Neural Network- and Fuzzy Control-Based Energy Optimization for the Switching in Parallel Hybrid Two-Wheeler

Abstract

Optimization of a two-wheeler hybrid electric vehicle (HEV) is a typical challenge compared to that for four-wheeler HEVs. Some of the challenges which are particular to two-wheeler HEVs are throttle integration, smooth switching between power sources, add-on weight compensation, efficiency improvisation in traffic, and energy optimization. Two power sources need to be synchronized skillfully for optimum energy utilization. A prominent variant of HEV is that it easily converts conventional scooters into parallel hybrids by “Through-the-Road (TTR)” architecture. This paper focuses on three switching control strategies of HEVs based on the state of charge, fuzzy logic, and neural network. Further, to optimize energy usage, all these control strategies are compared. Energy management control for the TTR model is developed with vehicle parameters in the Simulink environment and simulated using the “World Harmonized Motorcycle Test Cycle” (WMTC) drive cycle. The multivariable input model is presented with a fuzzy rule-based hybrid switching control. A similar system is also modeled with a neural network-based decision control and the observations are tabulated for the fuel economy and energy management. Simulation results show that the neural network-based optimization results in minimal energy consumption among all three hybrid operations.