Power Control in Wireless Ad Hoc Networks Background An ad hoc network is a group of self configuring wireless nodes that lack infrastructure. Motivation—Power Control There is no arbiter to regulate the power of each node. When determining their power level, nodes face the too much vs. too little dilemma. Too much – introduces too much interference into the network Too little – may not be enough power to transmit successfully Most nodes involved in wireless ad hoc networks have a limited or finite power supply. System Description Variable-length packets Channel Access Protocol RTS/CTS/DATA/ACK AB RTS ACK CTS Signal vs. Interference In addition to a node transmitting a signal or data to another node, it also transmits interference that is received by other nodes in the network. This interference may harm their chance of receiving a good signal. A DC B EINR Energy to Interference plus Noise Ratio where and Received Energy Spreading Factor InterferenceNoise path loss formula Power Schemes Fixed Power Each node uses a fixed power, P max No calculations Used for RTS, CTS, and DATA transmissions Simple Adaptive Power Simple calculation Used for RTS, CTS, and DATA transmissions interference margin depends on distance 3.6W Because the Simple Adaptive Power Scheme only depends on the distance between two nodes and fails to take into account the amount of interference in the network, the transmissions in the above network will fail because of excessive interference in the network. Genie-Aided Adaptive Power Actual interference is used from the environment. The power is iteratively adjusted based on these values such that the final power is the minimum amount required for a successful transmission. The power is updated every time the interference environment changes. Used for DATA transmissions D A B C ACK CTS RTS CTS RTS * ** DATA * Unlike the Simple Adaptive power scheme which calculates the power only at the beginning of transmission (*), the genie-aided power scheme also re-calculates the transmit power whenever the interference changes (*). Simulation Model Simulated in C. 100 nodes placed at random locations Each node contains a network layer which has a queue. First in First Out Maximum limit of 50 packets The network layer uses Dijkstra’s algorithm to determine the routes with the fewest number of relays or hops. Packet Generation Each node generates a packet in a slot with a probability of p. The destination for this node is randomly chosen among the other nodes in the network Transmission Power (P max ) Average Diameter 1.0W9.9 hops 1.4W8.5 hops 2.2W7.2 hops 7.1W4.9 hops 41.5W3.0 hops Throughput Results Number of packets per slot that are received at their destination Throughput for Fixed Power Throughput for Simple Adaptive Power Throughput for Genie-Aided Power Conclusion Simple Adaptive Power Scheme better than Fixed Power Scheme Approximately same throughput 200% increase in throughput efficiency Genie-Aided Power Scheme Significant improvement in throughput over Simple Adaptive Power Scheme Some increase in throughput efficiency Impractical to implement because you would need to know the power level of every node and path loss of every link in the network to determine the actual interference value. Future Work Modify Simple Adaptive Power Scheme in order to improve throughput Crystal Jackson, Clemson University Advisor: Dr. Harlan Russell A D C B Diameter: distance between two nodes that are furthest from each other in the network Distance = 1.0 Distance = W 3.6W 21.9W 16.7W Distance = 0.7 calculated actual needed interference 19.5W Throughput Results Number of packets received at their destination per unit of energy