In our daily lives, you are not always able to make a decision by saying Yes or No. This is because you might face situations where there is not enough information to make a decision. Or, we may be a little confused ourselves. For example, if someone asks you whether you are available on a certain date next month, you probably won’t say Yes or No straight away. It’s because you are not completely sure that you won’t be busy on that very date next month.
Tricky isn’t it? Fuzzy Logic in artificial intelligence helps computers deal with similar situations where input data is a bit unclear.
What is Fuzzy Logic?
The term Fuzzy means something that is a bit vague. When a situation is vague, the computer may not be able to produce a result that is True or False. As per Boolean Logic, the value 1 refers to True and 0 means False. But a Fuzzy Logic algorithm considers all the uncertainties of a problem, where there may be possible values besides True or False.
The term Fuzzy Logic was first described by Lotfi Zadeh in 1965. He thought that traditional computer logic is not capable of handling unclear or vague data. Similar to humans, there are many possible values between True and False that a computer can incorporate. These can be:
- Certainly yes
- Possibly yes
- Can’t say
- Possibly no
- Certainly no
Check out this simple example of Fuzzy Logic:
Problem – Is it hot outside?
- Yes (1.0)
- No (0)
According to conventional Boolean Logic, the algorithm will take a definite input and produce a precise result Yes or No. This is represented by 0 and 1, respectively.
- Very hot (0.9)
- Little hot (0.20)
- Moderately hot (0.35)
- Not hot (1.0)
As per the above example, Fuzzy Logic has a wider range of outputs, such as very hot, moderately hot and not hot. These values between 0 and 1 display the range of possibilities.
So, in cases where accurate reasoning cannot be provided, Fuzzy Logic provides an acceptable method of reasoning. An algorithm based on Fuzzy Logic takes all available data while solving a problem. It then takes the best possible decision according to the given input.
Charles Elkan, Assistant Professor of the Computer Science and Engineering department at the University of California at San Diego, shed some light upon Fuzzy Logic. He said that Fuzzy Logic in artificial intelligence is a generalized form of standard logic, where any concept might have a truth degree ranging between 0.0 and 1.0. Fuzzy Logic can be used for vague concepts, such as the characteristic of tallness. For example, we can say that President Clinton is tall, and the concept can have a degree of truth of 0.9.
He further said that Fuzzy Logic is very useful in low-level machine control especially in consumer appliances. Some special-purpose microprocessors built on Fuzzy Logic perform fuzzy operations on their hardware.
Let us now understand Fuzzy Logic in artificial intelligence by looking at its architecture.
Fuzzy Logic Architecture
The architecture of Fuzzy Logic consists of the following components:
This is the set of rules along with the If-Then conditions that are used for making decisions. But, modern developments in Fuzzy Logic have reduced the number of rules in the rule base. These set of rules are also called a knowledge base.
This is the step where crisp numbers are converted into fuzzy sets. A crisp set is a set of elements that have identical properties. Based on certain logic, an element can either belong to the set or not. Crisp sets are based on binary logic – Yes or No answers.
Here, the error signals and physical values are converted into a normalized fuzzy subset. In any Fuzzy Logic system, the fuzzifier separates the input signals into five states that are:
- Large positive
- Medium positive
- Medium negative
- Large negative
The fuzzification process converts crisp inputs, such as room temperature, fetched by sensors and passes them to the control system for further processing. A Fuzzy Logic control system is based on Fuzzy Logic. Common household appliances, such as air-conditioners and washing machines have Fuzzy Control systems within them.
The inference engine determines how much the input values and the rules match. The rules are applied based on the input values received. Then, the rules are used to develop control actions. The inference engine and the knowledge base together are called a controller in a Fuzzy Logic system.
This is the inverse process of fuzzification. Here, the fuzzy values are converted into crisp values by mapping. There will be several defuzzification methods for doing this, but the best one is selected as per the input. This is a complicated process where methods, such as the maximum membership principle, weighted average method and centroid method, are used.
Advantages of Fuzzy Logic in Artificial Intelligence
The benefits of using Fuzzy Logic systems are as follows:
Natural Language Processing
- It is a robust system where no precise inputs are required
- These systems are able to accommodate several types of inputs including vague, distorted or imprecise data
- In case the feedback sensor stops working, you can reprogram it according to the situation
- The Fuzzy Logic algorithms can be coded using less data, so they do not occupy a huge memory space
- As it resembles human reasoning, these systems are able to solve complex problems where ambiguous inputs are available and take decisions accordingly
- These systems are flexible and the rules can be modified
- The systems have a simple structure and can be constructed easily
- You can save system costs as inexpensive sensors can be accommodated by these systems
Disadvantages of Fuzzy Logic in Artificial Intelligence
Let us look at the drawbacks of Fuzzy Logic systems:
- The accuracy of these systems is compromised as the system mostly works on inaccurate data and inputs
- There is no single systematic approach to solve a problem using Fuzzy Logic. As a result, many solutions arise for a particular problem, leading to confusion
- Due to inaccuracy in results, they are not always widely accepted
- A major drawback of Fuzzy Logic control systems is that they are completely dependent on human knowledge and expertise
- You have to regularly update the rules of a Fuzzy Logic control system
- These systems cannot recognize machine learning or neural networks
- The systems require a lot of testing for validation and verification
Applications of Fuzzy Logic
The applications of Fuzzy Logic are spread across several fields. They are as follows:
- Controlling arterial pressure when providing anaesthesia to patients
- Used in diagnostic radiology and diagnostic support systems
- Diagnosis of prostate cancer and diabetes
- Handling underground train operations
- Controlling train schedules
- Braking and stopping vehicles based on parameters, such as car speed, acceleration and wheel speed
- Locating and recognizing targets underwater
- Supports naval decision making
- Using thermal infrared images for target recognition
- Used for controlling hypervelocity interceptors
- Controlling water purification plants
- Handling problems in constraint satisfaction in structural design
- Pattern analysis for quality assurance
- Fuzzy Logic is used for tackling sludge wastewater treatment
- Steer ships properly
- Selecting the optimal or best possible routes for reaching a destination
- Autopilot is based on Fuzzy Logic
- Autonomous underwater vehicles are controlled using Fuzzy Logic
Washing systems powered by Fuzzy Logic
Modern washing machines powered by Fuzzy Logic are becoming popular these days. They have sensors that continuously track variations in temperature. It adjusts the controls and operations accordingly. These systems perform well, and are productive and cost efficient.
For best wash results, Fuzzy Logic controls the washing process, water temperature, spin speed, wash time, water intake and rinse performance. Advanced machines do the following:
- Check the load of the machine to prevent overloading
- Check the hardness of water and determine the type of cloth material
- Advises the user on the optimal amount of detergent. They also check whether the detergent is in liquid or powdered form
- They learn from previous washing experiences, and memorize algorithms to enhance the washing results
Most of the systems are based on energy-saving technology that helps you save power while washing clothes. These features help you save energy even if you wash lots of clothes 2 to 3 times a week. The sensors monitor the entire washing process, and make necessary adjustments and corrections for best washing outcomes.
Fuzzy Logic in these washing machines checks the amount of dirt and grime on clothes, the direction of the spin and the quantity of soap required. For better spinning, the washing load is properly balanced. In case an imbalance is detected, the spinning speed is decreased. Furthermore, balancing the washing load helps to reduce the spinning noise.
Companies such as Panasonic use similar technology in their dishwashers. Fuzzy Logic is used in adjusting the cleaning cycles of dishwashers, along with the wash and rinse methods. The performance of the machine also depends upon the number of dishes put into the washer.
Although Fuzzy Logic in artificial intelligence helps to mimic human reasoning, these systems need expert guidance to be built. This lets you rely on the experience of experts who have a better understanding of the system. Fuzzy Logic can also be used for enhancing the execution of algorithms. IBM Watson uses Fuzzy Logic and fuzzy semantics.
If you are interested to know more about machine learning and artificial intelligence, check out our PG Diploma in Machine Learning and AI program which is designed for working professionals and provide 30+ case studies & assignments, 25+ industry mentorship sessions, 5+ practical hands-on capstone projects, more than 450 hours of rigorous training & job placement assistance with top firms.