Power-control for Multimode Transceivers

The performance of a power-control algorithm [7] suitable for multi-mode transceivers is investigated using 1, 2 and 4 bit/symbol modems. It is shown that the algorithm is suitable for maintaining a target frame error rate, irrespective of the modulation mode employed. The transceiver parameters are summarised in Table 1, while the minimum and maximum required average transmitted powers in the cell for the di erent modes are given in Table 4. 1. MOTIVATION Intelligent adaptive multi-mode transceivers or software radios [1],[3]-[6] are intensively studied under the auspices of a number of initiatives, such as for example the European Union's ACTS programme. In this programme we have been investigating adaptive modulation schemes [3, 6], which can be re-con gured to operate using di erent number of bits per modulation symbol, in order to exploit the timevariant channel capacity of fading channels. In order to bene t from the arising variable bitrate, we also investigated the feasibility of invoking the programmable-rate H.263 standard video codec and contrive adaptive packetisers, allowing us to design an integrated multimode video transceiver [7, 8]. Furthermore, an e cient power-control scheme was required, which could support the operation of di erent modulation modes and the speci c requirements imposed by the video codec. 2. POWER CONTROL The multi-mode video transceiver and the power-control algorithm used were described in Reference [7], here only the rationale behind the scheme is stated. The transceiver parameters are summarised in Table 1. The power-control technique is essentially a bit error rate (BER) based algorithm, which ensures a reliable channel quality estimation. This estimate is based on the binary BCH channel codec's error rate monitoring and over-load detection capability. The BCH codecs employed in the di erent modemmodes were also summarised in Table 1. The algorithm reduces the transmitted power, if the number of transmission errors is well below the BCH codec's error correction capability and hence the BCH codec's error correction power is not fully exploited, while frequent BCH code over-load conditions indicate that the transmitted power has to be incremented. The power VTC'97, PHOENIX, USA, MOBILE COMMUNICATIONS SESSION control step-size of the algorithm was also adjustable. For a detailed set of algorithmic parameters and a ow-chart the interested reader is referred to [7]. Let us now concentrate on the multi-mode performance of the algorithm in the next Section. 3. PERFORMANCE OF MULTI-MODE POWER CONTROL The proposed multimode video transceiver of Table 1 using our power control algorithm was simulated and the worstcase scenario of a single interfer was employed in order to generate co-channel interference. The transmission frame error rate (FER) versus user distance and interferer distance from the base station (BS) is shown in Figure 1, when using 4-QAM and no power control. Since our H.263-based video transceiver exhibited a near-unimpaired perceptual video quality at a transmission FER of 5 %, we con gured the power control scheme to maintain this target FER. Given the transceiver parameters used in this experiment, the FER is lower than 5 % over most of the cell area, but it is above this threshold for the worst-case combinations of user and interferer distances, as demonstrated by the Figure. Observe also that when the interferer is at a distance of +200 m from its BS and hence the farthest from the serving BS, a maximum of 5 % FER is maintained for all user distances within the cell. By contrast, for an interferer distance of -200 m, when the interferer is closest to the serving BS, the maximum acceptable user distance is about 140 m. In other words, over the majority of the cell a better than required FER is maintained at the cost of a high transmitted power, while in certain cell areas the FER performance is inadequate. In Figures 2(a) and 2(b) we displayed two operational scenarios for the best and worst case situations, in which the power-control scheme was used. BPSK mode was used and the best case was, when the interferer was as far from the serving BS, as possible, ie at +200 m from its own BS, while the user was as close to its serving BSs, as possible, ie at 0 m. By contrast, the worst case is when they are both at the edge of their cells, with the interferer in its own cell, but as close to the serving BS, as possible. Explicitly, Figures 2(a) and 2(b) display the power-controlled transmission power, slow fading envelope, the signal-to-interference+noise-ratio (SINR) averaged over a timeslot and the Frame Error Flag (FEF) as a function of time for the best and worst case interferer and user positions, respectively. Observe that in the best-case situation the transmitted power is close to its minimum of -34 dBm, while the worst-case example requires a substantially increased transmitted power. The