| 20080268834 | Method of grouping users to reduce interference in mimo-based wireless network - In a system of MIMO communications in a wireless network, a number of wireless units are logically divided into a plurality of user groups, through operation of a semi-orthogonal user selection sub-system. For example, the user selection sub-system may implement a heuristic user selection algorithm based on near-orthogonality. Each user group is assigned a discrete transmission channel, which may be orthogonally defined in terms of frequency, time, or code. Data is transmitted over the channels (e.g., from network base stations) in a coherently coordinated manner, according to a zero-forcing beamforming operation. The system may be configured for operation in a time/frequency selective manner, e.g., over time/frequency selective fading channels. The wireless units may be allocated to the time/frequency slots based on prioritization of channel strength and considerations of fairness, in conjunction with the application of a semi-orthogonal user selection algorithm. | 10-30-2008 |
| 20080299981 | Method of coordinated wireless downlink transmission - To reduce intercell interference, a method of coherently coordinated downlink transmission in a wireless network involves coordinating transmissions from a plurality of base stations to a plurality of wireless units, for coherent, reinforced reception of the transmitted signals at the wireless units. Thus, transmissions from the base stations are coordinated such that signals received at each wireless unit's particular location constructively add, but cancel out at other locations. The signals are generated based on a zero-forcing operation, e.g., by applying zero-forcing complex antenna weight vectors to data symbols designated for transmission to the wireless units. For fairness, the signals are transmitted at no less than a guaranteed common rate. A convex optimization problem (which incorporates a per-base station power constraint or a per-transmission antenna power constraint) is solved to maximize the guaranteed common rate. Dirty paper coding may also be employed. | 12-04-2008 |
