C. Fennema, A. Hanson, E. Riseman, R. J. Beveridge, and R. Kumar. Model-directed mobile robot navigation. IEEE Trans. on Systems, Man and Cybernetics, 20: 1352#1369, 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....lines on the floor [1] or tracking buried wires or infrared beacons. The disadvantage of these approaches is that they require substantial engineering of the environment. Recently many researchers have employed landmarks either artificial [2, 3, 4, 5, 6, 7] or extracted from the environment [8, 9, 10, 11, 12] to guide the motion of a mobile robot in indoor environments. The most commonly used approach with artificial landmarks is heuristic: landmarks are designed and placed so that landmark detection under normal circumstances is straightforward. For example, some systems use simple patterns ....

C. Fennema, A. Hanson, E. Riseman, J. R. Beveride, and R. Kumar, "Model-directed mobile robot navigation", IEEE Transactions on Systems, Man, and Cybernetics, vol. 20, no. 6, pp. 1352--1369, 1990.

....on the floor [11] or tracking buried wires or infrared beacons. The disadvantage of these approaches is that they require substantial engineering of the environment. Recently many researchers have employed landmarks either artificial [1, 7, 10, 15, 17, 18] or extracted from the environment [4, 6, 13, 14, 16] to guide the motion of a mobile robot in indoor environments. The most commonly used approach with artificial landmarks is heuristic: landmarks are designed and Support for this work was provided in part by the National Science Foundation under POWRE grant EIA 9806108, VTEPSCoR grant ....

C. Fennema, A. Hanson, E. Riseman, J. R. Beveride, and R. Kumar. Model-directed mobile robot navigation. IEEE Transactions on Systems, Man, and Cybernetics, 20(6):1352--1369, 1990.

.... robot to follow (see, e.g. 15] Second, since unavoidable odometer errors, e.g. due to wheel slippage, render it impossible for any mobile robot to precisely follow a planned trajectory, localization techniques are needed to precisely determine the robot s position and orientation (see, e.g. [4, 5, 7, 6, 10, 11, 18, 21, 22, 24]) In this paper, we describe the design and implementation of a framework for autonomous mobile robot navigation in a known indoor environment that requires only inexpensive range sensors. Our approach: integrates and coordinates the path planning and localization modules with the aid of a ....

....axes correspond to Eigen values. This method is considered to be statistically more accurate than the conventional min max approach [20] which was used in our experiments reported in Section 5. 3. 3 Localization Many di erent localization methods have been proposed in the literature (see, e.g. [4, 5, 6, 7, 10, 11, 18, 21, 22, 24]) We have selected a method we developed that provides fast localization using only range sensor data (i.e. distance measurements) which is based on simple geometric properties of the environment [16] During preprocessing, the workspace is partitioned into sectors using simple visibility ....

C. Fennema, A. Hanson, E. Riseman, Beveridge, and R. Kuman. Model-directed mobile robot navigation. IEEE Trans. Sys., Man, Cybern., 20(6):1352-1369, 1990.

....Medical Institute. 1 or tracking buried wires or infrared beacons. The disadvantage of these approaches is that they require substantial engineering of the environment. Recently many researchers have employed landmarks either artificial [1, 6, 9, 14, 16, 17] or extracted from the environment [3, 5, 12, 13, 15] to guide the motion of a mobile robot in indoor environments. The most commonly used approach with artificial landmarks is heuristic: landmarks are designed and placed so that landmark detection under normal circumstances is straightforward. For example, some systems use simple patterns ....

C. Fennema, A. Hanson, E. Riseman, J. R. Beveride, and R. Kumar. Model-directed mobile robot navigation. IEEE Transactions on Systems, Man, and Cybernetics, 20(6):1352--1369, 1990.

....including those for free flying planar robots, car like robots, and robots with high degrees of freedom. Finally, a Markov model is used by Simmons and Koenig [18] to plan navigation strategies in partially observable environments. Rather than relying on landmarks extracted from the environment [4, 5, 11, 13, 17], the approach taken in this paper and by a number of other research groups [1, 6, 10, 15, 19, 20] is to use artificial landmarks that can be easily and unobtrusively added to the environment. Becker et al. 1] use simple landmarks attached to the ceiling of the environment, and use a recognition ....

C. Fennema, A. Hanson, E. Riseman, J. R. Beveride, and R. Kumar. Model-directed mobile robot navigation. IEEE Transactions on Systems, Man, and Cybernetics, 20(6):1352--1369, 1990.

....was developed in [2] that uses a stereo pair of cameras to determine the correspondence between the observed landmarks and the preloaded map and to estimate the two dimensional location of the sensor from the correspondence. A system for navigation in a partially modeled environment is outlined in [8], and in [3] trinocular stereo and three dimensional line features are used for building, registering, and fusing noisy visual maps. Most of the vision based approaches use landmarks, whose locations are previously defined in world coordinates. Some of the map based localization techniques only ....

C. Fennema, A. Hanson, E. Riseman, Beveridge, and R. Kuman. Model-directed mobile robot navigation. IEEE Trans. Sys., Man, Cybern., 20(6):1352--1369, 1990.

....was developed in [2] that uses a stereo pair of cameras to determine the correspondence between the observed landmarks and the preloaded map and to estimate the two dimensional location of the sensor from the correspondence. A system for navigation in a partially modeled environment is outlined in [9], and in [3] trinocular stereo and three dimensional line features are used for building, registering, and fusing noisy visual maps. Most of the vision based approaches use landmarks, whose locations are previously defined in world coordinates. Some of the map based localization techniques only ....

C. Fennema, A. Hanson, E. Riseman, Beveridge, R. Kuman, "Model-directed mobile robot navigation ", IEEE Transactions on Systems, Man and Cybernetics,Vol. 20, No. 6, pp. 1352-1369, 1990

....of the robot is predetermined, and a pre storage of the entire path is needed. This is particularly problematic if the starting position of the robot may vary. On line methods (e.g [3] exist, but are commonly limited to the 2 D plane, or need the storage of a 3 D model of the environment (e. g [2, 4]) Also of relevance is work on image based visual servoing (see reviews in [7, 6] Below we present a method for visual navigation under the perspective imaging model. In our method the robot is instructed to reach a desired pose specified by a single image taken from that pose. The method then ....

C. Fennema, A. Hanson, E. Riseman, R. J. Beveridge, and R. Kumar. Model-directed mobile robot navigation. IEEE Trans. on Systems, Man and Cybernetics, 20: 1352--1369, 1990.

....Onoguchi et al. 15] use a stereo system to recover a depth map of the observed scene. In order to align the stereo image with the model a set of landmarks is first located by the system and their positions are used to derive the transformation that relates the model to the image. Fennema et al. [7]) compares the 3D models of the scene to sequences of 2D images. Gray scaled templates of selected landmarks are generated from the model, and the location of these landmarks is computed by means of correlation and tracking. The method presented in this paper does not generate signatures of the ....

C. Fennema, A. Hanson, E. Riseman, R. J. Beveridge, and R. Kumar. Model-directed mobile robot navigation. IEEE Trans. on Systems, Man and Cybernetics, Vol. 20, pp. 1352-- 1369, 1990.

....regained popularity, particularly in the area of computer vision. Most recent work in autonomous robot navigation has been done in structured indoor environments. A sampling of this work can be found in [Crowley, 1985, Tsubouchi and Yuta, 1987, Kriegman et al. 1989, Kosaka and Kak, 1992, Fennema et al. 1990, Ferrari et al. 1990, Atiya and Hager, 1993, D Orazio et al. 1992] Of the work in outdoor navigation, the following is associated most closely with this research. Dickmanns experimental five ton van VaMoRs navigated in outdoor environments, but remained (understandably) on paved roads, ....

C. Fennema, A. Hansen, E. Riseman, J. R. Beveridge, and R. Kumar. Model-directed mobile robot navigation. IEEE Transactions on Systems, Man and Cybernetics, 20(6):1352--1369, November/December 1990.

....planning, obstacle avoidance, place identification and position and motion estimation. A number of researchers provide their autonomous mobile robots with models instead of requiring their robots to perform mapping to learn models of their world (e.g. Kriegman et al. 1987; Kosaka and Kak, 1993; Fennema et al. 1990]) Others require mapping capability of their autonomous mobile robots. Numerous autonomous mobile robots capable of mapping have been developed, some of which are described in this section. Various approaches to mapping and spatial representations have been studied or implemented. The traditional ....

C. Fennema, A. Hanson, E. Riseman, J. R. Beveridge & R. Kumar, "Model-directed mobile robot navigation," in IEEE Transactions on Systems, Man, and Cybernetics, 20(6), Nov/Dec 1990, pp. 1352--1369.

....to its starting position and that the path is free of obstacles. Neither of these assumptions however are reasonable in real world situations. One way to eliminate the first assumption is to use additional information (e.g. landmarks) for correcting the current heading and position of the vehicle [6]. In order to account for the presence of unexpected obstacles we need some means to detect them and realize an appropriate avoidance maneuver. For obstacle detection we have adopted an idea proposed by Mallot [12] Obstacles are detected through the difference between a pair of stereo images ....

C. Fennema, A. Hanson, E. Riseman, J.R. Beveridge, and R. Kumar. Model-directed mobile robot navigation. Technical Report COINS TR 90-42, University of Massachusetts, Computer and Information Science, June 1990.

.... processing with control either concentrated on gaze control strategies separately from navigation [9, 10] or typically paid more attention to the issues of fast real time visual processing than to the control itself and control goals were usually very simple (e.g. keeping the vehicle on the road) [11, 12, 13, 14]. In this paper, in addition to the development of vision algorithms with real time performance, we investigate models of visually guided behaviors which tightly couple visual processing with control architecture. In general, we distinguish two kinds of behaviors: one where the perception action ....

....relative to its starting position and that the path is free of obstacles, neither of which is reasonable to assume in real world situations. One way to overcome the first assumption is to use additional information (e.g. landmarks) for correcting the current heading and position of the vehicle [13]. In order to account for the presence of unexpected obstacles we need some means to detect them and realize an appropriate avoidance maneuver. 4.1 Obstacle Detection For obstacle detection we adopted an idea proposed by Mallot [32] Obstacles are detected through the difference between a pair of ....

C. Fennema, A. Hanson, E. Riseman, J. Beveridge, and R. Kumar, "Model-directed mobile robot navigation," Tech. Rep. COINS TR 90-42, University of Massachusetts, Computer and Information Science, June 1990.

....some kind of internal spatial representation of the environment must be employed. These representations can be divided into three different paradigms: ffl Geometric ffl Sensor based 3 ffl Topological The geometric paradigm has been the traditional approach to spatial representation; see [42, 28, 1, 18, 48] for examples in mobile robotics. Here, low level geometric primitives are extracted from sensor data to build up high level abstractions of the environment. Thus, the world might be modeled using wire frame, surface boundary, or volumetric representations, for example. This approach is appealing ....

....remaining areas can be excluded from detailed mapping. Furthermore, dead reckoning is only employed within the context of a locale; no global coordinate system is used. Instead, the robot s metric position is specified relative to the local coordinate frame of the locale, if needed. Fennema et al. [28] use a similar approach in their spatial representation scheme in order to circumvent global inaccuracies arising from dead reckoning error. Likewise, we assert that if locales are of limited size, the robot will not travel long distances within the local coordinate system, and cumulative metric ....

[Article contains additional citation context not shown here]

Claude Fennema, Allen Hanson, Edward Riseman, J. Ross Beveridge, and Rakesh Kumar. Model-directed mobile robot navigation. IEEE Transactions on Systems, Man, and Cybernetics, 20(6):1352--1369, December 1990.

.... obvious way to reduce this problem is to observe the environment, and use the information in these observations to improve our estimate of R s position; cf. the work using Kalman filters (Kosaka Kak 1992; Cox Wilfong 1990) and other techniques (Smith Cheeseman 1987; Kuipers Levitt 1988; Fennema et al. 1990; Engelson 1992) We will model the environment using only a set of landmarks , each a (potentially visible) real world object at a known location; these objects could be doors, corners and pictures when specifying the hallways within building, or major buildings, junctions and prominent signs ....

Fennema, C.; Hanson, A.; Riseman, E.; Beveridge, J.; and Kumar, R. 1990. Model-directed mobile robot navigation.

.... understood that odometry is not sufficient because it leads to unbounded position error [13] A qualitative solution by tracking which side of landmark defined lines the robot is on is proposed in [20] If the robot is equipped with vision, then matching 3D models with 2D scene images is possible [9,16]. Because of the enormous computational requirements of using image data, using twodimensional laser range scans has also been proposed and demonstrated successfully [5] The robot Blanche assumes that a metric map of the environment consisting of polygonal obstacles is available, and it matches ....

C. Fennema, A. Hanson, E. Riseman, J. R. Beveridge, and R. Kumar. Model-directed mobile robot navigation. IEEE Transaction on Systems, Man, and Cybernetics, 20(6):1352--1369, 1990.

....by (A d A) evaluated at the final pose estimate. 3. 7 Least Squares Results Using Line Data The development of the algorithms presented in this paper are part of a larger effort to enable the UMASS robot Harvey to navigate the sidewalks and interior hallways of a part of the UMASS campus [6]. Consequently, results using line data are presented for both indoor hallway images and outdoor sidewalk images. Figures 9 and 10 are examples of outdoor and indoor images respectively. The indoor model was built by measuring distances with a tape measure and is accurate to approximately 0.1 feet ....

....Consequently, results using line data are presented for both indoor hallway images and outdoor sidewalk images. Figures 9 and 10 are examples of outdoor and indoor images respectively. The indoor model was built by measuring distances with a tape measure and is accurate to approximately 0. 1 feet [6]. The outdoor 3D model was built over two passes. In the first pass, blueprints of the campus, drawn to a scale of 40 feet to an inch, were used. Errors of up to 10 feet were found in the resulting 3D model; errors of this magnitude are unacceptable for our navigation goals. An error of i foot in ....

Fennema, C., A. Hanson, E. Riseman, J. R. Beveridge and R.Kumar, "Model-Directed Mobile Robot Navigation," IEEE Transactions on Systems, Man and Cybernetics, Vol. 20, No. 6, Nov./Dec. 1990.

....work was supported in part by the DARPA and Army ETL under contract DACA76 89 C 0017 and in part by NSF under grant DCR 8500332 1. INTRODUCTION The problem of mobile robot navigation has many approaches including methods of motion analysis[1] associative homing[2] and model based navigation[3]. This paper extends the work of Hong, Tan, et al. [4] on navigation through image based local homing. Homing is a navigation task in which the goal is one of a fixed set of target locations known to the robot. Unlike most homing systems, the navigation problem is treated here as a sequence of ....

....that best matched its current view. Both methods are classified as associative homing, because the actions are associated or derived in some manner from the stored set of patterns. Biological systems utilize associative processing and it is a natural paradigm for parallel hardware. Fennema et al. [3] use three dimensional models to generate projections of landmarks expected to 1 be seen from the estimated current location; the robot then servoes directly on the image features, tracking them via correlation. This kind of 3D model based navigation has the advantage that it can use precise ....

[Article contains additional citation context not shown here]

C. Fennema, A. Hanson, E. Riseman, J. Beveridge and R. Kumar. "Model-Directed Mobile Robot Navigation", To appear in IEEE Transactions on Systems, Man and Cybernetics.

....1 This work was supported in part by DARPA and TACOM under contract DAAE07 91 C R035 1. INTRODUCTION Robot navigation has been the focus of attention of many researchers in recent years. Some of the systems need to be very accurately calibrated in order to perform navigation successfully[5, 3]. Calibration is the process of determining the relationship of sensor output to the actual value of the input. In the system described in [3] it was necessary to position the robot camera system by hand to within 0.1 inches and 0.1 ffi of a known position in a world coordinate system. This ....

....of attention of many researchers in recent years. Some of the systems need to be very accurately calibrated in order to perform navigation successfully[5, 3] Calibration is the process of determining the relationship of sensor output to the actual value of the input. In the system described in [3], it was necessary to position the robot camera system by hand to within 0.1 inches and 0.1 ffi of a known position in a world coordinate system. This problem falls under the category of extrinsic camera calibration [10] There are at least two reasons why automatic calibration is needed. First ....

[Article contains additional citation context not shown here]

C. Fennema, A. Hanson, E. Riseman, J. Beveridge, and R. Kumar. Model-directed mobile robot navigation. Trans. Systems, Man, and Cybernetics, 20(6):1352--1369, 1990.

No context found.

Claude Fennema, Allen Hanson, Edward Riseman, J. R. Beveridge, and R. Kumar. Modeldirected mobile robot navigation. IEEE Trans. on Syst., Man, Cybern., 20(6):1352 -- 1369, November/December 1990.

....increase the probability of finding a near optimal solution. Our contribution has been to adapt these ideas to geometric matching problems [4] 5] 6] associated with object recognition. Our local search algorithms have been used for semiautonomous photo interpretation [7] robot navigation [8], 9] 10] and scene understanding [11] A matching system based upon the ideas presented here is now included in the KBVision system produced by AAI in Amherst Massachusetts. When estimates for object position and orientation are available, a 3D version finds optimal matches in domains with ....

Claude Fennema, Allen Hanson, Edward Riseman, J. R. Beveridge, and R. Kumar, "Model-directed mobile robot navigation, " IEEE Trans. on Syst., Man, Cybern., vol. 20, no. 6, pp. 1352 -- 1369, November/December 1990.

No context found.

C. Fennema, A. Hanson, E. Riseman, R. J. Beveridge, and R. Kumar. Model-directed mobile robot navigation. IEEE Trans. on Systems, Man and Cybernetics, 20: 1352#1369, 1990.

No context found.

C. Fennema, A. Hanson, E. Riseman, J. R. Beveridge, and R. Kumar. Model-Directed Mobile Robot Navigation. IEEE Trans. on Systems, Man, and Cybernetics, 20(6):1352--1369, November 1990.

No context found.

C. Fennema, A. Hanson, E. Riseman, J. R. Beveridge, and R. Kumar. Model-Directed Mobile Robot Navigation. IEEE Trans. on Systems, Man, and Cybernetics, 20(6):1352--1369, November 1990.

No context found.

C. Fennema, A. Hanson, E. Riseman, J.R. Beveridge, and R. Kumar. Model-directed mobile robot navigation. Technical Report COINS TR 90-42, University of Massachusetts, Computer and Information Science, June 1990. 66

First 50 documents

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