K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....such as localization, path planning, collision avoidance, and people finding. The basic occupancy grid map paradigm has been applied successfully in many different ways. For example, some systems use maps locally, to plan collision free paths or to identify environment features for localization [1, 10, 19, 20]. Others, such as many of the systems described in [11, 23] rely on global occupancy grid maps for global path planning and navigation. f Figure 1: A set of noise free sonar measurements that a robot may receive while passing an open door. While the measurements are perfectly consistent, ....

K. Konolige and K. Chou. Markov localization using correlation. In Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence (IJCAI), Stockholm, Sweden, 1999. IJCAI.

....environment maps overlap. Identifying the environment could then be solved by a global localization approach such as Markov Localization [1] where from the estimated pose the corresponding environment could be determined. Taking this point of view, map correlation as presented by Konolige and Chou [7] could be qualified as a related approach. In their work, the robot builds a patch (a small map) consisting of the last few frames of a laser range finder and correlates this patch with an occupancy grid like map in order to estimate the pose of the robot. A similar approach is the ....

K. Konolige and K. Chou. Markov localization using correlation. In Proc. International Joint Conference on Artificial Intelligence (IJCAI'99), Stockholm, 1999.

....a state of the art localizer in dynamic environments. 1 Introduction In the past decade, mobile robot localization and mapping has received substantial interest in AI and robotics [1, 12] The localization problem concerns itself with estimating the pose of a robot relative to a fixed map [9, 11], whereas the mapping addresses the problem of learning a map from sensor data [4, 13, 16] A striking characteristic of the rich literature on this topic is that virtually all published work assumes that the environment is static. This is in contrast to most robotic environments, which usually ....

K. Konolige and K. Chou. Markov localization using correlation. In Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence (IJCAI), Stockholm, Sweden, 1999. IJCAI.

....robot environments. They represent an environment by a fine grained, metric grid of variables that reflect the occupancy of the environment. Once acquired, they facilitate various key aspects of mobile robot navigation, such as localization, path planning, collision avoidance, and people finding [1, 5, 9]. Existing occupancy grid mapping algorithms suffer a key problem. They often generate maps that are inconsistent with the data, particularly in cluttered environments and when learned from sonar data. This problem is due to the fact that existing algorithms decompose the highdimensional mapping ....

K. Konolige and K. Chou. Markov localization using correlation. Proceedings of IJCAI, 1999

....of Registering Strength Signal Strength Figure 1: Two examples of signal strength distributions, measured over time at a constant location. the existence of a good linearization. Arguably the most powerful algorithms to date are based on Bayesian inference, in particular Markov models [15, 8] and Monte Carlo localization [7, 23] Various discretization schemes can be employed but often occupancy grids and point sets are employed. If no usable theoretical model is available, conditional probability distributions can be sampled directly. Alternately, the environment can be modeled with ....

....off the ground. We had a fairly precise map of the building that we had processed to mark off free space and obstacles. The pixel resolution was roughly six centimeters in this map. Our Model The localizer that we implemented operates in the general framework of Bayesian inference localization [22, 8, 15]. We chose a state space and observation space, estimated the required conditional probability distributions and explicitly integrated. We chose various sets of points in the map for the state space. A point for our experiments was chosen as a tuple############# on the floor of the building our ....

K. Konolige and K. Chou. Markov Localization using Correlation. In Proc. of the Seventeenth International Joint Conference on Artificial Intelligence (IJCAI), pages 1154--1159, Seattle, Washington, August 1999.

....In practice, obtaining linearizations for many sensing systems is difficult and errors can propagate very quickly through the system. Bayesian Approaches. Possibly the most powerful family of global localization algorithms to date is based on Bayesian inference, in particular Markov localization [24, 15] and Monte Carlo localization [13, 37] These are generalizations of the Kalman filter. These algorithms estimate posterior distributions over robot poses which are approximated by piecewise constant functions instead of Gaussians, enabling them to represent highly multi modal distributions. In ....

K. Konolige and K. Chou. Markov localization using correlation. In Proc. of the Seventeenth International Joint Conference on Artificial Intelligence (IJCAI), pages 1154--1159, Seattle, Washington, Aug. 1999.

....In practice, obtaining linearizations for many sensing systems is difficult and errors can propagate very quickly through the system. Bayesian Approaches Possibly the most powerful family of global localization algorithms to date is based in Bayesian inference, in particular Markov localization [23, 14] and Monte Carlo localization [12, 35] These are generalizations of the Kalman filter. These algorithms estimate posterior distributions over robot poses which are approximated by piecewise constant functions instead of Gaussians, enabling them to represent highly multi modal distributions. In ....

K. Konolige and K. Chou. Markov Localization using Correlation. In Proc. of the Seventeenth International Joint Conference on Artificial Intelligence (IJCAI), pages 1154-- 1159, Seattle, Washington, August 1999.

....movement a at position l # . Furthermore, the quantity l) denotes the probability of making observation o given the robot s current location is l. It highly depends on the information the robot possesses about the environment and the sensors used. Different kinds of realizations can be found in [12, 7, 18, 2, 8]. In this paper, p(o l) is computed using the image retrieval system described in Section 3. To represent the belief of the robot about its current position we apply a variant of Markov localization denoted as Monte Carlo localization [3, 5] In Monte Carlo localization, the update of the ....

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

....that the p(l o) sum up to 1 over all l. The term p(o l) denotes the probability of making observation o given that the robot s current location is l. It highly depends on the information the robot possesses about the environment and the sensors used. Different kinds of realizations can be found in [2,12,13,19]. Additionally, Markov localization uses a well known formula coming from the domain of Markov chains to update the belief p(l) whenever the robot performs a movement action a: p(l a) p(l a, l # ) p(l # ) dl # . 7) In this equation the term p(l a, l # ) describes the probability that the ....

K. Konolige, Markov localization using correlation, in: Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI-99), Stockholm, Sweden, 1999.

....always yields an inconsistent map. The characteristics of the Kalman filter approaches is that the motion model has to be known and these approaches also dependent on a good initialization. 2. 2 Consistent Pose Estimation approaches Another approach, the Consistent Pose Estimation method [8, 5], is not based on an incremental estimation procedure but keeps sensory observations over a longer period to avoid inconsistencies. Based on pairs of observations, the relative positions of the robot are derived. From the network of relative positions the absolute positions are derived by ....

K. Konolige and K. Chou. Markov localization using correlation. In Proc. International Joint Conference on Artificial Intelligence, pages 1154--1159. Morgan Kau#mann, 1999.

....data. This paper proposes an efficient probabilistic approach for collaborative multi robot localization. Our approach is based on Markov localization [51, 62, 34, 9] a family of probabilistic approaches that have recently been applied with great practical success to single robot localization [7, 39, 23, 67]. In contrast to previous research, which relied on grid based or coarse grained topological representations of a robot s state space, our approach adopts a sampling based representation [17, 21] which is capable of approximating a wide range of belief functions in real time. To transfer ....

....moments of the density. These approaches are unable to localize robots under global uncertainty a problem which Engelson called the kidnapped robot problem [19] Recently, several researchers proposed Markov localization, which enables robots to localize themselves under global uncertainty [9, 34, 51, 62, 39]. Global approaches have two important advantages over local ones: First, the initial location of the robot does not have to be specified and, second, they provide an additional level of robustness, due to their ability to recover from localization failures. Among the global approaches those using ....

[Article contains additional citation context not shown here]

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

.... observation model which gives the probability of the sensor measurement given the location of the robot and the parameterization of the environment (the map) Sometimes this parameter vector describes explicit properties of the environment (such as positions of landmarks [20] or occupancy values [8]) or describes an implicit relation between sensor pattern and location (such as neural networks [13] radial basis functions [21] or look up tables [2] In our laboratory we use a vision sensor for robot localization. One possible approach for modeling the environment is to make a full 3D CAD ....

....model p(x t ju; x t 1 ) repeatedly. If both models are known we can combine them and decrease the motion uncertainty by having a new observation. In this paper we will focus on the observation model. Often the parameter vector is an explicit map of the environment, such as an occupancy grid [20,8] or the positions of landmarks [19] Such a map can be provided from prior knowledge, but here the key issue is to learn this model from data. The approach that we will adopt is that we are not going to estimate the parameters of some sort of CAD model, but our map will consist of an implicit ....

K. Konolige and K. Chou. Markov localization using correlation. In Proc. International Joint Conference on Articial Intelligence, pages 1154-1159. Morgan Kaumann, 1999.

....The perceptual model, denoted P (ojm; models the likelihood of observing o in situations where both the world m and the robot s pose are known. For low dimensional sensors such as proximity sensors, perceptual models can readily be found in the literature [Burgard et al. 1996; Moravec, 1988; Konolige, 1999] . Figure 2 illustrates a perceptual model for a robot that can detect landmarks and that can measure, with some uncertainty, their relative orientations and distances. Figure 2 (a) shows an example map m, in which the dark spots indicate the locations of two indistinguishable landmarks. Figure 2 ....

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

....of being at location l after incorporating o (t) n is obtained by multiplying the observation likelihood P (o (t) n j L (t) n = l) with the prior belief. This likelihood is also called the environment perception model of robot n. Typical models for different types of sensors are described in [11, 9, 18]. 2. Odometry: Now suppose the last item in d (t) n is an odometry measurement, denoted a (t) n . Using the Theorem of Total Probability and exploiting the Markov property, we obtain the following incremental update scheme: Bel (t) n (L = l) Gamma Z P (L (t) n = l j a (t Gamma1) n ....

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

....2. Sensor Models: In addition to the di erent representations of the state space various perception models have been developed for di erenttypes of sensors (see for example Moravec, 1988; Kortenkamp Weymouth, 1994; Simmons Koenig, 1995; Burgard et al. 1996; Dellaert et al. 1999; and Konolige, 1999). These sensor models di er in the way how they compute the probabilityofthe current measurement. Whereas topological approaches suchas(Kortenkamp Weymouth, 1994; Simmons Koenig, 1995; Kaelbling et al. 1996) rst extract landmark information out of a sensor scan, the approaches in (Moravec, ....

....the probabilityofthe current measurement. Whereas topological approaches suchas(Kortenkamp Weymouth, 1994; Simmons Koenig, 1995; Kaelbling et al. 1996) rst extract landmark information out of a sensor scan, the approaches in (Moravec, 1988; Burgard et al. 1996; Dellaert et al. 1999; Konolige, 1999) operate on the raw sensor measurements. The techniques for proximity sensors described in (Moravec, 1988; Burgard et al. 1996; Konolige, 1999) mainly di er in their eciency and how they model the characteristics of the sensors and the map of the environment. In order to combine the strengths of ....

[Article contains additional citation context not shown here]

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Articial Intelligence (IJCAI), 1999.

....data. This paper proposes an efficient probabilistic approach for collaborative multi robot localization. Our approach is based on Markov localization [51, 62, 34, 9] a family of probabilistic approaches that have recently been applied with great practical success to single robot localization [7, 39, 23, 67]. In contrast to previous research, which relied on grid based or coarse grained topological representations of a robot s state space, our approach adopts a sampling based representation [17, 21] which is capable of approximating a wide range of belief functions in real time. To transfer ....

....moments of the density. These approaches are unable to localize robots under global uncertainty a problem which Engelson called the kidnapped robot problem [19] Recently, several researchers proposed Markov localization, which enables robots to localize themselves under global uncertainty [9, 34, 51, 62, 39]. Global approaches have two important advantages over local ones: First, the initial location of the robot does not have to be specified and, second, they provide an additional level of robustness, due to their ability to recover from localization failures. Among the global approaches those using ....

[Article contains additional citation context not shown here]

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

....2. Sensor Models: In addition to the di erent representations of the state space various perception models have been developed for di erent types of sensors (see for example Moravec, 1988; Kortenkamp Weymouth, 1994; Simmons Koenig, 1995; Burgard et al. 1996; Dellaert et al. 1999; and Konolige, 1999). These sensor models di er in the way how they compute the probability of the current measurement. Whereas topological approaches such as (Kortenkamp Weymouth, 1994; Simmons Koenig, 1995; Kaelbling et al. 1996) rst extract landmark information out of a sensor scan, the approaches in ....

....the probability of the current measurement. Whereas topological approaches such as (Kortenkamp Weymouth, 1994; Simmons Koenig, 1995; Kaelbling et al. 1996) rst extract landmark information out of a sensor scan, the approaches in (Moravec, 1988; Burgard et al. 1996; Dellaert et al. 1999; Konolige, 1999) operate on the raw sensor measurements. The techniques for proximity sensors described in (Moravec, 1988; Burgard et al. 1996; Konolige, 1999) mainly di er in their eciency and how they model the characteristics of the sensors and the map of the environment. In order to combine the strengths of ....

[Article contains additional citation context not shown here]

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Articial Intelligence (IJCAI), 1999.

....are subject to inaccuracies as a result of slip. Therefore, external sensors are needed to update the position estimate. Because sensor signals are noisy, a wide variety of probabilistic methods has been developed to obtain a robust estimate of the location of the robot given its sensory inputs [15, 6, 11, 2, 4]. In our laboratory we use an omni directional vision sensor for robot localization. Modeling the relation between robot location and visual observation is not a trivial task, particularly since the images are high dimensional data vectors. For an accurate modeling of these data we should need an ....

K. Konolige and K. Chou. Markov localization using correlation. In Proc. International Joint Conference on Articial Intelligence, pages 1154{ 1159. Morgan Kaumann, 1999.

....CAD map of the environment. He assigns scan points to line segments based on closest neighborhood and then searches for a translation and rotation that minimizes the total squared distance between scan points and their target lines. Scans can also be matched by correlating occupancy grids [21, 22, 15]. However accuracy and run time performance depend heavily on the grid resolution. Weiss et al. 27] use histograms for matching a pair of scans. They first compute a so called angle histogram for determining the rotation of the two scans and then use x and y histograms for computing the ....

K. Konolige. Markov localization using correlation. In Proc. International Joint Conference on Artificial Intelligence (IJCAI '99), Stockholm, 1999.

....single scans in rejecting false positives, but it leaves open the question of how to efficiently perform matching, since single scan techniques are no longer applicable. Fortunately, one author has recently investigated correlational concepts for matching map patches in the context of localization [12]. The resultant techniques have been shown to be both efficient and reliable, and we make use of them here to determine topological correctness in map building with cycles. Correlation operates in a background mode, checking for matches against the old map whenever the robot moves to an ....

....constraint on map matching is that it must be efficient, since we intend to run it constantly in the background as the robot starts cycling back to places previously visited. Recent investigations by one of the authors has provided a fast and accurate matching technique based on correlation [12]. The justification for this technique lies in a Bayesian analysis of the match probability. For any given new map patch r and old map m, we seek the posterior probability p#ljr;m# that the robot is at pose l.Using Bayes rule, we have: p#ljr;m#=k # p#rjl; m#p#l; m#: The sensor response ....

[Article contains additional citation context not shown here]

K. Konolige. Markov localization using correlation. In Proceedings International Joint Conference on Artificial Intelligence (IJCAI'99), Stockholm, 1999.

No context found.

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

No context found.

K. Konolige. Markov localization using correlation. n Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999. 113

No context found.

K. Konolige and K. Chou, "Markov localization using correlation," in Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI'99), 1999.

No context found.

K. Konolige and K. Chou, "Markov localization using correlation," in Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI'99), 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence (IJCAI), Stockholm, Sweden, 1999. IJCAI.

No context found.

K. Konolige and K. Chou, "Markov localization using correlation," in Proc 17th Int. JoYY CoY . Artificial Intelligence, Seattle, WA, Aug. 1999, pp. 1154--1159.

No context found.

K. Konolige and K. Chou, "Markov localization using correlation," in Proceedings of the International Joint Conference on Artificial Intelligence, 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proceedings of the International Joint Conference on Artificial Intelligence, 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proceedings of the International Joint Conference on Artificial Intelligence, 1999.

No context found.

K. Konolige. Markov localization using correlation. n Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999. 113

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proc. Int. Joint Conf. on Artificial Intelligence (IJCAI), pages 1154--1159, 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In International Joint Conference on Articial Intelligence, pages 1154--1159, 1999.

No context found.

K. Konolige. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

No context found.

K. Konolige and K. Chou. Markov localization using correlation. In Proc. of the International Joint Conference on Artificial Intelligence (IJCAI), 1999.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC