Maximizing the throughput of a network while supporting fairness among nodes is one of the most critical issues in designing wireless networks. In single-hop networks, a lot of schemes have been proposed to satisfy this criterion, and efficient protocols like the IEEE 802.11 and the HiperLAN/2 standards have been established for wireless LAN. In multi-hop wireless networks, however, throughput and fairness have different characteristics from those of single-hop networks. In this paper, the tradeoff between throughput and fairness on multi-hop networks is studied by computer simulation, assuming three node distribution models, namely, normal, constant, and uniform distribution and four different bandwidth (channel) scheduling methods, i.e., first-in first-out buffer based, weighted traffic model based, bandwidth reservation based, and maximum throughput based scheduling. Furthermore, as a realistic model, a hybrid scheme is investigated where partial bandwidth is allocated to the bandwidth reservation based scheduling and the remaining to the maximum throughput based one.

It is well-known that the bit error rate (BER) of digital transmission through a mobile/portable communication channel is strongly dependent on the multipath delay spread. In this paper, we propose an antenna pattern diversity reception system with a new selection strategy based on the measurement of cross-channel-interference which has been confirmed to be a linear function of the multipath rms delay spread assuming BPSK as an example. The coherence bandwidth and BER performances using such a system are studied theoretically and considerable improvement is observed. In addition, the field experiment using a four-direction antenna in 400 MHz band was made to measure error patterns for an individual antenna pattern, and the proposed system was shown to work well even in a real urban area based on the measured data by computer simulation.