## Localization for mobile sensor networks (2004)

### Cached

### Download Links

Citations: | 212 - 0 self |

### BibTeX

@INPROCEEDINGS{Hu04localizationfor,

author = {Lingxuan Hu and David Evans},

title = {Localization for mobile sensor networks},

booktitle = {},

year = {2004},

pages = {45--57}

}

### Years of Citing Articles

### OpenURL

### Abstract

Many sensor network applications require location awareness, but it is often too expensive to include a GPS receiver in a sensor network node. Hence, localization schemes for sensor networks typically use a small number of seed nodes that know their location and protocols whereby other nodes estimate their location from the messages they receive. Several such localization techniques have been proposed, but none of them consider mobile nodes and seeds. Although mobility would appear to make localization more difficult, in this paper we introduce the sequential Monte Carlo Localization method and argue that it can exploit mobility to improve the accuracy and precision of localization. Our approach does not require additional hardware on the nodes and works even when the movement of seeds and nodes is uncontrollable. We analyze the properties of our technique and report experimental results from simulations. Our scheme outperforms the best known static localization schemes under a wide range of conditions.

### Citations

1335 | RADAR: An Inbuilding RF-based User Location and Tracking System
- Bahl, Padmanabhan
- 2000
(Show Context)
Citation Context ...hniques are described in Section 2. They all suffer from one or both of these problems: 1. Dependence on special hardware. Techniques that depend on measuring ranging information from signal strength =-=[2]-=-, time of arrival [43], time difference of arrival [40] or angle of arrival [36] require hardware that is typically not available on sensor nodes. Adding the required hardware increases the cost and s... |

1084 |
A.F.M.: Novel approach to nonlinear/non-Gaussian Bayesian state estimation
- Gordon, Salmond, et al.
- 1993
(Show Context)
Citation Context ...he importance weights will increase [27], re-sampling techniques [39] are used to eliminate trajectories with small normalized importance weights. SMC has been successfully applied in target tracking =-=[17]-=-, robot localization [10] and computer vision [23]. We provide a brief introduction below. A more detailed introduction can be found in [13], and an overview and discussion of SMC’s properties can be ... |

759 | GPS-less Low Cost Outdoor Localization for Very Small Devices
- Bulusu, Heidemann, et al.
- 2000
(Show Context)
Citation Context ... local techniques that rely on a high density of seeds so that every node can hear several seeds, and hop counting techniques that rely on flooding a network. Local Techniques. In the Centroid method =-=[6]-=-, each node estimates its location by calculating the center of the locations of all seeds it hears. If seeds are well positioned, location error can be reduced [7], but this is not possible in ad hoc... |

724 | Location systems for Ubiquitous Computing
- Hightower, Borriello
- 2001
(Show Context)
Citation Context ...n simulated experiments and compares our results with other localization techniques. 2. BACKGROUND Extensive research has been done on localization for wireless networks. A general survey is found in =-=[20]-=-. Here, we provide only a brief survey focusing only on localization techniques suitable for ad hoc sensor networks. The approaches taken to achieve localization in sensor networks differ in their ass... |

623 | Dynamic fine-grained localization in Ad-hoc networks of sensors
- Savvides, Han, et al.
(Show Context)
Citation Context ...om one or both of these problems: 1. Dependence on special hardware. Techniques that depend on measuring ranging information from signal strength [2], time of arrival [43], time difference of arrival =-=[40]-=- or angle of arrival [36] require hardware that is typically not available on sensor nodes. Adding the required hardware increases the cost and size of the nodes. 2. Requirement for particular network... |

614 | Robust monte carlo localization for mobile robots
- Thrun, Fox, et al.
- 2001
(Show Context)
Citation Context ...ed environments, but assumes fixed seed locations and requires a learning phase, so is not well suited to mobile sensor network applications. Our work adapts the Monte Carlo localization (MCL) method =-=[10, 42]-=- developed for use in robotics localization for use in mobile sensor network applications. MCL is a particle filter combined with probabilistic models of robot perception and motion. It outperforms ot... |

599 | V.: A key-management scheme for distributed sensor networks
- Eschenauer, Gligor
- 2002
(Show Context)
Citation Context ...f locations) until the next key in the hash chain is released. It would also require loose synchronization among the seeds. Alternative key establishment mechanisms such as random key predistribution =-=[14, 9]-=- can establish keys using symmetric cryptography. These techniques require bidirectional messages, however, and are not suitable for localization techniques where seeds transmit further than nodes. Fu... |

481 | Packet leashes: a defense against wormhole attacks in wireless networks
- Hu, Perrig, et al.
- 2003
(Show Context)
Citation Context ...5, 26, 32] and improve caching behavior for applications where requests may be location dependent [28]. Security can also been enhanced by location awareness (for example, preventing wormhole attacks =-=[22, 24]-=-). However, putting GPS receivers in every node or manually configuring locations is not cost effective for most sensor network applications. Recently some localization techniques have been proposed t... |

370 | A survey on position-based routing in mobile ad hoc networks
- Mauve, Widmer, et al.
(Show Context)
Citation Context ...ng, vehicle tracking and mapping depend on knowing the locations of sensor nodes. In addition, location-based routing protocols can save significant energy by eliminating the need for route discovery =-=[25, 26, 32]-=- and improve caching behavior for applications where requests may be location dependent [28]. Security can also been enhanced by location awareness (for example, preventing wormhole attacks [22, 24]).... |

335 | Ad hoc positioning system (APS) using AOA
- Niculescu, Nath
- 2003
(Show Context)
Citation Context ...roblems: 1. Dependence on special hardware. Techniques that depend on measuring ranging information from signal strength [2], time of arrival [43], time difference of arrival [40] or angle of arrival =-=[36]-=- require hardware that is typically not available on sensor nodes. Adding the required hardware increases the cost and size of the nodes. 2. Requirement for particular network topologies. Most techniq... |

325 |
Bayesian inference in econometric models using Monte Carlo integration
- Geweke
- 1989
(Show Context)
Citation Context ... and p ( o | i ) > 0 } Filtering (3.3) Rfiltered = { i t i t L t = choose (L t ∪ Rfiltered, N) a set of m weighted samples, and to update them recursively in time using the importance sampling method =-=[16]-=-. Since the unconditional variance of the importance weights will increase [27], re-sampling techniques [39] are used to eliminate trajectories with small normalized importance weights. SMC has been s... |

296 |
Global Positioning System, Theory and Practice
- Hofmann-Wellenhof, Lichtenegger, et al.
- 1994
(Show Context)
Citation Context ... in Section 2. They all suffer from one or both of these problems: 1. Dependence on special hardware. Techniques that depend on measuring ranging information from signal strength [2], time of arrival =-=[43]-=-, time difference of arrival [40] or angle of arrival [36] require hardware that is typically not available on sensor nodes. Adding the required hardware increases the cost and size of the nodes. 2. R... |

266 | Organizing a global coordinate system from local information on an ad hoc sensor network
- NAGPAL, SHROBE, et al.
(Show Context)
Citation Context ...he whole network. But prior deployment of seeds is not possible in many sensor network applications (for example, sensor nodes dropped from plane over a hostile territory). Hop count based techniques =-=[35, 34]-=- avoid the need for a large number of seeds, but instead require dense and uniform node distribution. We are interested in performing localization in a more general network environment where no specia... |

236 |
Stochastic Models, Estimation, and Control Volume 1
- Maybeck
- 1994
(Show Context)
Citation Context ...n based on its motion and sensor data. If both the motion and the measurement model can be described using a Gaussian density, and the initial state is also a Gaussian, then a classical Kalman filter =-=[33]-=- can be used. Grid based Markov localization [4, 5] has been proposed to deal with multimodal and non-Gaussian densities. However, the grid representation can impose a significant memory and computati... |

187 |
Sequential Imputations and Bayesian Missing Data Problems
- Kong, Liu, et al.
- 1994
(Show Context)
Citation Context ...∪ Rfiltered, N) a set of m weighted samples, and to update them recursively in time using the importance sampling method [16]. Since the unconditional variance of the importance weights will increase =-=[27]-=-, re-sampling techniques [39] are used to eliminate trajectories with small normalized importance weights. SMC has been successfully applied in target tracking [17], robot localization [10] and comput... |

185 | Robotics-based location sensing using wireless ethernet
- Ladd, Bekris, et al.
- 2002
(Show Context)
Citation Context ...tion. Ladd et al. used a robotics localization approach to achieve accurate localization for wireless networks by using learned variation in RF signal strengths received using standard Ethernet cards =-=[29, 30]-=-. Their approach performs well for indoor localization in fixed environments, but assumes fixed seed locations and requires a learning phase, so is not well suited to mobile sensor network application... |

155 |
Using the SIR algorithm to simulate posterior distributions. In Bayesian statistics 3. Edited by
- Rubin
- 1988
(Show Context)
Citation Context ...ighted samples, and to update them recursively in time using the importance sampling method [16]. Since the unconditional variance of the importance weights will increase [27], re-sampling techniques =-=[39]-=- are used to eliminate trajectories with small normalized importance weights. SMC has been successfully applied in target tracking [17], robot localization [10] and computer vision [23]. We provide a ... |

122 | Using Directional Antennas to Prevent Wormhole Attacks
- Hu, Evans
(Show Context)
Citation Context ...rypted. Known defenses for wormhole attacks rely on either nodes already knowing their locations [22], or require specialized hardware such as tightly synchronized clocks [22] or directional antennas =-=[21]-=-. MCL offers several desirable properties for security adaptation. Since it does not depend on seeds with powerful transmitters, bidirectional verification and key establishment is possible. When node... |

63 | Creemers. Integrating global position estimation and position tracking for mobile robots: the dynamic markov localization approach
- Burgard, Derr, et al.
- 1998
(Show Context)
Citation Context ...e motion and the measurement model can be described using a Gaussian density, and the initial state is also a Gaussian, then a classical Kalman filter [33] can be used. Grid based Markov localization =-=[4, 5]-=- has been proposed to deal with multimodal and non-Gaussian densities. However, the grid representation can impose a significant memory and computational burden, especially if one is interested in hig... |

58 |
Secure Routing in Sensor Networks: Attacks and Countermeasures
- Karlof, Wagner
- 2003
(Show Context)
Citation Context ...5, 26, 32] and improve caching behavior for applications where requests may be location dependent [28]. Security can also been enhanced by location awareness (for example, preventing wormhole attacks =-=[22, 24]-=-). However, putting GPS receivers in every node or manually configuring locations is not cost effective for most sensor network applications. Recently some localization techniques have been proposed t... |

46 | Localization in sensor networks with fading and mobility
- Bergamo, Mazzini
(Show Context)
Citation Context ...are not designed with any consideration for how mobility can be exploited to achieve localization. The only work we are aware of that considers localization with mobile nodes is Bergamo and Mazzini’s =-=[3]-=-. They considered mobility of nodes only and assumed a network with two seeds at fixed locations that can transmit across the entire network, and nodes that are able to measure signal strength accurat... |

36 |
Monte Carlo techniques for prediction and filtering of non-linear stochastic processes
- Handschin
- 1970
(Show Context)
Citation Context ...f new landmarks) are incorporated to filter and update data. The process repeats and the robot continually updates its predicted location. Like our work, MCL applies the Sequential Monte Carlo method =-=[18]-=- to achieve localization. However, there are substantial differences between robot localization and node localization for sensor networks. While robot localization locates a robot in a predefined map,... |

36 | On the feasibility of using wireless ethernet for indoor localization
- Ladd, Bekris, et al.
- 2004
(Show Context)
Citation Context ...tion. Ladd et al. used a robotics localization approach to achieve accurate localization for wireless networks by using learned variation in RF signal strengths received using standard Ethernet cards =-=[29, 30]-=-. Their approach performs well for indoor localization in fixed environments, but assumes fixed seed locations and requires a learning phase, so is not well suited to mobile sensor network application... |

32 |
Tarek Abdelzaher. Range-free localization schemes for large scale sensor networks
- He, Huang, et al.
- 2003
(Show Context)
Citation Context ...e surrounding environment changes). Because of the hardware limitations of sensor devices, range-free localization algorithms are a cost effective alternative to more expensive range-based approaches =-=[19]-=-. There are two main types of range-free localization algorithms that have been proposed for sensor networks: local techniques that rely on a high density of seeds so that every node can hear several ... |

31 | Exploiting Location Information for Infostation-Based Hoarding. MobiCom 2001
- Kubach, Rothermel
(Show Context)
Citation Context ...ion-based routing protocols can save significant energy by eliminating the need for route discovery [25, 26, 32] and improve caching behavior for applications where requests may be location dependent =-=[28]-=-. Security can also been enhanced by location awareness (for example, preventing wormhole attacks [22, 24]). However, putting GPS receivers in every node or manually configuring locations is not cost ... |

29 |
Contour Tracking by Stochastic
- Isard, Blake
- 1996
(Show Context)
Citation Context ...ing techniques [39] are used to eliminate trajectories with small normalized importance weights. SMC has been successfully applied in target tracking [17], robot localization [10] and computer vision =-=[23]-=-. We provide a brief introduction below. A more detailed introduction can be found in [13], and an overview and discussion of SMC’s properties can be found in [12]. 3.1 Location Estimation Algorithm T... |

27 |
Boleng and Vanessa Davies, A Survey of Mobility Models for Ad Hoc Network
- Camp, Jeff
- 2002
(Show Context)
Citation Context ... and use a fixed nd = 10 for other experiments. • Seed density (sd), the average number of seeds in one hop transmission range (considered in Section 5.4). We adopt the random waypoint mobility model =-=[8]-=- for both nodes and seeds. It is one of the most commonly used mobility models for mobile ad hoc networks. In the random waypoint model, a node randomly chooses its destination, its speed of movement,... |

20 |
Density Adaptive Algorithms for Beacon Placement
- Bulusu, Heidemann, et al.
- 2001
(Show Context)
Citation Context ... Techniques. In the Centroid method [6], each node estimates its location by calculating the center of the locations of all seeds it hears. If seeds are well positioned, location error can be reduced =-=[7]-=-, but this is not possible in ad hoc deployments. The APIT method [19] isolates the environment into triangular regions between beaconing nodes, and uses a grid algorithm to calculate the maximum area... |

12 | DV Based Positioning in Ad hoc Networks. Kluwer Journal of Telecommunication Systems - Niculescu, Nath - 2003 |

5 |
Fromherz. Localization from Mere Connectivity. MobiHoc
- Shang, Ruml, et al.
- 2003
(Show Context)
Citation Context ...erty, Pister and Ghaoui developed a centralized technique using 46 convex optimization to estimate positions based only on connectivity constraints given some nodes with known positions [11]. MDS-MAP =-=[41]-=- improves on these results by using a multidimensional scaling approach, but still requires centralized computation. Requiring central computation would be infeasible for mobile applications because o... |

4 |
Perrig and Dawn Song. Random Key Predistribution Schemes for Sensor Networks
- Chan, Adrian
- 2003
(Show Context)
Citation Context ...f locations) until the next key in the hash chain is released. It would also require loose synchronization among the seeds. Alternative key establishment mechanisms such as random key predistribution =-=[14, 9]-=- can establish keys using symmetric cryptography. These techniques require bidirectional messages, however, and are not suitable for localization techniques where seeds transmit further than nodes. Fu... |

4 |
Guangyu Pei and ChingChuan Chiang. A Group Mobility Model for Ad Hoc Wireless Networks
- Hong, Gerla
- 1999
(Show Context)
Citation Context ...ions, the motion of nodes and seeds may be correlated and demonstrate some group behavior, and this may affect the performance of our algorithm. We use the Reference Point Group Mobility model (RPGM) =-=[15]-=- to investigate the effect of group behavior on our algorithm. In RPGM, the motion of a node is the combination of a group motion vector and a random motion vector. The random motion is based on a ref... |

4 |
Using Proximity and Quantized RSS for Sensor Localization
- Patwari, Hero
- 2003
(Show Context)
Citation Context ...estimates in location calculations, while a range-free solution depends only on the contents of received messages. Range-based approaches have exploited time of arrival [43], received signal strength =-=[2, 37]-=- time difference of arrival of two different signals (TDOA) [40], and angle of arrival (AOA) [36]. Though they can reach fine resolution, either the required hardware is expensive (ultrasound device f... |

2 |
Pister and Laurent El Ghaoui. Convex Position Estimation in Wireless Sensor Networks
- Doherty, Kristofer
- 2001
(Show Context)
Citation Context ...each node. Doherty, Pister and Ghaoui developed a centralized technique using 46 convex optimization to estimate positions based only on connectivity constraints given some nodes with known positions =-=[11]-=-. MDS-MAP [41] improves on these results by using a multidimensional scaling approach, but still requires centralized computation. Requiring central computation would be infeasible for mobile applicat... |

2 |
Godsill and Christophe Andrieu. On Sequential Monte Carlo Sampling Methods for Bayesian Filtering
- Doucet, Simon
- 2000
(Show Context)
Citation Context ... and observation is not easy to determine. Except for a few special cases including linear Gaussian state space models (Kalman filter [33]), it is impossible to evaluate the distribution analytically =-=[12]-=-. The Sequential Monte Carlo (SMC) method [18] provides simulation-based solutions to estimate the posterior distribution of nonlinear discrete time dynamic models. The key idea of SMC is to represent... |

2 |
Sound Mobility Models. MobiCom 2003
- Yoon, Liu, et al.
(Show Context)
Citation Context ...on, its speed of movement, and its pause time after arriving at the destination. We assume nodes are unaware of their velocity and direction, but have a known maximum velocity vmax. As pointed out in =-=[44]-=-, the random waypoint model suffers from the decay of average speed, and this will provide an unsound basis for simulation. We used a modified random waypoint model to maintain the average speed. Inst... |

1 |
Hennig and Timo Schmidt. Estimating the Absolute Position of a Mobile Robot Using Position Probability Grids
- Burgard, Fox, et al.
- 1996
(Show Context)
Citation Context ...e motion and the measurement model can be described using a Gaussian density, and the initial state is also a Gaussian, then a classical Kalman filter [33] can be used. Grid based Markov localization =-=[4, 5]-=- has been proposed to deal with multimodal and non-Gaussian densities. However, the grid representation can impose a significant memory and computational burden, especially if one is interested in hig... |

1 |
Nando de Freitas and Neil Gordon. An Introduction to Sequential Monte Carlo Methods
- Doucet
- 2001
(Show Context)
Citation Context ...ights. SMC has been successfully applied in target tracking [17], robot localization [10] and computer vision [23]. We provide a brief introduction below. A more detailed introduction can be found in =-=[13]-=-, and an overview and discussion of SMC’s properties can be found in [12]. 3.1 Location Estimation Algorithm The mobile localization problem can be stated in a state space form as follows. Let t be th... |