Self-Organized Sensor Network with Optimized Organization and Communication

OCO: OPTIMIZED COMMUNICATION AND ORGANIZING FOR TARGET TRACKING IN WIRELESS SENSOR NETWORK

1. Introduction

Wireless sensor networks have significant impact to the efficiency of military and civil applications such as environment monitoring, target surveillance, industrial process observation, and tactical systems. In which, target tracking is one of the most important applications of wireless sensor networks.

So, there happen a demand of self-organizing and routing in the network. However, the current computer network protocols could not apply to sensor network because sensor nodes are constrained in energy supply, performance, and band-width. Hence, optimized computation and energy dissipation are the highest requirements to maximize lifetime of the sensor network. Existing methods, however, suffer redundant in data and sensor node deployment, or require complex computation on sensor node. Those contribute to use energy inefficient or request complex calculation for sensor nodes.

In this paper, we devise and evaluate a multiple target tracking method, called OCO (Optimized Communication and Organizing), that ensures equal accuracy as when all nodes are turn on. However, energy dissipation is many times less than LEACH-based algorithm. Especially, OCO demands very low computation on sensor nodes.

2. Survey of existing methods a. Existing routing protocols: (http://www.cs.umbc.edu/~kemal1/mypapers/Akkaya_Youni s_JoAdHocRevised.pdf , http://vulcan.ee.iastate.edu/~kamal/Docs/kk04.pdf)

i. Data-centric protocols:

o Direct communication:

 How (one-hop communication): Change power of the node transmitter to send the message directly to destination nodes.  Pros: - Simple.  Cons: - Waste energy. - Need auto-calibration for the transmitter. - Size of the monitoring zone is small. o Flooding:

 How: Broadcast data to all neighbors. The neighbors, then, continue broadcast the received data to all their neighbors until the data reach the destination or run out the message time-to- live.  Pros: - Simple.  Cons: - Data redundancy. o Gossip:(S. Hedetniemi and A. Liestman, “A survey of gossiping and broadcasting in communication networks,” Networks, Vol. 18, No. 4, pp. 319-349, 1988)

 How: Try to reduce the data redundancy by sending data to one randomly selected neighbor.

 Pros: - Reduce data overlapping compare to flooding method.  Cons: - Data is still overlapped - Delay. - Data is not guaranteed to reach the destination. - Inefficient energy consumption. o SPIN (SENSOR PROTOCOL FOR INFORMATION VIA NEGOTIATION):(W. Heinzelman, J. Kulik, and H. Balakrishnan, “Adaptive protocols for information dissemination in wireless sensor networks,” in the Proceedings of the 5th Annual ACM/IEEE International Conference on Mobile Computing and Networking (MobiCom’99), Seattle, WA, August 1999)

 How: (Figure 1) The idea is to use meta-data (a short description of data  small size). Before transmission, meta-data are broadcasted to all neighbors by advertisement message (ADV). Interested neighbors (those who do not have the data) retrieve the data by sending back a request message (REQ). The node only sends the real data to the interested neighbors. Figure 1: SPIN phases  Pros: - Consume energy 3.5 times less than flooding.  Cons: - How to make a meta-data? - ADV messages (meta-data) are redundant. - Need meta-data caches at each node to decide which new data should be retrieved. - Cannot guarantee the delivery of data because if there is at least a node between the source and the destination that is not interested in that data, such data will not be delivered to the destination. o Directed Diffusion: (C. Intanagonwiwat, R. Govindan and D. Estrin, "Directed diffusion: A scalable and robust communication paradigm for sensor networks", in the Proceedings of the 6th Annual ACM/IEEE International Conference on Mobile Computing and Networking (MobiCom'00), Boston, MA, August 2000)

 How: (Used to query information) When the base need information, it generates a message, called INTEREST, including all attribute-value pairs. For example: “Select all positions having temperature greater than 30”. Then, the message is sent to all neighbors. At each node, when receiving the message, it considers if the message is in the cache or not. If the message is not in the cache, cache the message with information (such as: timestamp, data rate, duration, expiration, and etc). The reply path is the neighbor from which the message was received (gradient). Then, the message is continue forwarded to all neighbors. In the case the message is already in the cache, the cache information (timestamp, data rate, and etc) is compared with those values of the message to select the best reply path. Finally, if the sensing information meets the query (temperature greater than 30), the answer is sent through the reply path to reach the base. (Figure 2) Figure 2: Directed diffusion phases  Pros: - No need for maintaining global network topology. - Suitable for query applications.  Cons: - On demand routing set up  Inefficient for continuous data application as environment monitoring. - More computation (caching and comparison) requirement for nodes. - Extra overhead for data matching and queries. o Energy-aware routing: (R. Shah and J. Rabaey, "Energy Aware Routing for Low Energy Ad Hoc Sensor Networks", in the Proceedings of the IEEE Wireless Communications and Networking Conference (WCNC), Orlando, FL, March 2002)

 How: Similar to directed diffusion in the way potential paths from data sources to the base are discovered. However, network survivability is the main metric of the approach. So, the cost each path is calculated and stored in the routing table. At each time, the reply path is selected randomly from the routing table based on the cost of path (the higher cost path has less probability of selection).

 Pros: - Increase 44% of network lifetime compare with directed diffusion.

 Cons: - - On demand routing set up  Inefficient for continuous data application as environment monitoring. - More computation demands on node than directed diffusion o Rumor routing:(D. Braginsky and D. Estrin, "Rumor Routing Algorithm for Sensor Networks," in the Proceedings of the First Workshop on Sensor Networks and Applications (WSNA), Atlanta, GA, October 2002)

 How: Rumor is another variation of Directed Diffusion. It argues that flooding query to all nodes as in directed diffusion is inefficient because replying data is small. Its approach is alternative, flooding the events if number of events is small and number of queries is large. To flood events through the network, the rumor routing algorithm employs long lived packets, called agents. When a node detects an event, it adds such event to its local table and generates an agent. Agents travel the network in order to propagate information about local events to distant nodes. When a node generates a query for an event, the nodes that know the route, can respond to the query by referring its event table (Note: The idea of routing query is likely of directed diffusion). Hence, the cost of flooding the whole network is avoided.

 Pros:

- Save energy significantly compare to directed diffusion when number of events is small.  Cons: - Inefficient if number of events is large. - Hard to tune time-to-live for queries and agents to prevent overhead. - On demand routing set up  Inefficient for continuous data application as environment monitoring. o Gradient-based routing:(C. Schurgers and M.B. Srivastava, “Energy efficient routing in wireless sensor networks,” in the MILCOM Proceedings on Communications for Network-Centric Operations: Creating the Information Force, McLean, VA, 2001)

 How: It is another version of Directed Diffusion. The idea is to keep the number of hops when the interest is diffused through the network. Hence, each node can discover the minimum number of hops to the base, which is called height of the node. When there are two or more next hops with the same height, the node chooses one of them at random. When a node’s energy drops below a certain threshold, it increases its height so that other sensors are discouraged from sending data to that node.

 Pros: - Save energy compare to Directed Diffusion. - Increase network lifetime.  Cons: - Nodes have to consider height of their neighbors before sending  exchange asking height messages. So, when number of neighbors is large  increase the exchange messages  inefficient. o Information-driven sensor querying (IDSQ) and Constrained anisotropic diffusion routing (CADR): ( M. Chu, H. Haussecker, and F. Zhao, "Scalable Information-Driven Sensor Querying and Routing for ad hoc Heterogeneous Sensor Networks," The International Journal of High Performance Computing Applications, Vol. 16, No. 3, August 2002 - http://www2.parc.com/spl/members/zhao/stanford- cs428/readings/CollaborativeProcessing/zhao_idsq_200 2.pdf)

 How: Generalized form of Directed Diffusion. Two techniques namely information-driven sensor querying (IDSQ) and constrained anisotropic diffusion routing (CADR) are proposed. The idea is to query sensors and route data in a network in order to maximize the information gain, while minimizing the latency and bandwidth. This is achieved by activating only the sensors that are close to a particular event and dynamically adjusting data routes. The major difference from Directed Diffusion is the consideration of information gain in addition to the communication cost. In CADR, each node evaluates an information/cost objective and routes data based on the local information/cost gradient and end-user requirements. The information utility measure is modeled using standard estimation theory. IDSQ is based on a protocol in which the querying node can determine which node can provide the most useful information while balancing the energy cost. While IDSQ provides a way of selecting the optimal order of sensors for maximum incremental information gain, it does not specifically define how the query and the information are routed between sensors and the sink. Therefore, IDSQ can be seen as a complementary optimization procedure.

 Pros:

- More energy efficient than Directed Diffusion.  Cons: - Too much calculation demand for nodes. o Cougar: (Y. Yao and J. Gehrke, “The cougar approach to in-network query processing in sensor networks,” in SIGMOD Record, September 2002)  How: The main idea is to use declarative queries in order to abstract query processing from the network layer functions such as selection of relevant sensors etc. and utilize in-network data aggregation to save energy. The abstraction is supported through a new query layer between the network and application layers. COUGAR proposes architecture for the sensor database system where sensor nodes select a leader node to perform aggregation and transmit the data to the gateway (sink). The architecture is depicted in Figure 3. The gateway is responsible for generating a query plan, which specifies the necessary information about the data flow and in-network computation for the incoming query and send it to the relevant nodes. The query plan also describes how to select a leader for the query. The architecture provides in-network computation ability for all the sensor nodes. Such ability ensures energy efficiency especially when the number of sensors generating and sending data to the leader is huge.

Figure 3: Query plan at a leader node: The leader node gets all the readings, calculates the average and if it is greater than a threshold sends it to the gateway (sink).

 Pros: - Could save energy because the aggregation.  Cons: - Add new layer to node (query layer) -> overhead, fault tolerant problems. - Need synchronization at leader node (wait for packets from nodes before aggregation). - Aggregation algorithm?

o Acquire:(ACtive QUery forwarding In sensoR networks - N. Sadagopan et al., “The ACQUIRE mechanism for efficient querying in sensor networks,” in the Proceedings of the First International Workshop on Sensor Network Protocol and Applications, Anchorage, Alaska, May 2003)

 How: Similar to COUGAR, ACQUIRE views the network as a distributed database where complex queries can be further divided into several sub queries. The operation of ACQUIRE can be described as follows. The BS node sends a query, which is then forwarded by each node receiving the query. During this, each node tries to respond to the query partially by using its pre- cached information and then forward it to another sensor node. If the pre-cached information is not up-to-date, the nodes gather information from their neighbors within a look- ahead of d hops. Once the query is being resolved completely, it is sent back through either the reverse or shortest-path to the BS. Hence, ACQUIRE can deal with complex queries by allowing many nodes to send responses. Note that directed diffusion may not be used for complex queries due to energy considerations as directed diffusion also uses flooding-based query mechanism for continuous and aggregate queries. On the other hand, ACQUIRE can provide efficient querying by adjusting the value of the look-ahead parameter d. When d is equal to network diameter, ACQUIRE mechanism behaves similar to flooding. However, the query has to travel more hops if d is too small.

 Pros:

- Solve complex queries.  Cons: - No validation result for energy efficiency. ii. Hierarchical protocols o LEACH: (W. Heinzelman, A. Chandrakasan and H. Balakrishnan, "Energy-Efficient Communication Protocol for Wireless Micro-sensor Networks," Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS '00), January 2000)  How: There are 2 phases: set-up phase and steady-phase. Set-up phase: Sensors may elect themselves to be a local cluster head at any time with a certain probability (reason: to balance the energy dissipation). A sensor node chooses a random number between 0 and 1. If this random number is less than the threshold T (optimal is 5%), the sensor node becomes a cluster-head.

After the cluster heads are selected, the cluster heads advertise to all sensor nodes in the network that they are the new cluster heads. Each node accesses the network through the cluster head that requires minimum energy to reach. Once the nodes receive the advertisements, they decide which head they belong to. The, the nodes inform the appropriate cluster heads that they will be member of the cluster. Finally, the cluster heads assign the time slot on which the sensor nodes can send data to them.

Steady-phase: Sensors begin to sense and transmit data to the cluster heads which aggregate data from the nodes in their clusters. After a certain period of time spent on the steady state, the network goes into start-up phase again and enters another round of selecting cluster heads.

 Pros: - Save energy. - Increase network lifetime.  Cons: - LEACH assumes that all nodes have enough power to direct communicate with the base  Can not apply for large areas. - LEACH cannot show how a node know if the numbers of cluster heads reach the threshold. - Some area has no cluster head. - Cluster heads directly communicate with the base  Need too many channels. - Aggregation algorithm? o Power-Efficient Gathering in Sensor Information Systems (PEGASIS): (S. Lindsey and C. S Raghavendra, "PEGASIS: Power Efficient GAthering in Sensor Information Systems," in the Proceedings of the IEEE Aerospace Conference, Big Sky, Montana, March 2002)

 How: An enhancement over LEACH. The basic idea of the protocol is that in order to extend network lifetime, nodes need only communicate with their closest neighbors and they take turns in communicating with the base-station. When the round of all nodes communicating with the base-station ends, a new round will start and so on. To locate the closest neighbor node in PEGASIS, each node uses the signal strength to measure the distance to all neighboring nodes, then, adjust the signal strength so that only one node can be heard. The chain in PEGASIS will consist of those nodes that are closest to each other and form a path to the base-station. The aggregated form of the data will be sent to the base-station by any node in the chain and the nodes in the chain will take turns in sending to the base-station. The chain construction is performed in a greedy fashion.

 Pros: - Elimination of the overhead caused by dynamic cluster formation in LEACH.

- Simulation results showed that PEGASIS is able to increase the lifetime of the network twice as much the lifetime of the network under the LEACH protocol.

 Cons: - Need time and energy to estimate the closest node. - PEGASIS introduces excessive delay for distant node on the chain. - The single leader can become a bottleneck. - Require dynamic topology adjustment since a sensor node needs to know about energy status of its neighbors in order to know where to route its data. - PEGASIS assumes that all nodes have enough power to direct communicate with the base  Can not apply for large areas. o Threshold-sensitive Energy Efficient Protocols (TEEN): (A. Manjeshwar and D. P. Agarwal, "TEEN: a routing protocol for enhanced efficiency in wireless sensor networks," In 1st International Workshop on Parallel and Distributed Computing Issues in Wireless Networks and Mobile Computing, April 2001)

 How: TEEN pursues a hierarchical approach along with the use of a data-centric mechanism. The sensor network architecture is based on a hierarchical grouping where closer nodes form clusters and this process goes on the second level until base station (sink) is reached. The model is depicted in Figure 4. After the clusters are formed, the cluster head broadcasts two thresholds to the nodes. These are hard and soft thresholds for sensed attributes. Hard threshold is the minimum possible value of an attribute to trigger a sensor node to switch on its transmitter and transmit to the cluster head. Thus, the hard threshold allows the nodes to transmit only when the sensed attribute is in the range of interest, thus reducing the number of transmissions significantly. Once a node senses a value at or beyond the hard threshold, it transmits data only when the value of the attribute changes by an amount equal to or greater than the soft threshold. As a consequence, soft threshold will further reduce the number of transmissions if there is little or no change in the value of sensed attribute. One can adjust both hard and soft threshold values in order to control the number of packet transmissions.

Figure 4: Hierarchical Clustering in TEEN/APTEEN

 Pros: - Soft threshold will further reduce the number of transmissions if there is little or no change in the value of sensed attribute.  Cons: - TEEN is not good for applications where periodic reports are needed since the user may not get any data at all if the thresholds are not reached. - Overhead and complexity of forming clusters in multiple levels, implementing threshold-based functions and dealing with attribute-based naming of queries. o Adaptive Threshold sensitive Energy Efficient sensor Network protocol (APTEEN): (A. Manjeshwar and D. P. Agarwal, "APTEEN: A hybrid protocol for efficient routing and comprehensive information retrieval in wireless sensor networks," Parallel and Distributed Processing Symposium., Proceedings International, IPDPS 2002, pp. 195-202)

 How: APTEEN is an extension to TEEN and aims at both capturing periodic data collections and reacting to time-critical events. The architecture is same as in TEEN. When the base station forms the clusters, the cluster heads broadcast the attributes, the threshold values, and the transmission schedule to all nodes. Cluster heads also perform data aggregation in order to save energy. APTEEN supports three different query types: historical, to analyze past data values; one-time, to take a snapshot view of the network; and persistent to monitor an event for a period of time.

 Pros: - Better than TEEN  Cons: - Overhead and complexity of forming clusters in multiple levels, implementing threshold-based functions and dealing with attribute-based naming of queries. - Aggregation algorithm? o Energy-aware routing for cluster-based sensor networks: ( M. Younis, M. Youssef and K. Arisha, “Energy-Aware Routing in Cluster-Based Sensor Networks”, in the Proceedings of the 10th IEEE/ACM International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS2002), Fort Worth, TX, October 2002)

 How: Sensors are grouped into clusters prior to network operation. The algorithm employs cluster heads, namely gateways, which are less energy constrained than sensors and assumed to know the location of sensor nodes. Gateways maintain the states of the sensors and sets up multi-hop routes for collecting sensors’ data. The base communicates only with the gateways. The routing is based on the cost (energy consumption, delay, and etc) of links between node and the gateway.

 Pros: - Increase lifetime of the network.  Cons: - Overhead of the cost computation and re-routing. - Two types of sensor nodes: gateway and sensor  Lost generality. o Self-organizing protocol: (L. Subramanian and R. H. Katz, "An Architecture for Building Self Configurable Systems," in the Proceedings of IEEE/ACM Workshop on Mobile Ad Hoc Networking and Computing, Boston, MA, August 2000)

 How: The routing architecture is hierarchical where groups of nodes are formed and merge when needed. In order to support fault tolerance, Local Markov Loops (LML) algorithm, which performs a random walk on spanning trees of a graph, is used in broadcasting. The algorithm for selforganizing the router nodes and creating the routing tables consists of four phases:

• Discovery phase: The nodes in the neighborhood of each sensor are discovered.

• Organization phase: Groups are formed and merged by forming a hierarchy. Each node is allocated an address based on its position in the hierarchy. Routing tables of size O(log N) are created for each node. Broadcast trees that span all the nodes are constructed.

• Maintenance phase: Updating of routing tables and energy levels of nodes is made in this phase. Each node informs the neighbors about its routing table and energy level. LML are used to maintain broadcast trees.

• Self-reorganization phase: In case of partition or node failures, group reorganizations are performed.

 Pros: - Save energy compare to SPIN.  Cons: - Complex computations are required to nodes. iii. Location-based protocols

o Minimum Energy Communication Network (MECN): (V. Rodoplu and T.H. Ming, "Minimum energy mobile wireless networks," IEEE Journal of Selected Areas in Communications, Vol. 17, No. 8, pp. 1333-1344, 1999)

 How: The protocol has two phases:

• It takes the positions of a two dimensional plane and constructs a sparse graph (enclosure graph), which consists of all the enclosures of each transmit node in the graph. This construction requires local computations in the nodes. The enclose graph contains globally optimal links in terms of energy consumption.

• Finds optimal links on the enclosure graph. It uses distributed Belmann-Ford shortest path algorithm with power consumption as the cost metric. In case of mobility the position coordinates are updated using GPS.

 Pros: - MECN is self-reconfiguring and thus can dynamically adapt to nodes failure or the deployment of new sensors  Cons: - Required complex computation for nodes. - The network is assumed to be full connected.

o Small Minimum Energy Communication Network (SMECN): (L. Li and J. Y Halpern, “Minimum energy mobile wireless networks revisited,” in the Proceedings of IEEE International Conference on Communications (ICC’01), Helsinki, Finland, June 2001)

 How: SMECN is an extension to MECN. In MECN, it is assumed that every node can transmit to every other node, which is not possible every time. In SMECN possible obstacles between any pair of nodes are considered. The subnetwork constructed by SMECN for minimum energy relaying is provably smaller (in terms of number of edges) than the one constructed in MECN if broadcasts are able to reach to all nodes in a circular region around the broadcaster. As a result, the number of hops for transmissions will decrease. Simulation results show that SMECN uses less energy than MECN and maintenance cost of the links is less. However, finding a sub- network with smaller number of edges introduces more overhead in the algorithm.

 Pros: - More energy efficient compare to MECN.  Cons - The proposed algorithm is local in the sense that it does not actually find the minimum-energy path, it just constructs a sub-network in which it is guaranteed to exist. - Finding a sub-network with smaller number of edges introduces more overhead in the algorithm. - The network is still assumed to be full connected. o Geographic Adaptive Fidelity (GAF): (Y. Xu, J. Heidemann, and D. Estrin, "Geography-informed energy conservation for ad hoc routing," in the Proceedings of the 7th Annual ACM/IEEE International Conference on Mobile Computing and Networking (MobiCom’01), Rome, Italy, July 2001)

 How: GAF conserves energy by turning off unnecessary nodes in the network without affecting the level of routing fidelity. It forms a virtual grid for the covered area. Each node uses its GPS-indicated location to associate itself with a point in the virtual grid. Nodes associated with the same point on the grid are considered equivalent in terms of the cost of packet routing. Such equivalence is exploited in keeping some nodes located in a particular grid area in sleeping state in order to save energy. Thus, GAF can substantially increase the network lifetime as the number of nodes increases. A sample situation is depicted in Figure 5. In this figure, node 1 can reach any of 2, 3 and 4 and nodes 2, 3, and 4 can reach 5. Therefore nodes 2, 3 and 4 are equivalent and two of them can sleep. Nodes change states from sleeping to active in turn so that the load is balanced. There are three states defined in GAF. These states are discovery, for determining the neighbors in the grid, active reflecting participation in routing and sleep when the radio is turned off. The state transitions in GAF are depicted in Figure 6. Which node will sleep for how long is application dependent and the related parameters are tuned accordingly during the routing process. In order to handle the mobility, each node in the grid estimates its leaving time of grid and sends this to its neighbors. The sleeping neighbors adjust their sleeping time accordingly in order to keep the routing fidelity. Before the leaving time of the active node expires, sleeping nodes wake up and one of them becomes active. GAF is implemented both for non-mobility (GAF-basic) and mobility (GAF-mobility adaptation) of nodes.

Figure 5: Example of virtual grid in GAF

Figure 6: State transitions in GAF

 Pros: - Simulation results show that GAF performs at least as well as a normal ad hoc routing protocol in terms of latency and packet loss and increases the lifetime of the network by saving energy.  Cons: - Considered as a hierarchical protocol without aggregation  Having same weaknesses of hierarchical protocol as talk above. o Geographic and Energy Aware Routing (GEAR): (Y. Yu, D. Estrin, and R. Govindan, “Geographical and Energy- Aware Routing: A Recursive Data Dissemination Protocol for Wireless Sensor Networks,” UCLA Computer Science Department Technical Report, UCLA-CSD TR-01-0023, May 2001)

 How: The idea is to restrict the number of interests in Directed Diffusion by only considering a certain region rather than sending the interests to the whole network. In GEAR, each node keeps an estimated cost and a learning cost of reaching the destination through its neighbors. The estimated cost is a combination of residual energy and distance to destination. The learned cost is a refinement of the estimated cost that accounts for routing around holes in the network. A hole occurs when a node does not have any closer neighbor to the target region than itself. If there are no holes, the estimated cost is equal to the learned cost. The learned cost is propagated one hop back every time a packet reaches the destination so that route setup for next packet will be adjusted. There are two phases in the algorithm: • Forwarding packets towards the target region: Upon receiving a packet, a node checks its neighbors to see if there is one neighbor, which is closer to the target region than itself. If there is more than one, the nearest neighbor to the target region is selected as the next hop. If they are all further than the node itself, this means there is a hole. In this case, one of the neighbors is picked to forward the packet based on the learning cost function. This choice can then be updated according to the convergence of the learned cost during the delivery of packets. • Forwarding the packets within the region: If the packet has reached the region, it can be diffused in that region by either recursive geographic forwarding or restricted flooding. Restricted flooding is good when the sensors are not densely deployed. In high-density networks, recursive geographic flooding is more energy efficient than restricted flooding. In that case, the region is divided into four sub regions and four copies of the packet are created. This splitting and forwarding process continues until the regions with only one node are left. An example is depicted in Figure 7. Figure 7: Recursive Geographic Forwarding in GEAR

 Pros: - Better in term of packet delivery.  Cons: - Complex computation demands for nodes. iv. Network flow and QoS-aware protocols

o Maximum lifetime energy routing: (J.-H. Chang and L. Tassiulas, "Maximum Lifetime Routing in Wireless Sensor Networks," in the Proceedings of the Advanced Telecommunications and Information Distribution Research Program (ATIRP'2000), College Park, MD, March 2000)

 How: The main objective of the approach is to maximize the network lifetime by carefully defining link cost as a function of node remaining energy and the required transmission energy using that link. By using Bellman-Ford shortest path algorithm for the above link costs, the least cost paths to the destination (gateway) are found. The least cost path obtained is the path whose residual energy is largest among all the paths.

 Pros: - Increase lifetime of network  Cons: - Using Bellman-Ford leads to require high-performance nodes when the number of node is large. o Maximum lifetime data gathering:(K. Kalpakis, K. Dasgupta and P. Namjoshi, “Maximum lifetime data gathering and aggregation in wireless sensor networks,” in the Proceedings of IEEE International Conference on Networking (NETWORKS '02), Atlanta, GA, August 2002)

 How: The lifetime “T” of the system is defined as the number of rounds or periodic data readings from sensors until the first sensor dies. The data- gathering schedule specifies for each round how to get and route data to the sink. A schedule has one tree for each round, which is directed from the sink and spans all the nodes in the system. The system lifetime depends on the duration for which the schedule remains valid. The aim is to maximize the lifetime of the schedule. An algorithm called Maximum Lifetime Data Aggregation (MLDA) is proposed. The algorithm considers data aggregation while setting up maximum lifetime routes. In this case, if a schedule “S” with “T” rounds is considered, it induces a flow network G. The flow network with maximum lifetime subject to the energy constraints of sensor nodes is called an optimal admissible flow network. Then, a schedule is constructed by using this admissible flow network.  Pros: - System lifetime is significantly better than hierarchical-PEGASIS.  Cons: - Delay is slightly greater that PEGASIS. - Computation is so complex. o Minimum cost forwarding: (M. Chu, H. Haussecker, and F. Zhao, "Scalable Information-Driven Sensor Querying and Routing for ad hoc Heterogeneous Sensor Networks," The International Journal of High Performance Computing Applications, Vol. 16, No. 3, August 2002)

 How: Minimum cost forwarding protocol aims at finding the minimum cost path in a large sensor network. The cost function for the protocol captures the effect of delay, throughput and energy consumption from any node to the sink. There are two phases in the protocol. First phase is a setup phase for setting the cost value in all nodes. It starts from the sink and diffuses through the network. Every node adjusts its cost value by adding the cost of the node it received the message from and the cost of the link. Such cost adjustment is not done through flooding. Instead, a back-off based algorithm is used in order to limit the number of messages exchanged. The forwarding of message is deferred for a preset duration to allow the message with a minimum cost to arrive. Hence, the algorithm finds optimal cost of all nodes to the sink by using only one message at each node. Once these cost fields are set, there will be no need to keep next hop states for the nodes. This will ensure scalability. In the second phase, the source broadcasts the data to its neighbors. The nodes receiving the broadcast message, adds its transmission cost (to sink) to the cost of the packet. Then the node checks the remaining cost in the packet. If the remaining cost of the packet is not sufficient to reach the sink, the packet is dropped. Otherwise the node forwards the packet to its neighbors.

 Pros: - The protocol does not require any addresses and forwarding paths. - Simulation results show that the cost values for each node obtained by the proposed protocol is same as flooding. - The average number of advertisement messages in flooding could be reduced by a factor of 50 using the back off based algorithm with a proper setting of the back off timer.  Cons: - Computation is expensive. o Sequential Assignment Routing (SAR): (I. F. Akyildiz et al., “Wireless sensor networks: a survey”, Computer Networks, Vol. 38, pp. 393- 422, March 2002 & K. Sohrabi, et al., "Protocols for self-organization of a wireless sensor network,” IEEE Personal Communications, Vol. 7, No. 5, pp. 16-27, October 2000)

 How: the first protocol for sensor networks that includes the notion of QoS in its routing decisions. It is a table-driven multi-path approach striving to achieve energy efficiency and fault tolerance. The SAR protocol creates trees rooted at one-hop neighbors of the sink by taking QoS metric, energy resource on each path and priority level of each packet into consideration. By using created trees, multiple paths from sink to sensors are formed. One of these paths is selected according to the energy resources and QoS on the path. Failure recovery is done by enforcing routing table consistency between upstream and downstream nodes on each path. Any local failure causes an automatic path restoration procedure locally.

 Pros: - Simulation results show that SAR offers less power consumption than the minimum-energy metric algorithm, which focuses only the energy consumption of each packet without considering its priority.  Cons: - SAR maintains multiple paths from nodes to sink. Although, this ensures fault-tolerance and easy recovery, the protocol suffers from the overhead of maintaining the tables and states at each sensor node especially when the number of nodes is huge. o Energy-Aware QoS Routing Protocol: (K. Akkaya and M. Younis, “An Energy-Aware QoS Routing Protocol for Wireless Sensor Networks,” in the Proceedings of the IEEE Workshop on Mobile and Wireless Networks (MWN 2003), Providence, Rhode Island, May 2003)

 How: The proposed protocol extends the routing approach and finds a least cost and energy efficient path that meets certain end-to-end delay during the connection. The link cost used is a function that captures the nodes’ energy reserve, transmission energy, error rate, and other communication parameters. In order to support both best effort and real time traffic at the same time, a class-based queuing model is employed. The queuing model allows service sharing for real-time and non-real-time traffic. The bandwidth ratio r, is defined as an initial value set by the gateway and represents the amount of bandwidth to be dedicated both to the real-time and non-real-time traffic on a particular outgoing link in case of a congestion. As a consequence, the throughput for normal data does not diminish by properly adjusting such “r” value. The queuing model is depicted in Figure 8. The protocol finds a list of least cost paths by using an extended version of Dijkstra’s algorithm and picks a path from that list which meets the end-to-end delay requirement. Figure 8: Query model in a particular sensor node

 Pros: - Simulation results show that the proposed protocol consistently performs well with respect to QoS and energy metrics.  Cons: - How to define r-value for each node. - Computation is so expensive. o SPEED: (T. He et al., “SPEED: A stateless protocol for real-time communication in sensor networks,” in the Proceedings of International Conference on Distributed Computing Systems, Providence, RI, May 2003)

 How: The protocol requires each node to maintain information about its neighbors and uses geographic forwarding to find the paths. In addition, SPEED strive to ensure a certain speed for each packet in the network so that each application can estimate the end-to-end delay for the packets by dividing the distance to the sink by the speed of the packet before making the admission decision. Moreover, SPEED can provide congestion avoidance when the network is congested. The routing module in SPEED is called Stateless Geographic Non-Deterministic forwarding (SNFG) and works with four other modules at the network layer, as shown in Figure 9. The beacon exchange mechanism collects information about the nodes and their location. Delay estimation at each node is basically made by calculating the elapsed time when an ACK is received from a neighbor as a response to a transmitted data packet. By looking at the delay values, SNGF selects the node, which meets the speed requirement. If such a node cannot be found, the relay ratio of the node is checked. The Neighborhood Feedback Loop module is responsible for providing the relay ratio which is calculated by looking at the miss ratios of the neighbors of a node (the nodes which could not provide the desired speed) and is fed to the SNGF module. If the relay ratio is less than a randomly generated number between 0 and 1, the packet is dropped. And finally, the backpressure-rerouting module is used to prevent voids, when a node fails to find a next hop node, and to eliminate congestion by sending messages back to the source nodes so that they will pursue new routes.

Figure 9: Routing components of SPEED

 Pros: - SPEED performs better in terms of end- to-end delay and miss ratio. - The total transmission energy is less due to the simplicity of the routing algorithm (i.e. control packet overhead is less).  Cons: - SPEED does not consider any further energy metric in its routing protocol. Therefore, for more realistic understanding of SPEED’s energy consumption, there is a need for comparing it to a routing protocol, which is energy-aware. - Expensive computation. b. Existing target tracking techniques: (http://mnet.cs.nthu.edu.tw/paper/934355tbl/050526-- Survey%20on%20Target%20Tracking%20in%20wireless %20sensor%20newworks.pdf)

i. Tree-based: o Scalable Tracking Using Networked Sensors (STUN): (H. T. Kung and D. Vlah. “Efficient Location Tracking Using Sensor Networks.” WCNC, March 2003 & Chih-Yu Lin and Yu-Chee Tseng “Structures for In-Network Moving Object Tracking in Wireless Sensor Networks” BROADNETS’04 -- -- Manish Kochhal, Loren Schwiebert, and Sandeep Gupta (2004). "Integrating Sensing Perspectives for Better Self Organization of Ad Hoc Wireless Sensor Networks". JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 20, 449-475 (2004))

 How: Use Voronoi diagram to build a hierarchy tree as Figure 10 and 11. The leaves are sensors, the querying point as the root, and the other nodes are communication nodes. The routing is based on the tree. The communication nodes can aggregate data to reduce redundancy.

Figure 10: Using Voronoi diagram to build graph

Figure 11: Hierarchy tree T from the graph

 Pros: - Data could be aggregated. - Routing follows the tree.  Cons: - All nodes are ON  waste energy and noise sensitive. - Nodes are supposed direct communicate to the base (root)  Can not apply for large area. - Data can be aggregated only when there is one intruder. - Build the tree is so expensive when number of nodes is large. o Dynamic Convoy Tree-Based Collaboration (DCTC): (Wensheng Zhang and Guohong Cao, “DCTC: Dynamic Convoy Tree-Based Collaboration for Target Tracking in Sensor Networks” IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 3, NO. 5, SEPTEMBER 2004 & Wensheng Zhang and Guohong Cao, “Optimizing Tree Reconfiguration for Mobile Target Tracking in Sensor Networks” Infocom 2004)

 How: DCTC relies on a tree structure called “convoy tree”. The tree is dynamically configured to add some nodes and prune some nodes as the target moves. When a target shows up for the first time, an initial convoy tree is constructed (Figure 12). The root collects data from nodes surrounding the target, process the data. When the target moves, the membership of the tree is changed. The structure of the tree is reconfigured if necessary.

Figure 12: Using convoy tree to track the target

 Pros: - Save energy  Cons: - DCTC did not mention how to detect the target at first time. If all nodes are turn ON  Inefficient and noise sensitive. - Nodes in DCTC can create, maintain or reconfigure the tree  demand high performance for the nodes and energy consumption. - At each target moving, the tree and the routes are re-calculated  Inefficient. - Calculation booms when there appear many targets. ii. Cluster-based: o Dynamic Clustering for Acoustic Target Tracking: (Wei- Peng Chen, Jennifer C. Hou, and Lui Sha, Fellow, IEEE “Dynamic Clustering for Acoustic Target Tracking in Wireless Sensor Networks” IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 3, NO. 3, JULY SEPTEMBER 2004 -- Xiang Ji, Hongyuan Zha, John J. Metzner, and George Kesidis , “Dynamic cluster structure for object detection and tracking in wireless ad-hoc sensor networks” ICC 2004)

 How: Based on LEACH algorithm. However, it aims to select the cluster head (CH) that closest to the target (ideally). To do that, it uses the strength of a received acoustic signal to estimate the distance from a node to target. A CH volunteers to become active when the strength exceeds a predetermined threshold. If there are more than one CH volunteers, a random delay- based broadcast mechanism is used to select one. When a CH is active, it will broadcast a packet that contains the energy and the extracted signature of the detected signal to sensors, receive replies from sensors, construct two Voronoi diagram (one for set of neighbor sensors and one for set of neighbor CHs) to estimate the location of the target based on replies, and send the result to subscriber(s).

 Pros: - More energy efficiency than original LEACH. - Estimate target position more accuracy.  Cons: - Still require all node turn ON. - Noise sensitive because it estimates distance by strength of acoustic signal. - High performance required for nodes, especially, when number of nodes is huge. - Still have LEACH weaknesses. iii. Prediction-based: o Prediction-based: (Yingqi Xu Winter, J. Wang-Chien Lee “Prediction-based strategies for energy saving in object tracking sensor networks” Mobile Data Management, 2004. Proceedings. 2004 IEEE International Conference -- Xu, Y., Winter, J., Lee, W.-C. “Dual predictionbased reporting for object tracking sensor networks” MOBIQUITOUS 2004)

 How: Cluster-based with prediction models:

- Heuristics INSTANT: Current node assumes that moving objects will stay in the current speed and direction for the next (T-X) seconds. - Heuristics AVERAGE: By recording some history, the current node derives the object’s speed and direction for the next (T-X) seconds from the average of the object movement history. - Heuristics EXP_AVG: Assigns different weights to the different stages of history. -  Pros: - Could save energy.  Cons: - Have the weaknesses of the based protocol. - Expensive computation for node. - Inaccuracy and unstable results. - Did not mention how to detect intruders at the first time.

3. Detailed description of OCO a. Assumptions:

i. The base station: The base station is likely a node with unlimited energy, high performance computer, and can one way direct communication to any node.

ii. Sensor node ID and location: Nodes are assumed having a unique ID and knowing their location (by attached GPS or etc).

iii. Data package size: 2000-bit per data packet, 64-bit per signal (advertising or neighbor wake-up) packet.

iv. Data rate: It is assumed 2Kbits/s.

v. Energy consumption formula:

o Energy consumption for module: Based on the MICA2DOT (MPR500) of UCLA as in Figure 13 (re-drawn from http://www.xbow.com/Support/Support_pdf_files/MPR- MIB_Series_Users_Manual.pdf page 21): Figure 13: MOTE’s specification at 38.4Kbits/s and Vcc=3v (page 4)

From the Figure 13, we can determine that current consumption of sensor board (Esensor) is equal to 2/3 of the current of the radio in receive mode.

o Energy consumption formula for transmitter:

To transmit k bits to a distance d, we need:

Etx= Eelec* k(bits) + Eamp*k(bits)*d^2

Eelc: Energy is used for transmit electronic circuit. Eamp: Energy is used for transmit amplifier.

Figure 13 does not tell clearly about current consumption of electronic board and the amplifier. We get the value from LEACH simulation ( http://academic.csuohio.edu/yuc/mobile03/0403- heinzelman.pdf):

Eelc = 50nJ/bit Eamp = 100pJ/bit/m2 = 0.1nJ/bit/m2 Esensor = Epro = Eelc = 50nJ/bit

(Epro = Energy consumption for processor module. And, we assume that the modules take zero energy in sleep mode). vi. Sensor node modes: There are 3 modes:

o ACTIVE: All module of the node is turned on (Processor = ON, Radio = ON, Sensor Board =ON/full operation). In practical, the processor is in sleep mode, it becomes active when having an eternal interrupt. So, in the simulation, we assume that in ACTIVE mode, the processor is in sleep mode. It just switches to full operation when having an event. It means when creating a new message, an energy as follow is consumed:

Epro_data = 2000(bit)*50(nJ) = 100 µJ/message Epro_signal = 64(bit)*50(nJ) = 3.2 µJ/message ~ 3 µJ/message

The radio board in this mode is assumed in idle mode (~ receive mode). Hence, energv is supposed to take: 2*10^3(bit)*50(nJ) = 100 µJ/s

100 µJ * 2/3 (assumed that current of sensor board is 2/3 of radio board) = 66 µJ is also the energy for sensor board (Esensor) at each second. The transiver board is likely processor board assumption. It just takes energy for each sent message:

Erecv_data = 2000(bit)*50(nJ) = 100 µJ Erecv_signal = 64(bit)*50(nJ) = 3.2 µJ ~ 3 µJ

Etrans_data = 2000(bit)*(50(nJ) + 0.1(nJ)*d^2) Etrans_signal = 64(bit)*(50(nJ) + 0.1(nJ)*d^2)

We assume that, the optimized distance is 60m. So:

Etrans_data = 2000(bit)*(50(nJ) + 0.1(nJ)*60*60) = 820 µJ/message

Etrans_signal = 64(bit)*(50(nJ) + 0.1(nJ)*60*60) = 26.2 µJ/message ~ 26 µJ/message

It also means that the transmitter consumes 820 µJ/message for all distances d <= 60m.

Create/Receive a data message 100 µJ Create/Receive a signal message 3 µJ Send a data message (d<= 60m) 820 µJ Send a signal message (d<=60m) 26 µJ Send a message (d > 60m) 100 µJ + 0.1*d^2 Sensor board (full operation) 66 µJ/s Radio board (idle mode) 100 µJ/s Summary table

o FORWARD: In this mode, the sensor module of the node is in sleep mode. The node only forwards all messages received from its neighbors.

o SLEEP: In this mode, the node is totally turned off and almost consumes zero energy. It just turns on each long period to processes commands from the base. b. Collecting positions phase: (Happen at nodes)

i. Purpose: The base collects all reachable nodes in the network.

ii. How: When millions of sensor nodes are thrown randomly, they are all in FORWARD mode. The base station assigned its ID as a father to all nodes in its coverage (neighbor nodes) and asks them about their positions. At the neighbors, after sending their IDs and positions to their father (the base), they mark itself as recognized and do as the base does with their neighbors and so on. Note that, when a node gets the information (position and ID) from its neighbor, it just forwards the message to its father and by this way the message will reach the base. So, after this step, the base got all information about reachable nodes (ID and positions) in the network. c. Processing phase: (Happen at the base ~ high performance computer and unlimited energy)

i. Purpose: Clean up the redundant nodes, assign mission for nodes, and route. ii. How: o Clean up the redundant nodes:

 Define: A redundant node is a node that its sensing coverage zone is occupied by one or more other nodes

 Algorithm:

- Initialize a list of node that is supposed to cover all network area (a rectangular of xmin, ymin, xmax, ymax), called Area_List.

- Assign Area_List = null.

- Add the base node to the Area_List.

- For each point in the network area. If the point isn’t covered by the node in the Area_List -> Add the node that contains the point to the Area_List.

- Nodes aren’t in the Area_list called redundant nodes.

o Assign missions for nodes:

 Define: Classify redundant nodes (SLEEP mode), border nodes (ACTIVE mode), and forwarding nodes (FORWARD mode). The base assigns tasks for nodes by broadcasting with node ID.

 Algorithm: - Redundant nodes are classified in part 1.

- Border nodes: Nodes that stay at the border of the network area. To find out these nodes, firstly, we build a geographical image about the coverage zone of the network. Then, we apply border detection for the image to find out a list of points that stay at the border of the image, called border points. Finally, find all nodes in the Area_List that contain at least one border point. These nodes are called border node. See Figure 14, 15, 16 for more detail

Figure 14: Nodes in the network area

Figure 15: Coverage zone and border.

Figure 16: Border nodes

- Forwarding node: Nodes are in the Area_List but not a border node. o Route:

 Purpose: Find the shortest path (the least hops) from every node in the Area_List to the base.  Algorithm: - Work only with nodes in the Area_List.

- Assign father_ID = 0 for all nodes.

- Initialize 2 processing list, called Process_List[2]. A boolean variable, called active, is used to recognize which processing list is in active.

- Assign active = 0. Add the base to Process_List[0]. Assign Process_List[1] = null.

- While Process_List[active] != null { Foreach node pn in Process_List[active] { Neighbor_list = all neighbors of pn Foreach node nn in Neighbor_list If(nn.father_ID == 0) { nn.father_ID = pn.ID add nn to Process_List[1- active] } } Process_List[active].Clear active = 1- active }

- After the while loop, each node in the Area_list has a father_ID. When a node want to send a message to the base, it just delivery the message to its father_ID and so on. The algorithm ensures that by this way, all the messages will reach the base with minimum hops. Figure 17 shows the routing paths from nodes to the base by following father_ID. Figure 17: Routes follow father_ID.

d. Tracking phase: Objects are assumed coming from outside. Normally, only the border nodes are ACTIVE. When a border node detects an object, it periodically sends its position information to the base by using father_ID (as said above). When it lost the object, it will turn all its neighbors (forwarding nodes) to ACTIVE (assumed that the delay time is smaller than sensing radius / object speed). If a neighbor detects the object, it will send its position to the base and right after it lost the object, it turns all its neighbors to ACTIVE and so on. If activated neighbors detect nothing, they automatically switch to the previous state (FORWARD) after a short interval. By this way, the objects are tracked during the time in the network area.

e. Maintenance phase: (Happen at nodes)

i. Purpose: Reconfigure the network when events happen.

ii. Define events and estimate: o Nodes in Area_List run out energy  frequently. o Nodes are broken suddenly  rarely. o Nodes change position by physical events, such as earthquake, explosives, or etc  rarely.

iii. Algorithm: o Event No. 1: When energy level of a node is below a threshold, it turns all its sons to SLEEP and sends a report to the base. When the base gets the report, it performs the processing phase with dead node are deleted and re-assign tasks for nodes. o Event No. 2: After an interval (long period: hours or days), nodes require their sons to send their ID (small size message) to them  They can detect dead node IDs. o Event No. 3: When a node changes position, it automatically turns to SLEEP mode  Become event No. 2.

4. Design experiments to compare OCO to LEACH & direct communication a. Scenario: 200  1000 sensor nodes are thrown randomly in area of 640m x 540m. Each node has 2J (2*10^6 µJ) of energy with sensing radius = 30m and communication radius = 60m. Intruder objects are supposed moving specific paths. No data aggregation is allowed.

b. Utilized tools and module descriptions

i. Tools: OMNET++, C#, Matlab.

ii. Module description under OMNET++: Sensor node module: (Figure 18)

Application

Sensor Coordinator Module MAC Radio

Energy

Layer 0

Figure 18: A node structure

o Layer 0 module: Represented for physical layer. It consists of gates (in/out) and be responsible for making connection between the node and its neighbors. Its behaviors include forward messages from higher layer to its neighbors and vice versa.

o MAC module: Represented for pre-processing packet layers. It consists of gates (in/out) and queues (incoming queue and outgoing queue). When the queue is full, it deletes some latest messages to make sure that there is enough room in the queue for new messages. It helps to evaluate performance of the node. (Note: In current simulation, this module is temporary eliminated to speech up the performance)

o Application module: Represented for application layer consisting of gates (in/out). Note that, at anytime, after sending a message, the module automatically sends a DECREASE_ENERGY message to energy module (through the coordinator) to let the module decrease the energy by one unit.

o Coordinator module: An interface to connect all modules together. It categories incoming messages to delivery them to the right module. For example, when receiving a DECREASE_ENERGY message, it will forward the message to energy module.

o Sensor module: Represented for sensor board in a sensor node. If SENSOR_SWITCH parameter is ON (=1), the module consumes energy, so, after an interval (timer), the module send DECREASE_ENERGY message to the energy module (through the coordinator). When the timer ticks, the waiting timer decreases. The waiting timer is set by SENSOR_REFRESH messages from application module. If the waiting timer is zero, the module will turn off (SENSOR_SWITCH parameter is set to 0). o Radio module: Represented for the radio board in a sensor node. If RADIO_SWITCH parameter is ON (=1), the module consumes energy, so, after an interval (timer), the module send DECREASE_ENERGY message to the energy module (through the coordinator). o Energy module: Represented for battery in a sensor node. At the beginning, each sensor node is set to a specific energy level (ENERGY parameter). If the module receives a DECREASE_ENERGY message, it decreases the energy level by one. o Parameters:  CNNCTVTY: Maximum connections a node has.  OCCUPATION: Task of the node  PX: Position by X.  PY: Position by Y.  ID: ID of node.  FATHER: ID of node for forwarding message.  SENSING_RADIUS: Radius of zone that node can sensing.  COMMUNICATION_RADIUS: Radius of zone that node can communication to.  ENERGY: Energy level.  SENSOR_SWITCH: Turn ON/OFF the sensor module.  POWER_SWITCH: Turn ON/OFF the node.

Object module: (figure 19)

objectApplication

Layer 0

Figure 19: An object structure o Layer 0 module: Similar to layer 0 of the sensor node. However, the connections are re-created after each moving (the manager module handles this task). o objectApplication module: The object is moved by reading position from a text file. The sensing is simulated by creating connection between the object and the sensor nodes in ACTIVE mode, then, the object sends SENSOR_INFO messages to all the connected nodes after an interval.

Manager module: This module aims to help the simulation. Firstly, it read the network.txt (file stores all information of position, task, routing for all nodes) to place the nodes. Then, it makes connections among nodes (it checks the coverage zone of all nodes to see if any node is in the coverage and make connections between the node and the covered nodes). Secondly, at each time of the object movement, it will consider if any node is in the object zone (nodes can sense the object) to create the connection among them so that the object can send SENSOR_INFO messages to the nodes. Also, it manages the broadcast from the base to all nodes (create connections from the base to all nodes). Finally, it controls the power switch (POWER_SWITCH parameter) for all nodes in the network.

Network module: The simulation network consists of sensor nodes, objects, and a manager module and place in the area of a rectangular (0, 0, xmax, ymax).

c. Apply direct communication for target tracking on the scenario: In this case, all nodes are in ACTIVE mode and send to information about intruders to the base by direct communicating.

d. Apply LEACH-based for target tracking on the scenario: In this case, nodes are ACTIVE all the time and organized by LEACH- based algorithm. The number of cluster head is selected 5% (the optimal cluster head number of LEACH) of the total number of nodes.

e. Apply OCO for target tracking on the scenario: In this case, nodes are organized as the proposed method above.

5. Simulation results

a. Metrics

i. Energy consumption: Count the total energy consumption after the simulation.

ii. Accuracy: The number of detected positions of methods is compared to the number of detected positions in the case when all sensor nodes in the network are ACTIVE.

iii. Cost per detected position: Count the ratio between Energy consumption and number of detected positions.

b. Diagrams

The object path is supposed as Figure 20. The energy consumption of methods is collected in the cases of 200, 400, 600, 800, and 1000 nodes (area: 640x540). The result is as Figure 21, Figure 22, and Figure 23.

Figure 20: The testing object path Figure 21: Energy consumption for each method Figure 22: Accuracy of methods

Figure 23: Cost per detected point of methods c. Explanations and comparisons

The Figure 21 shows that proposed method consumed less energy. When the number node is smaller than 400, the border in this case is longer so he OCO consumes more energy. When the number node increases, the energy consumption of OCO goes toward a threshold. Meanwhile, the energy consumption of the direct communication and LEACH-based method increase forever.

6. Related problems

a. Living time: The border nodes seem to drain energy first. However, we can reassign task for nodes (the network coverage may be shrunk down by the time). In addition, the neighbors of the base are used to forward messages for all network, so, they could run out energy faster than others. To cope with this problem, skeleton routing can be used.

b. Noise sensitive: The proposed method is very noise insensitive because, normally, only the border nodes are active. The noise effects just at the border.

c. Security: The proposed method can support shared key. For other security stuff, we need node-node communication support. The skeleton-based protocol could do that. 7. Summary:

 The proposed method seems to consume less energy than others.  The new method demands less computation on node. The nodes is only keep father ID, son ID (for maintenance phase).  Noise insensitive is also a strong feature of the method.

8. Future work:  Do more testing.  Simulate the skeleton method.  Evaluate network life time for the OCO method and the skeleton method.