Crowley J. L. (1996). Mathematical Foundations of Navigation and Perception for an Autonomous Mobile Robot, In: Reasoning with Uncertainty in Robotics (L. Dorst, ed.), Springer-Verlag.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....Equations 3 and 4. The initial probability distribution over H is then calculated using P (h 0 j ) L(S j h 0 j ) P k L(S j h 0 k ) 6) After initialisation, localisation proceeds as follows. This algorithm is best explained as a three step predict match update cycle (after Crowley [4]) 4.2.2 Predict Step Firstly, the robot waits until it has travelled a further 50 cm, then the coordinates (x h i ; y h i ) of each of the prior hypotheses h i are translated to take into account the robot motion, using x h i (t) x h i (t 1) 4x; 7) y h i (t) y h i (t 1) 4y; 8) where ....

James L. Crowley. Mathematical foundations of navigation and perception for an autonomous mobile robot. In Workshop on Reasoning With Uncertainty in Robotics, 1995. Tutorial paper.

....to self positioning, and there are different sensor modalities as well, vision, laser range finders, sonars, etc. In this context we shall only consider vision, and in particular we use a single camera and landmarks. Landmarks can be used several ways. One approach, presented by Crowley96 [2], involves a Kalman filter framework for updating robot position using measured heading to a single landmark. It must be noted though, that robot position cannot be computed from heading to a single landmark, but only updated. Another use of landmarks involves triangulation based on three ....

....has to be changed. We thought we can use the angular separation between the predicted and the observed line of sight to the landmark as input to the pose computation procedure, and from this angle, we can compute a new robot position, by using a kalman filter framework as presented by Crowley96 [2]. Figure 5 shows a schematization of the pose computation procedure with both the proposed and the alternative method. Fig. 5. Schematization of the pose computation procedure with both the proposed and the alternative method. A drawback to the alternative method is that we can only compute a ....

J.L.Crowley. Mathematical foundation of navigation and perception for an autonomous mobile robot. In L.Dorst editor, Reasoning with Uncertainty in Robotics, Springer Verlag, 1996. This article was processed using the T E X macro package with SIRS98 style

....to self positioning, and there are di erent sensor modalities as well, vision, laser range nders, sonars, etc. In this context we shall only consider vision, and in particular we use a single camera and landmarks. Landmarks can be used several ways. One approach, presented by Crowley [3], involves a Kalman lter framework for updating robot position using measured heading to a single landmark. It must be noted though, that robot position cannot be computed from heading to a single landmark, but only updated. Another use of landmarks involves triangulation based on three ....

....has to be changed. We thought we can use the angular separation between the predicted and the observed line of sight to the landmark, as input to the pose computation procedure, and, from this angle, we can compute a new robot position by using a kalman lter framework as presented by Crowley [3]. Figure 5 shows a schematic representation of the posecomputation procedure with both the proposed and the alternative method. Figure 5: A schematic representation of the posecomputation procedure with both the proposed and the alternative method. A drawback to the alternative method is that ....

J.L. Crowley, Mathematical foundation of navigation and perception for an autonomous mobile robot, In: L.Dorst editor, Reasoning with Uncertainty in Robotics, Springer Verlag, 1996.

....and complementary aspects of uncertainty. 1.2 Uncertainty in robot localization In robot localization, the probabilistic framework is largely predominant. Bayesian methods are proposed in [ Burgard et al. 1998; Kortenkamp and Weymouth, 1994; Thrun et al. 1998 ] In [ Cox and Leonard, 1994; Crowley, 1995; Leonard et al. 1992; Moutarlier and Chatila, 1989; Smith and Cheeseman, 1986 ] Kalman filter methods are used to update the positions of geometric features of the environment in the presence of uncertainty. In [ Burgard et al. 1998; Kaelbling et al. 1996; Nourbakhsh et al. 1995; Simmons and ....

J. L. Crowley. Mathematical foundations of navigation and perception for an autonomous mobile robot. In Workshop on Reasoning with Uncertainty in Robotics (RUR'95). University of Amsterdam, December 4-6 1995.

....platform: State vector: x = x y ] T ; State Space Model: 8 : x = v cos y = v sin = 6) where (x; y) describe the robot position in pixels, denotes the robot orientation and (v; indicate the robot linear (forward) and angular velocities. The path, Psi , to be followed [2] is defined as a collection of points x Psi = x Psi ; y Psi ; Psi ) expressed in the same coordinate system and units as the robot state vector, x. At each time instant and depending on the robot position (x; y) the motion planning module, must determine a reference point on the ....

James L. Crowley. Mathematical foundations of navigation and perception for an autonomous mobile robot. Workshop on Reasoning with Uncertainty in Robotics 4-6 Dec. 1995; also in wwwprima. imag.fr/Prima/Homepages/jlc/ navigation.tutorial.word.ps, 1995.

....neural network model that performs adaptive probabilistic clustering of the robot s free configuration space, based on the localization information at any instant. We assume an autonomous mobile robot that moves on a Cartesian global map, while a dynamic localization procedure, e.g. Kalman filter [4], gives at any instant an estimation of the robot s current configuration, together with the associated uncertainty. This information is forwarded to the neural network which adapts its parameters appropriately. At any instant, the robot s total free space C free is partitioned into a number of ....

J. L. Crowley. Mathematical foundations of navigation and perception for an autonomous mobile robot. In Tutorial at the Workshop on Reasoning with Uncertainty in Robotics, University of Amsterdam, The Netherlands, Dec. 1995.

....argument of the exponent is a constant, given by the relation 2 1 z X C X X T = D D . This is the equation of an ellipsoid in N dimensions. Incidentally, the value 2 z d = is known as the Mahalanobis distance, and is a measure of the magnitude of the error vector, normalized by covariance [10]. For a given value of z, the cumulative probability of an error vector being inside the ellipsoid is P. For N=3 dimensions, the ellipsoid defined by z=3 corresponds to a cumulative probability P of approximately 97 2 . For a six dimensional pose X, the covariance matrix C X is 6x6, and the ....

J. L. Crowley, "Mathematical Foundations of Navigation and Perception for an Autonomous Mobile Robot," in Reasoning with Uncertainty in Robotics, Lecture Notes in Artificial Intelligence, L. Dorst, M. van Lambalgen, and F. Voorbraak, Ed., New York, Springer-Verlag, pp. 9-51, 1996.

....of the robot. As exact matches between the perceived and stored objects are unlikely, some form of approximate matching must be performed. Assuming normally distributed errors, Kalman filtering techniques can be used to correct odometric (or position estimation) errors based on perceptual inputs [Crowley, 1995]. Other such approaches include the fuzzy localization algorithm [Saffiotti Wesley, 1995] and the use of a hill climbing process to match evidence grids [Yamauchi Beer, 1996] On the other hand, animals demonstrate exceptional abilities for acquiring and responding to spatial information, ....

Crowley, J. (1995). Mathematical Foundations of Navigation and Perception for an Autonomous Mobile Robot. Pages 9--51 of: Proceedings of the International Workshop on Reasoning with Uncertainty in Robotics. Springer-Verlag.

.... generated in our experiments correspond more closely to those reported by Collett et al. 1986) Other Robot Localization Approaches Owing to the Kalman filtering framework, our computational model of hippocampal spatial learning is directly related to KF approaches for robot localization (Crowley, 1995; Leonard and Durrant Whyte, 1992) However, these KF based approaches require a sensor model of the environment (as shown in Figure 3) and often run into matching problems in environments with multiple identical landmarks and limited sensor ranges. The hippocampal model, on the other hand, ....

Crowley, J. (1995). Mathematical foundations of navigation and perception for an autonomous mobile robot. In Proceedings of the International Workshop on Reasoning with Uncertainty in Robotics, pages 9--51. SpringerVerlag.

....or may provide incorrect or inconsistent data. These sensor discrepancies have to be handled in some framework to allow the robot to visualize a unified view of its environment. Most advances in sensor fusion have largely entailed rediscovery and adaptation of techniques from estimation theory [17]. The early work characterized the sensor fusion problem as one of incremental combination of geometric information. Most of these techniques were ad hoc and did not characterize the uncertainties in the system. The first work on incorporating uncertainty in an explicit manner in sensor fusion was ....

....the use of Bayesian estimation theory and derived a combination function that was an equivalent form of Kalman Filter. This caused a rapid paradigm shift towards probabilistic estimation theories closely related to Bayesian estimation, maximum likelihood estimation and least squares methods [17]. While these methods have been successful in combining sensor data that determine a common subset of the robot state vector (e.g. robot location derived from GPS, odometry and gyroscope readings) there seems to be no obvious extension to multi modal sensor fusion processes. The advantage of ....

J.L. Crowley. Mathematical Foundations of Navigation and Perception For an Autonomous Mobile Robot. Reasoning With Uncertainty in Robotics 1995, Amsterdam, The Netherlands, 4-6 Dec, 1995, pp 9-51.

No context found.

Crowley J. L. (1996). Mathematical Foundations of Navigation and Perception for an Autonomous Mobile Robot, In: Reasoning with Uncertainty in Robotics (L. Dorst, ed.), Springer-Verlag.

No context found.

J. L. Crowley, \Mathematical foundations of navigation and perception for an autonomous mobile robot." Workshop on Reasoning with Uncertainty in Robotics, Dec. 4-6, 1995.

No context found.

J. L. Crowley, Mathematical foundations of navigation and perception for an autonomous mobile robot. Reasoning with Uncertainty in Robotics (1995), 9-51. Available at http://citeseer:nj:nec:com/crowley95mathematical:html.

No context found.

J.L. Crowley, "Mathematical foundations of navigation and perception for an autonomous mobile robot", Lecture notes in Artificial Intelligence 1093: Reasoning with Uncertainty in Robotics, L. Dorst, M. van Lambalgen, F. Voorbraak (eds.), Springer Verlag 1996.

No context found.

J.L. Crowley. "Mathematical Foundations of Navigation and Perception For an Autonomous Mobile Robot," Reasoning With Uncertainty in Robotics 1995, Amsterdam, The Netherlands, 4-6 Dec, 1995, pp 9-51.

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