Optimization of Intelligent Control System Characteristics for Unmanned Ground Vehicles with the Use of Genetic Algorithms

This paper studies the possibilities of optimizing the main characteristics of the developed intelligent control system for unmanned ground vehicles with the use of genetic algorithms. Authors presented the software architecture designed on the basis of the developed simulation model of the behaviour of unmanned ground vehicles (UGVs) interacting with other road users such as manned ground vehicles (MGVs), pedestrians, etc. A two-objective optimization problem was formulated and solved, the objective functions of which are the output traffic flow and the number of potential accidents with a participation of UGVs. The Monte Carlo experiments were performed to confirm the sensitivity of the objective functions’ values to the control parameters of the model under various configurations of the street road network. The Pareto fronts are computed with the use of the previously created real-coded genetic optimization algorithms. It allows choosing the best trade-offs that provide the most preferable operating modes of the intelligent transportation system under various configurations of digital road networks (DRNs).