Lecture Note
University
California State UniversityCourse
CS 3590 | Data Communications and NetworkingPages
1
Academic year
2023
Jithin Jacob Issac
Views
0
p {margin: 0; padding: 0;} .ft00{font-size:21px;font-family:NimbusSansBold;color:#000000;} .ft01{font-size:18px;font-family:NimbusSansBold;color:#000000;} .ft02{font-size:15px;font-family:OpenSymbol;color:#000000;} .ft03{font-size:18px;font-family:NimbusSans;color:#000000;} UNIVERSAL THEORETICAL WIRELESS SIGNAL PROPAGATION PREDICTION MODELLING Introduction Wireless networks becoming ubiquitous with growing demand Various methods used for modeling wireless signal propagation Traditional empirical models include One Slope, Dual Slope, Hata etc. AI methods like ANN, ANFIS being explored for prediction accuracy Literature Review Obstacles affect indoor/outdoor signal propagation differently Neural networks transform inputs like distance, obstacles to RSSI output ANFIS combines neural networks and fuzzy logic for function approximation Has learning capability of neural networks and knowledge base of fuzzy logic ANFIS architecture has 5 layers - input, 3 hidden, output Premise and consequent parameters modified during training PSO optimizes ANFIS parameters based on swarm intelligence Objectives Compare PSO trained LOG10D-ANFIS, LOG10D-ANFIS and ANFIS Obtain RSSI using One Slope and Two Ray Ground Reflection models Analyze performance using RMSE and graphical comparisons Conclusion AI methods like PSO-ANFIS useful for accurate wireless signal prediction Enhances modeling of complex real-world radio propagation environments More research needed on integrating AI with empirical models Key Highlights ANFIS combines strengths of neural networks and fuzzy logic systems PSO provides efficient training of ANFIS parameters Comparative study analyzes different ANFIS frameworks for RSSI modeling AI and computational intelligence crucial for future wireless networks
Universal Theoretical Wireless Signal Propagation Prediction Modelling
Please or to post comments