The main goal of the thesis is the proposition of highly accurate curved boundary conditions aiming to reduce the need for grid refinement or non-Cartesian grids meanwhile maintaining a highly favorable profile for massively parallel implementation on CPU or GPU systems. The novelly proposed boundary conditions are the nonunique outcome of a broader work of analysis, unification, and improvement of the lattice Boltzmann method directional boundary conditions (DBC). In particular, the thesis presents, additionally to the novel family of DBC, named ELI (Enhanced Local Interpolation), a set of new corrections to DBC. These corrections grant advanced accuracy, especially in the low Reynolds number case. Further, both ELI and the new adjustments are local from the algorithmic and the physical viewpoint: they do not need to recover information outside the boundary node and can model narrow gaps where only a single computational node is present.