Towards Prediction Model For IoT-Wireless Sensor Network Packet Transmission Using Soft Computing Technique

Abstract

Packet transmission is a function of signal strength and distance in a wireless sensor network (WSN). The interval between the source node and sink node is of high significance in a wireless sensor network. This work developed a soft computing model that relates distance, signal strength, packet received and packet transmitted. An experiment was carried out using the CC2430 module for data acquisition. A learning process is carried out using the neural network (NN) tool toolbox. The entire datasets for training, testing, and validation are 60%, 25%, and 15% respectively. Regression result from training, testing, and validation was realized with R yielding 0.99972, 0.99974 and 0.99954 respectively. The summation of R gave 0.99869 in the work. The significance of the R value is that when it tends towards 1, the result demonstrates the accuracy of the model. This means that when the output of the trained and the test data are compared, it shows closeness symmetry. The closeness depicts how well the model performs. At convergence, the result shows that the coefficient for transmitted packet, distance and signal strength are predicted as 0.9969, 1.10800, and 0.7 respectively.