2013/09/23 by Adriano Zanin Zambom, Zambom, Adriano Zanin, Julian A. A. Collazos +3
Computer Science · #FOS: Computer and information sciences #Metaheuristic Optimization Algorithms Research #Methodology (stat.ME) #Optimization and Search Problems #Robotic Path Planning Algorithms
paper · pdf · doi:10.48550/arxiv.1309.5999
openalex publication_date 2013/09/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
In real-time trajectory planning for unmanned vehicles, on-board sensors,\nradars and other instruments are used to collect information on possible\nobstacles to be avoided and pathways to be followed. Since, in practice,\nobservations of the sensors have measurement errors, the stochasticity of the\ndata has to be incorporated into the models. In this paper, we consider using a\ngenetic algorithm for the constrained optimization problem of finding the\ntrajectory with minimum length between two locations, avoiding the obstacles on\nthe way. To incorporate the variability of the sensor readings, we propose a\nmore general framework, where the feasible regions of the genetic algorithm are\nstochastic. In this way, the probability that a possible solution of the search\nspace, say x, is feasible can be derived from the random observations of\nobstacles and pathways, creating a real-time data learning algorithm. By\nbuilding a confidence region from the observed data such that its border\nintersects with the solution point x, the level of the confidence region\ndefines the probability that x is feasible. We propose using a smooth penalty\nfunction based on the Gaussian distribution, facilitating the borders of the\nfeasible regions to be reached by the algorithm.\n