IAPS equipped with the new Hybrid Fuzzy Controller with Fuzzy Decision Tree (HFC-FDT) demonstrates superior performance characterized by faster convergence and minimal deviation from the optimal parking trajectory. This research contributes to advancing autonomous parking systems through the detailed design and evaluation of the Supervisory Fuzzy Decision Tree Controller.

Introduction:

The development of autonomous vehicles has been a major focus in the field of transportation. One key aspect of this technology is intelligent parking systems, which aim to automate the process of finding and maneuvering into a parking spot. In order to achieve this, precise control over speed and steering angle is crucial. In their research paper titled "Design of Hybrid Fuzzy Controller for Speed and Steering Angle Control in Intelligent Autonomous Parking System", authors Yu Wang and Xiaoxi Zhu propose a new controller that combines fuzzy logic with decision tree theory to improve the performance of an Intelligent Autonomous Parking System (IAPS).

Overview of Previous Work:

The use of fuzzy logic control for unmanned parking has been extensively studied by researchers in recent years. This approach allows for efficient trajectory planning based on linguistic rules rather than complex mathematical models. However, one limitation of traditional fuzzy controllers is their sensitivity to noise and uncertainty in real-world environments.

To address this issue, Wang and Zhu build upon previous work by incorporating decision tree theory into the design of their controller. Decision trees are widely used in machine learning algorithms due to their ability to handle noisy data effectively. By integrating decision tree theory with fuzzy inference mechanisms, the proposed controller aims to enhance system robustness in noisy environments.

Designing the Hybrid Fuzzy Controller:

The proposed Hybrid Fuzzy Controller with Fuzzy Decision Tree (HFC-FDT) consists of two main components: a Base Fuzzy Controller (BFC) and a Supervisory Fuzzy Decision Tree Controller (SFDTC). The BFC is responsible for efficient trajectory planning while the SFDTC supervises its operation by making decisions based on input from sensors.

The BFC uses linguistic rules defined by experts to generate appropriate speed and steering angle commands based on inputs such as distance from obstacles, desired parking spot location, and vehicle velocity. These rules are then converted into membership functions using Mamdani's method.

On the other hand, the SFDTC utilizes a decision tree to make decisions based on the current state of the vehicle and its environment. The decision tree is constructed using training data obtained from simulations, which allows for improved control performance in real-world scenarios.

Evaluation through Simulations:

To evaluate the performance of the proposed controller, simulations were conducted using MATLAB. The IAPS equipped with HFC-FDT was compared to a traditional fuzzy controller and a PID controller commonly used in autonomous parking systems.

The results showed that the HFC-FDT outperformed both controllers in terms of convergence time and deviation from the optimal parking trajectory. This demonstrates the effectiveness of combining fuzzy logic with decision tree theory in improving control performance.

Conclusion:

In conclusion, Wang and Zhu's research paper presents a detailed design and evaluation of a Hybrid Fuzzy Controller for speed and steering angle control in an Intelligent Autonomous Parking System. By incorporating decision tree theory into traditional fuzzy logic control, their proposed controller offers improved robustness in noisy environments while maintaining efficient trajectory planning. The use of data training also contributes to better control performance, as demonstrated by simulation results. This research contributes to advancing autonomous parking systems and provides valuable insights for future developments in this field.