Increasing coverage of cell-corner users is one of the essential requirements to improve the performance of current cellular systems. Unlike the typical network user, a user located at the cell boundary suffers from severe inter-cell interference, a problem which is aggravated by the deployment of small cells in heterogeneous networks. To address this issue, inter/intra-cell base station cooperation has been proposed in wireless standards as a technique to limit the amount of inter-cell interference and hence improve coverage.Motivated by this ongoing debate in standard settings, this thesis presents a stochastic geometry model to study base station cooperation in heterogeneous cellular networks. Within this model, base stations are assumed to be randomly located on a plane according to a Poisson point process, such that the so-called coverage probability, i.e., the probability that the signal-to-interference ratio at the designated receiver exceeds a desired threshold value, can be analytically characterized as a function of the network parameters. In contrast, traditional models either ignore the presence of interferes or make approximate models for the aggregate interference power distribution. Unlike any other work in the literature, the proposed model allows base stations and receivers to be equipped with multiple antennas. Specific cooperation schemes that are considered include joint transmission, base station silencing, cooperativere-transmission, and a distributed version of the well-known Alamouti space-time code. For each of these schemes, integral expressions for the coverage probability are characterized for both the typical network user and the typical cell-corner user.The expressions derived are used to quantify the relative coverage gain due to cooperation at the typical cell-corner user vs the typical network user. At a high level, the theoretical analysis reveals the existence of two qualitatively different operating regimes, depending on whether the coverage probability is close to 1 or 0. In the high-coverage regime where the coverage probability is close to 1, the typical network user and the typical cell-corner user are diversity-limited, so cooperation techniques exploiting spatiotemporal diversity such as the Alamouti code or cooperative re-transmission are highly effective in increasing coverage. In the low-coverage regime where the coverage probability is close to 0, on the other hand, the typical network user and the typical cell-corner user are interference-limited, so cooperation techniques such as joint transmission and base station silencing are effective in increasing coverage as they suppress part of the interference power.