Projects

 

 

  1. Optimal resource reservation for handoffs in cellular systems

 

There is a fundamental tradeoff between "call blocking probability" and "call dropping probability" in cellular systems. While both blocked calls and dropped calls are undesirable, dropped calls are annoying and must be eliminated if possible. Call dropping probability can be decreased at the cost of increased call blocking probability. For this, some resources are reserved in each cell for calls that are transferred from neighboring cells. In this project, OPNET simulations were used to determine optimal resource that need to be reserved. Two different scenarios were modeled - Metropolitan model and Highway model. In metropolitan model, people were assumed to be moving in random directions at various speeds - walking, vehicular. The highway model represents cars moving on a typical highway. It was determined that fixed resource allocation does not perform very well. Resource reservation schemes that take into account the speed and general direction of movement of callers perform better.

 

 

  1. Performance comparison of DSR and TORA adhoc routing protocols

 

The objective of this project was to study and compare the performance of two MANET routing protocols: Temporarily Ordered Routing Algorithm (TORA) and Dynamic Source Routing (DSR). The simulation model consists of a set of nodes moving in random directions at variable speeds. Number of nodes and average speed are the variables that affect the protocol performance. Results show that DSR outperforms TORA in terms of the number of MANET control packets used to maintain/erase routes. Also, it is concluded that TORA is a better choice than DSR for fast moving highly connected set of nodes. It is observed that DSR provides better data throughput than TORA and that routes can be created faster in DSR than in TORA.

 

 

  1. Performance of 802.11b Wireless LAN protocol in lightly loaded networks

 

This project involves modification of standard 802.11b protocol for improved performance in lightly loaded conditions. An event based simulation program written in C is used to model a standard 802.11b network in adhoc configuration. Varying the Number of Stations, Packet Size and Arrival Rate changes network load. The effect of these variations on parameters like data throughput and packet delay is studied and the protocol is modified for lightly loaded networks. For lightly loaded networks with small sized packets, a reduction in minimum contention window produces favorable results. But on the other hand, small contention windows are not very helpful in lightly loaded networks with long packets.

 

 

  1. Quality of Service Issues in 3G Wireless Networks

 

Literature survey of QoS issues in 3G wireless Networks.

 

 

  1. Performance of TCP over RLC/MAC of UMTS

 

Literature search on TCP over wireless networks:  TCP Snoop, M-TCP etc. and UMTS RLC/MAC operation.

 

 

 

  1. Multiuser detection techniques for synchronous CDMA

 

The inadequacy of conventional receivers in Multiple Access Interference (MAI) environment spurs the research in multiuser receiver architectures. Multiuser receivers mitigate MAI, improve accuracy of the signal estimate, and therefore result in reduced bit error rates. In this project we investigate the various multiuser receivers namely: Optimal, Decorrelator, Minimum mean square error, Parallel interference cancellation and Successive interference cancellation receivers. MATLAB simulations are used to compare the performance of these receivers with conventional receiver. 

 

 

  1. Cellular Path loss simulator in MATLAB

 

The simulation program uses transmitter power, radio frequency, path loss exponent and standard deviation of log normal distribution as inputs to generate samples of propagation path loss. It also uses receiver antenna gain, receiver bandwidth and noise figure to determine noise power at receiver. This information is used to determine the percentage of time the SNR at the receiver is greater than minimum. This is used to determine cell size and minimum separation between cells. 

 

 

  1. Design and Implementation of concurrent HTTP 1.1 Server

 

WINSOCK programming was used to implement the server. The server supports GET/POST methods and persistent connections.

 

  1. Performance Comparison of real time scheduling algorithms – DASA, EDF and RMA

 

DASA, EDF and RMA were implemented using uClinux/RTlinux on a Lieno uCsimm kit (that uses motorola’s Dragonball processor). It was shown that even though DASA should theoretically perform best in overload conditions, it performs worse than RMA because of the processing overhead involved.