Smart monitoring of off-road vehicles is cursed by complex and expensive IoT sensors technologies. High dependency on the cloud/fog computation, availability of the network and expert knowledge are a handicap in rural off-network areas. The use of edge devices such as smartphones with dedicated computational capabilities near the data source represents a powerful solution that is yet to be developed at the commercial level.
This article presents a hybridized computational intelligence methodology to develop an edge-device-enabled AI technology for the health monitoring and diagnosis (HM&D) of off-road vehicles, taking use of very economic microphones as sensors. Enhanced Selection and Log-scaled Mutation Genetic Algorithms (ESALOGA) is used to evolve the structure of the Artificial Neural Network (ANN) toward an optimally lightweight structure. Each evolved Lightweight ANN (LANN) structure is then trained by a scaled conjugate gradient back-propagation (SCGB) training algorithm to optimize corresponding weights and biases.