Fuzzy Logic Examples | Fuzzy Logic Part 3

Summary

In this video, the presenter walks through the design of a fuzzy logic controller for the cart-pole problem, demonstrating how intuitive, experience-based design can replace complex mathematical models. By balancing a pole on a moving cart, the controller adjusts the force applied to the cart based on the pole's angle, angular velocity, and cart position. The design uses simple fuzzy rules and membership functions to maintain balance, and the video explains how fuzzy inference systems can be applied to various real-world scenarios like autonomous driving and medical devices. The importance of experience-driven control design is highlighted throughout.

Takeaways

• 😀 Fuzzy inference systems can be designed without mathematical models, using intuition and experience with the system.
• 😀 The cart-pole problem is used as an example, where a fuzzy logic controller can balance a pole on a moving cart.
• 😀 The system has four state variables: pole angle, angular velocity, cart position, and cart velocity.
• 😀 The fuzzy logic controller can be designed using MATLAB, with a Mamdani fuzzy inference object and simple rules.
• 😀 Initially, the fuzzy controller uses only the pole angle (Theta) as input to balance the pole.
• 😀 Membership functions are defined for inputs and outputs, such as 'Z' and 'S' shapes for angle and bell-shaped functions for force.
• 😀 Simple fuzzy rules are created, such as 'If Theta is negative, then force is negative medium.'
• 😀 By considering both pole angle and angular velocity (Theta_dot), the controller can improve performance by reducing the speed of pole rotation.
• 😀 Additional input variables like cart position and velocity can be added to further refine the controller's response and prevent the cart from drifting.
• 😀 The design can be done either using code in MATLAB or graphically through the Fuzzy Logic Designer app, offering flexibility for users.
• 😀 Practical applications of fuzzy inference systems include autonomous driving, smoother train acceleration in Japan's Sendai Subway, and artificial pancreas systems for personalized insulin delivery.

Q & A

What is the primary goal of the fuzzy inference system in the script?

The primary goal is to design a fuzzy inference-based controller to balance a pole on a cart using fuzzy logic, which doesn't require a mathematical model of the system.

How does the fuzzy logic controller balance the pole on the cart?

The fuzzy logic controller adjusts the cart's position based on the pole's angle, moving the cart in the direction of the tilt to counterbalance the pole.

What type of fuzzy membership functions are used for the pole angle in the design?

The pole angle uses two types of fuzzy membership functions: a Z-shaped function for negative values and an S-shaped function for positive values, both designed to control the pole within a specific range of angles.

Why is a fuzzy logic controller useful when there is no accurate model of the system?

A fuzzy logic controller is beneficial in such cases because it allows the design of control systems based on experience and intuition about the system's behavior, without needing a precise mathematical model.

What is the significance of adding angular velocity (Theta_dot) to the controller?

Adding angular velocity helps slow down the pole's rotation as it reaches the balance point, preventing overshooting and improving the stability of the system.

What was the result of the first version of the controller that used only the pole angle?

The first version of the controller kept the pole upright but caused oscillation due to overshooting, as the cart tried to move under the pole to correct its position.

How does the fuzzy logic controller adjust when the cart begins to wander slightly?

The controller uses additional inputs like cart position and velocity to apply small corrective forces, keeping the cart steady without disrupting the pole's balance.

What practical application of fuzzy logic is mentioned in the script related to autonomous driving?

Fuzzy logic is used in autonomous driving for a parking system, where sensors help guide the car into a parking spot, handling the car's movement constraints intuitively.

Why did the designer choose bell-shaped membership functions for the output force?

Bell-shaped membership functions were chosen because they provide a smooth transition for the output force, which is limited to a specific range (about 12 newtons in either direction), ensuring appropriate adjustments to balance the pole.

How is fuzzy logic applied in the medical field according to the script?

In the medical field, fuzzy logic is used in the development of an artificial pancreas for insulin delivery, where it adjusts to individual patient needs without requiring complex mathematical models, offering a personalized treatment approach.