A straightforward analytical scheme is proposed for computing the long-time, asymptotic mean velocity and dispersivity (effective diffusivity) of a particle undergoing a discrete biased random walk on a periodic lattice amongst an array of immobile, impenetrable obstacles. The results of this Taylor-Aris dispersion-based theory are exact, at least in an asymptotic sense, and furnish an analytical alternative to conventional numerical lattice Monte Carlo simulation techniques. Results obtained for an obstacle-free lattice are employed to establish generic relationships between the transition probabilities, lattice size and jump time. As an example, the dispersivity is computed for a solute moving through an isotropic array of obstacles under the influence of a finite external field. The calculation scheme is also shown to agree with existing zero-field results, the latter obtained elsewhere either by first-passage time analysis or use of the Nernst-Einstein equation in the zerofield limit. The generality of this scheme permits the study of more complex lattice structures, in particular trapping geometries.