| Popis: |
Path planning algorithms for mobile robots are analyzed from the point of view of the possibility of their use for autonomous systems. Particular attention is paid to the analysis of the classical approaches possibility for the implementation of autonomous path replanning based on sensory data in the incompleteness and fuzziness of information conditions. Among the algorithms of the classical approach, the most adaptive ones turned out to be those that use artificial potential fields and the Monte Carlo method (sampling-based). It is shown that most algorithms, including those based on the approach of computational intelligence, provide replanning only on the basis of constant updating of global information about the environment. It was revealed that the hybrid approach to solving the path planning problem is the most adaptive, combining the techniques of global and local planning. The methods of this approach combine classical planning models with models based on computational intelligence. In addition, it is shown that the question of the homogeneity of the integration of solutions to various navigation tasks and the mutual influence of errors arising at each of the stages remains unexplored. |
