Rajat Prakash

Adaptive hard handoff algorithms

The design of hard handoff algorithms based on optimizing the tradeoff between link quality and rate of handoffs is considered. For handoff algorithms based on this criterion, adaptation is precisely defined in terms of remaining on a locus of desirable operating points as system parameters (such as mobile velocity) change. A rule based on a linear cost criterion is used to select desirable operating points. For this rule, it is shown that the optimal handoff algorithm, which is impractical, is easily adapted by fixing a single tradeoff parameter at an appropriate value. The same adaptation property is shown to hold for an easily implementable approximation to the optimal algorithm, the locally optimal (LO) handoff algorithm. This is in contrast to the poor adaptation of hysteresis based approaches which require lookup tables for adaptation. Practical estimators for all relevant system parameters based on a short window of pilot signal strength measurements are also discussed. It is shown that the LO algorithm adapts well when these simple estimators are used. A hysteresis-threshold approximation to the adaptive LO algorithm is also developed.

Locally optimal soft handoff algorithms

The design of soft handoff algorithms for cellular radio systems is considered. The design problem is posed as a tradeoff between three metrics: the rate of handoffs, the mean size of the active set, and the link quality. It is argued that the algorithm that optimizes the tradeoff among these metrics is impractical. Hence, a locally optimal (LO) handoff algorithm is derived as a practical approximation to the optimal handoff algorithm. The LO algorithm is shown to yield a significantly better tradeoff than the static threshold handoff algorithm used in second-generation code-division multiple-access (CDMA) systems. It is also shown that the dynamic threshold algorithm, which is an ad hoc algorithm proposed for third-generation CDMA systems, achieves nearly the same performance as the LO algorithm. Thus, an analytical justification is developed for the dynamic threshold algorithm. Further, the handoff algorithm design is separated into independent design problems on the forward and reverse links. The forward link LO algorithm is shown to be computationally intensive but is also shown to be closely approximated by the simpler reverse link LO algorithm.

Accurate performance analysis of hard handoff algorithms

Previous work on estimating the performance of hard handoff algorithms has been based on Monte-Carlo simulations or asymptotic approximations. A new analysis technique for evaluating the performance of handoff algorithms is introduced. This technique may be used to accurately estimate the performance of both optimal and suboptimal handoff algorithms. Specifically, the technique allows for the calculation of an upper bound on handoff performance without the need for storing or simulating the complicated optimal solution.

A time-scale separation technique for the analysis of random access systems with incremental redundancy

A time-scale separation technique is developed for the analysis of multi-access packet data systems with a varying number of users. User arrival and departure processes are assumed to occur on a time-scale that is slow compared with the time-scale over which a packet is served. Time-scale separation is shown to considerably simplify system analysis by separating the queueing theoretic and the physical layer aspects of the multi-access system.

Locally optimal soft handoff algorithm

The performance of soft handoff algorithms is studied in terms of the tradeoff between three metrics: rate of handoffs, mean active set size and link quality. A locally optimal (LO) handoff algorithm is derived as an approximation to the optimal handoff algorithm. The LO algorithm is shown to yield a significantly better tradeoff than the static threshold handoff algorithm. A simple dynamic threshold handoff algorithm is shown to be an approximation to the LO algorithm, and result in performance only slightly inferior to the LO algorithm.

Traffic load based reverse link power allocation for cellular packet data systems

Power control typically constitutes the adaptation of transmit power to fading. It is argued that in packet data systems, there is another important aspect of power control, and that is the adaptation of the transmit power to the traffic load. The traffic load is defined as the number of users who wish to transmit data to the base station, and this traffic load changes with time. Adaptation of the reverse link transmit power to the traffic load is shown to improve the power-delay tradeoff. The optimal power control policy is computed using dynamic programming, and is compared with some simple suboptimal power control schemes.

Analysis of code division random multiple access systems with packet combining

A reverse link random access system is considered where CDMA with random spreading is used for reset separation. The receiver consists of either a matched filter (MF) or a minimum mean squared error (MMSE) detector followed by autonomous forward error correction (FEC) decoders for each user. The random access strategy combines slotted ALOHA with incremental redundancy (IR). Such a system is defined as a code division random multiple access (CDRMA) system. Two types of IR, namely code combining and maximal ratio combining (MRC) are considered. Bounds on the throughput of a CDRMA system are obtained for different detectors and IR schemes, when the number of users K and the spreading factor N, are both large (K,N/spl rarr//spl infin/, K/N=/spl alpha/). These bounds are derived using known results on the information theoretic capacity for a user within a slot. The bound on the throughput of a CDRMA system is shown to be equal to the bound on the throughput of an equivalent fixed access (conventional) CDMA system.

The impact of service rate fluctuations in wireless packet data systems

Following a model similar to that of Telatar and Gallager, multiuser communication systems with a varying number of users are studied. It is shown that fading causes variability in the service process, and reduces the effective service rate seen by a user. This result is used to design a power control scheme that balances the twin goals of low service variability and high mean service rate.

Reverse link analysis of cellular packet data networks with multiple receive antennas

The reverse link of a wireless packet data system with a varying number of users is considered. A time scale separation approximation is used to justify the analysis of this system based on a processor sharing model, and to compute the tradeoff between the offered load and the throughput seen by a typical user. When multiple receive antennas are available at the base station, it is shown that simple random multiaccess schemes with very little coordination between users can outperform more complex orthogonal schemes.