Presentation on theme: "Distributed Algorithms for Mobile Sensor Networks Chelsea Sanders Ben Tullis."— Presentation transcript:
Distributed Algorithms for Mobile Sensor Networks Chelsea Sanders Ben Tullis
What is a Mobile Sensor Network? Sensors are set up in a field and they pass the data collected through a network to a main location. They are used to monitor temperature, sound, vibration, pressure, motion, pollutants, and other things. Motivated by military applications such as battlefield surveillance. Today they are used in the military, industrial process monitoring and control, machine health monitoring, environment and habitat monitoring, healthcare applications, home automation, and traffic control.
Problems with Mobile Sensor Networks Where to deploy mobile sensor networks – Affects the performance and the lifetime of the network. – In unknown or hostile environments, is it is hard to determine a good way to optimally deploy the sensors. How to gather the data efficiently – Lots of communication to one node, the master node. – Since the sensors are either continuously gathering data or gathering often, how much data should be sent back? Loss in communication or failure in sensor – Can cause break in the coverage of the area – Can also cause communication to stop all together if it is set up in communication relay.
How Can Parallel Programming Help? The most efficient algorithms are the ones that are distributed algorithms. Distributed algorithms are designed to run on computers that are constructed from interconnected processors. Typically executed concurrently with the other processors having limited information about what the other parts of the algorithm are doing. It helps to cut down are communication between the processors and helps speed up the calculation time.
Using Parallel Programming to Solve Problems Where to deploy the sensors – A distributed algorithm can help figure out where to deploy sensors efficiently and quickly. – Distributed Self-Spreading Algorithm (DSSA) developed by Yong Zhang and Li Wang (and then later studied by Heo and Varshney).
Distributed Self-Spreading Algorithm First, a specified number of nodes are pre- deployed randomly in an area. Each node has a sensing range, communication range, and its initial location. Each node has to be able to communicate to: – Find what nodes are around it and the locations of the nodes around it. – Transmit and forward sensed data.
Distributed Self-Spreading Algorithm Next each node finds its expected density which is the average number of nodes required to cover the entire area and the initial local density which is the number of nodes within is communication range. – Formula: Where N is the number of nodes and cR is the communication range of each node, and A is the range of interest. It then uses a force function which takes into consideration how much force it would take to move to a location and then adds all the partial forces together. – Formula:
Distributed Self-Spreading Algorithm With that information, each node can decide its next movement. It then recollects the information to find the local density, uses the force function again, and decides its next movement. It keeps repeating that process until the nodes move less than a specified amount for a certain amount of moves or the node is just moving back and forth between two places, the movement is stopped and it is considered in a stable place.
Using Parallel Programming to Solve Problems Communication of data – A distributed algorithm can help figure out how much data to send to the master node and the best way to route it to the master node. – Zhao and Yang created a distributed algorithm, the Distributed Self-Spreading Algorithm, that focuses primarily on a sensor network that has master node that actually travels to specific anchor points to collect the data.
Distributed Self-Spreading Algorithm The algorithm allows the sensor nodes to be collecting data constantly and then summarizing it using algorithms. Then it sends its data to an anchor point every so often (as determined by the algorithm), and the master mobile node goes around and collects the data from the anchor points.
Using Parallel Programming to Solve Problems Loss in communication or failure in sensor – Lee and Lin suggest several situations. – The first is dynamic coverage with migration of redundant nodes. – The second is loss of a node is handled by increasing the radii of the nodes nearest to the lost area until the area is completely covered again.
Parallel Solution In both cases parallel programming quickly and efficiently regains the lost coverage by dynamically maintaining the worst case coverage distance
Worst case coverage distance Maximum breach path: a path where the minimum distance from points on the path to the sensor network is maximized.
Advantages Low communication complexity No need for a tight bound on message propagation delay
BlueCube Construction of a hypercube that incorporates communication environment over Bluetooth radio systems. The algorithms that are used focus on the already existing algorithms for hypercube and then shows how they can be applied to Bluetooth communication.
Hyper cube backbone Hypercube Backbone optimizes communication path While Bluetooth communication provides higher communication speed
Conclusion Mobile sensor networks are very widely used and are all around us. They are very important in our everyday life, but they do have some problems. Distributed algorithms have helped improve upon those problems and helped make mobile sensor networks more efficient and effective.
Works Cited Chao-Tsun Chang; Chih-Yung Chang; Jang-Ping Sheu;, "BlueCube: constructing a hypercube parallel computing and communication environment over Bluetooth radio system," Parallel Processing, 2003. Proceedings. 2003 International Conference on, vol., no., pp.447-454, 9-9 Oct. 2003 Heo, N.; Varshney, P.K.;, "A distributed self spreading algorithm for mobile wireless sensor networks," Wireless Communications and Networking, 2003. WCNC 2003. 2003 IEEE, vol.3, no., pp.1597-1602 vol.3, 20-20 March 2003 Miao Zhao; Yuanyuan Yang;, "An Optimization Based Distributed Algorithm for Mobile Data Gathering in Wireless Sensor Networks," INFOCOM, 2010 Proceedings IEEE, vol., no., pp.1-5, 14-19 March 2010 Yong Zhang; Li Wang;, "A distributed sensor deployment algorithm of mobile sensor network," Intelligent Control and Automation (WCICA), 2010 8th World Congress on, vol., no., pp.6963-6968, 7-9 July 2010