Abstract
Congestion, in India, is generally defined using volume-capacity (V/C) ratio. The passenger car unit (PCU) measures the volume, and capacity here is subjective. In such circumstances, volume and capacity cannot be directly measured. Hence, the determination of the actual capacity of any road remains debatable. As a result, the measure of the degree of congestion becomes subjective. This paper discusses conventional techniques for quantifying congestion and describes congestion using fuzzy tools and techniques. This paper uses two input variables that give direct and precise measures such as speed and inter vehicular distance (IVD) in the fuzzy model and volume and capacity as two input parameters for conventional methods. The congestions were calculated using conventional and fuzzy techniques on the roads of Central Kolkata and compared those quantified congestions on each road to find out the more reliable techniques among themselves.
TopMachine Learning Approach
Machine learning techniques are divided into four branches:
- •
Supervised Learning: It works on the set of unlabeled data. The various algorithms like Artificial Neural Networks (ANN), K-Means, and Bayesian Belief Networks (BBN) are working on the supervised learning technique. These algorithms are used to train the standard data, and while adjusting the parameters, the data classification is performed.
- •
Unsupervised Learning: The unsupervised learning method characterizes the data structure. It is not dependent on previously labeled data. An algorithm like clustering and outlier detection works well in unsupervised machine learning techniques (Sinha et al., 2019).
- •
Semi-Supervised Learning: Semi-Supervised learning technique is a mixture of supervised and unsupervised learning methodology. The training process required both labeled and unlabeled datasets (Sinha et al., 2016).
- •
Self-Training Learning: Self-Training is also a kind of self-machine learning technique, where the wrapper algorithm will use for both self-labeling and decision-directed learning. In starting, the unlabeled data is labeled based on the model. The unlabeled points are labeled and retrain the occurrences that a new model is learned. This process is repeated for the entire dataset until the model accuracy is not achieved (Sinha et al., 2021).
Key Terms in this Chapter
Random Forests: There seem to be various random forests in this classifier that offer a value. The matter of the most votes is the actual outcome—Researchers employed several machine learning classifiers to detect false news. The random forest is also one of those classifiers.
Operational Speed: The speed at which traffic flows on the road. The operational speed depends on the V/C ratio of the particular road.
Volume Of Traffic (V): Total number of vehicles flowing at a point on the road in one hour.
Neural Network: Various machine learning methods are employed to aid with categorization issues. The neural network is also one of the techniques used for allocation and optimization.
Spacing / Inter Vehicular Distance (IVD): Spacing is defined as the distance measured from the head to head of each successive vehicle.
Capacity Of Road (C): The maximum number of vehicles that can pass a point on the road in one hour under the most nearly ideal roadway and traffic conditions.
V/C Ratio: The V/C ratio is referred to as the ratio between the volume of traffic and the capacity of the road.
Levels of Service (L.O.S): It is the qualitative expression describing the traffic condition at a given state of the road.
Fuzzy Logic: Fuzzy Logic mimics complex human reasoning to arrive at realistic conclusions about reality's imprecise and often fuzzy nature. Fuzzy logic is an approach to computing based on “degrees of truth” rather than the usual “true or false” (1 or 0) Boolean logic on which the modern computer works.
Logistic Regression: Whenever the quantity is predictable, is definite, then the classifier is employed. For example, it can anticipate or provide a true or false result. Researchers in determine engaged this classifier