Novel Ellipsoidal Heights Predictive Models Based on Artificial Intelligence Training Algorithms and Classical Regression Models Techniques: A Case Study in the Greater Kumasi Metropolitan Area Local Geodetic Reference Network, Kumasi, Ghana: A Case Study in the Greater Kumasi Metropolitan Area (GKMA) Local Geodetic Reference Network, Kumasi, Ghana

Authors : Naa LAMKAI, Daniel ASENSOGYAMBIBI, Michael Stanley PEPRAH, Edwin Kojo LARBI, Benedict ASAMOAH, Philip OKANTEY

Pages : 493-515

View : 41 | Download : 11

Publication Date : 2022-12-30

Article Type : Research Paper

Abstract :The standard forward transformation for the direct conversion of curvilinear geodetic coordinates insert ignore into journalissuearticles values(φ, γ, Η); to its associated Cartesian coordinates insert ignore into journalissuearticles values(E, N, Z); has become a major challenge in most countries. This is due to the non-existence of the ellipsoidal height insert ignore into journalissuearticles values(h); in the modelling of their local geodetic reference network. Numerous studies in the past and recent years have suggested various mathematical techniques for predicting and estimating local ellipsoidal heights. Primary data used for the studies comprises of topographic data obtained from a survey in the Ghana urban water supply project in the Greater Kumasi Metropolitan Area insert ignore into journalissuearticles values(GKMA);.This study considered an empirical evaluation of soft computing techniques such as Back Propagation Artificial Neural Network insert ignore into journalissuearticles values(BPANN);, Generalized Regression Neural Network insert ignore into journalissuearticles values(GRNN);, Radial Basis Function Artificial Neural Network insert ignore into journalissuearticles values(RBFANN); and conventional methods such as Polynomial Regression Model insert ignore into journalissuearticles values(PRM);, Autoregressive Integrated Moving Average insert ignore into journalissuearticles values(ARIMA); and Least Square Regression insert ignore into journalissuearticles values(LSR);. The motive is to apply and assess for the first time in our study area, the working efficiency of the aforementioned techniques. Each model technique was assessed based on statistical hypothesis insert ignore into journalissuearticles values(F, t); tests and performance criteria indices such as arithmetic mean error insert ignore into journalissuearticles values(AME);, arithmetic mean square error insert ignore into journalissuearticles values(AMSE);, minimum and maximum error value, and arithmetic standard deviation insert ignore into journalissuearticles values(ASD);. The statistical analysis of the results revealed that, RBFANN, GRNN, BPANN, LSR, ARIMA and PRM, successfully estimated the ellipsoidal heights for the study area. However, the ANN models insert ignore into journalissuearticles values(RBFANN, BPANN, GRNN); outperforms the conventional models insert ignore into journalissuearticles values(LSR, PRM, ARIMA); in terms of accuracy and precision in estimating the local ellipsoidal heights. Also, statistical findings revealed that RBFANN produced more reliable results compared with the other methods. The main conclusion drawn from this study is that, the method of using soft computing is very much promising and can be adopted to solve some of the major problems related to height issues in Ghana. This study seeks to contribute to the existing knowledge on establishing a precise geodetic vertical datum in Ghana for national heightening purpose.
Keywords : Artificial Intelligence, Geodetic Reference Network, Height systems, Performance Criteria Indices, Statistical Hypothesis, Regression Models

ORIGINAL ARTICLE URL