Applied Computational Intelligence and Soft Computing

To enhance the performance of image classification and speech recognition, the optimizer is considered an important factor for achieving high accuracy. The state-of-the-art optimizer can perform to serve in applications that may not require very high accuracy, yet the demand for high-precision image classification and speech recognition is increasing. This study implements an adaptive method for applying the particle filter technique with a gradient descent optimizer to improve model learning performance. Using a pretrained model helps reduce the computational time to deploy an image classification model and uses a simple deep convolutional neural network for speech recognition. The applied method results in a higher speech recognition accuracy score—89.693% for the test dataset—than the conventional method, which reaches 89.325%. The applied method also performs well on the image classification task, reaching an accuracy of 89.860% on the test dataset, better than the conventional method, which has an accuracy of 89.644%. Despite a slight difference in accuracy, the applied optimizer performs well in this dataset overall.

In this paper, an improved particle swarm optimization method (PSO) is proposed to optimally size and place a DG unit in an electrical power system so as to improve voltage profile and reduce active power losses in the system. An IEEE 34 distribution bus system is used as a case study for this research. A new equation of weight inertia is proposed so as to improve the performance of the PSO conventional algorithm. This development is done by controlling the inertia weight which affects the updating velocity of particles in the algorithm. Matlab codes are developed for the adapted electrical power system and the improved PSO algorithm. Results show that the proposed PSO algorithm successfully finds the optimal size and location of the desired DG unit with a capacity of 1.6722 MW at bus number 10. This makes the voltage magnitude of the selected bus equal to 1.0055 pu and improves the status of the electrical power system in general. The minimum value of fitness losses using the applied algorithm is found to be 0.0.0406 while the average elapsed time is 62.2325 s. In addition to that, the proposed PSO algorithm reduces the active power losses by 31.6%. This means that the average elapsed time is reduced by 21% by using the proposed PSO algorithm as compared to the conventional PSO algorithm that is based on the liner inertia weight equation.

An agricultural model for allocation of crops is considered in this work using Pollination Intelligence Method. The model was constructed to solve farmer’s decision making in allocating crops to a piece of land using market price, known yield of crops, cost incurred during planting, and the total amount of land available. A new class of metaheuristic method called Flower Pollinated Algorithm is also presented in this work to solve the designed model. An improved version of the Flower Pollinated Algorithm called Pollination Intelligence Algorithm using an iterative scheme to override the switch parameter in Flower Pollinated Algorithm is also presented and used in solving the designed model. A case study of a farmer in Ife, Osun State, Nigeria, was used to implement the model, and the results obtained suggested that instead of allocating crops to land randomly based on farmer’s intuition, cost of planting, yield of crops, and market price were factors that must be considered by farmers for optimal profit before planting crops.

Nowadays, building intelligent systems for science, technology, engineering, and math (STEM) education is necessary to support the studying of learners. Intelligent problem solver (IPS) is a system that can be able to solve or tutor how to solve the problems automatically. Learners only declare hypothesis and goal of problems based on a sufficient specification language. They can request the program to solve it automatically or to give instructions that help them to solve it themselves. Knowledge representation plays a vital role in these kinds of intelligent systems. There are various methods for knowledge representation; however, they do not meet the requirements of an IPS in STEM education. In this paper, we propose the criteria of a knowledge model for an IPS in education. These criteria orient to develop a method for knowledge representation to meet actual requirements in practice, especially pedagogical requirements. For proving the effectiveness of these criteria, a knowledge model is also constructed. This model can satisfy these criteria and be applied to build IPS for courses, such as mathematics and physics.

Groundwater is a precious natural resource. Groundwater level (GWL) forecasting is crucial in the field of water resource management. Measurement of GWL from observation-wells is the principle source of information about the aquifer and is critical to its evaluation. Most part of the Udupi district of Karnataka State in India consists of geological formations: lateritic terrain and gneissic complex. Due to the topographical ruggedness and inconsistency in rainfall, the GWL in Udupi region is declining continually and most of the open wells are drying-up during the summer. Hence, the current research aimed at developing a groundwater level forecasting model by using hybrid long short-term memory-lion algorithm (LSTM-LA). The historical GWL and rainfall data from an observation well from Udupi district, located in Karnataka state, India, were used to develop the model. The prediction accuracy of the hybrid LSTM-LA model was better than that of the feedforward neural network (FFNN) and the isolated LSTM models. The hybrid LSTM-LA-based forecasting model is promising for a larger dataset.

Traffic law violation has been recognized as a major cause for road accidents in most parts of the world with majority occurring in developing countries. Even with the presence of rules and regulations stipulated against this, violators are still on the increase. This is due to the fact that the rules are not properly enforced by appropriate authorities in those parts of the world. Therefore, a system needs to be designed to assist law enforcement agencies to impose these rules to improve road safety and reduce road accidents. This work uses a Vehicle Plate Number Recognition (VNPR) system which is a real-time embedded system to automatically recognize license plate numbers. It provides an alternative means to VPNR using an open-source library known as openCV. The main aim of the system is to use image processing to identify vehicles violating traffic by their plate numbers. It consists of an IR sensor for detecting the vehicle. During testing, a minimum time was set for the sensor to detect the object which was recorded by the microprocessor. Once it was less than the set time, the camera was triggered to capture the plate number and store the image on the Raspberry Pi. The image captured is processed by the Raspberry Pi to extract the numbers on the image. The numbers on the capture imaged were viewed on a web page via an IP address. The system if implemented can be used to improve road safety and control traffic of emerging smart cities. It will also be used to apply appropriate sanctions for traffic law violators.