McKendall, R., Mintz, M.: Data fusion techniques using robust statistics. In Abidi, M., Gonzalez, R., eds.: Data Fusion in Robotics and Machine Intelligence. (1992) 211--244  Home/Search   Document Not in Database   Summary   Related Articles   Check

....second problem mentioned above that of cue selection. Some approaches explicitly describe interactions between visual modules by modeling the flow and priority of information between a variety of modules to specify how various cues are combined [4, 17, 18] Other methods use robust statistics [19, 7] or voting approaches [8] which take advantage of consistency among those cues which have not failed in selecting between cues or alternative interpretations. No aspect of current approaches for cue integration is geared towards timecritical implementations, which require all cue computations to ....

.... in a shape from shading calculation) and computations resulting in inappropriate local minima (such as not finding the best alignment that maps model features to image features) These errors are modeled using the probability distribution p(x y c ) although robust statistical procedures [27, 7] may be required for detection or exclusion of estimates that su#er from serious disruptions. y c e c p(y Y ) cc d c 2 d c 0 y c (b) Fig. 5. a) A model of errors in the partial results for cue c uses e c to represent the deviation in the partial result y c that will be ....

McKendall, R., Mintz, M.: Data fusion techniques using robust statistics. In Abidi, M., Gonzalez, R., eds.: Data Fusion in Robotics and Machine Intelligence. (1992) 211--244

.... using a more general form of the optical flow constraint equation to take radiometric variations into account [20] Second, and perhaps more generally applicable, would be the use of robust techniques for cue integration, which expect some (but not all) cues or computations to fail at any time [17]. 6 Conclusions We have presented a novel deformable model technique which uses residuals from a model based optical flow solution to refine the shape of the model. By using the relationship between the shape and motion parameterizations, small improvements to the parameters are made by ....

R. McKendall and M. Mintz. Data fusion techniques using robust statistics. In M.A. Abidi and R.C. Gonzalez, editors, Data Fusion in Robotics and Machine Intelligence, 1992.

....confidence set estimate for the pose of the mobile agent. The work in this dissertation constitutes the first application to the mobile robotics domain of optimal fixed size confidence interval decision theory developed by Zeytinoglu and Mintz [72, 73] McKendall [54] McKendall and Mintz [55], and Kamberova [40] The approach is evaluated in terms of theoretical capture probability and empirical capture frequency during actual experiments with the mobile agent. The method is compared to several other procedures including the Kalman Filter (minimum mean squared error estimate) and ....

....of V be taken into account 4. In cases where the CDF of V changes over time, how should statistical dependencies over time be taken into account The work presented in this chapter is based on decision theory developed by Zeytinoglu and Mintz [72, 73] McKendall [54] McKendall and Mintz [55], and Kamberova [40] which addresses these issues. The work in this chapter constitutes an application of these theoretical results to the mobile robotic domain. We discuss each of the above issues with respect to this domain. Confidence set size In a mobile robotic setting, the size of the ....

[Article contains additional citation context not shown here]

R. McKendall and M. Mintz. Data fusion techniques using robust statistics. In M. A. Abidi and R. C. Gonzalez, editors, Data Fusion in Robotics and Machine Intelligence, pages 211--244. Academic Press, 1992.

....features detected during the two scans are distinct. Owing to errors in robot localization, a simple check whether the observed features lie at the same world coordinates is insufficient. We are currently investigating an approach to solving this problem using confidence set based decision rules [MM92]. The idea is to compute fixed size confidence sets for the estimated locations of the candidate features; if these sets intersect, we accept the hypothesis that we have found two instances of the same feature. This approach requires delineation of the distributions of sensor and deadreckoning ....

R. McKendall and M. Mintz. Data fusion techniques using robust statistics. In M. A. Abidi and R. C. Gonzalez, editors, Data Fusion in Robotics and Machine Intelligence, pages 211--244, Academic Press, 1992.

No context found.

R. McKendall and M. Mintz. Data fusion techniques using robust statistics. In M. A. Abidi and R. C. Gonzalez, editors, Data Fusion in Robotics and Machine Intelligence, pages 211--244. Academic Press, 1992.

....performance guarantees using fixed geometry confidence intervals. The technique is optimal against bounded location and a very wide range of observation noise distributions. The work described in this document constitutes the first application of the decision theory developed in [15] [11] and [7] to this setting. It can also be extended to the multi dimensional Minimax Confidence Set Estimation (see [10, 8] The empirical results confirm the utility of the theory and attest to the value of the MCIE procedure. ....

R. McKendall and M. Mintz. Data fusion techniques using robust statistics. In M. A. Abidi and R. C. Gonzalez, editors, Data Fusion in Robotics and Machine Intelligence, pages 211--244. Academic Press, 1992.

....performance guarantees using fixed geometry confidence sets. The technique is optimal against bounded location and a very wide range of observation noise distributions. The work described in this document constitutes the first application of the decision theory developed in [8] 9] 6] [7] and [3] to a mobile robotic setting. The MCSE approach is very appropriate for the domain of mobile robot pose estimation. The empirical results confirm the utility of the theory and attest to the value of the MCSE procedure. ....

R. McKendall and M. Mintz. Data fusion techniques using robust statistics. In M. A. Abidi and R. C. Gonzalez, editors, Data Fusion in Robotics and Machine Intelligence, pages 211--244. Academic Press, 1992.

No context found.

R. McKendall and M.Mintz, Data fusion techniques using robust statistics, in: M.A. Abidi and R.C. Gonzalez, eds., Data Fusion in Robotics and Machine Intelligence (Academic Press, 1992) 211--244.

....performance guarantees using fixed geometry confidence sets. The technique is optimal against bounded location and a very wide range of observation noise distributions. The work described in this document constitutes the first application of the decision theory developed in [8] 9] 6] [7] and [3] to a mobile robotic setting. The MCSE approach is very appropriate for the domain of mobile robot pose estimation. The empirical results confirm the utility of the theory and attest to the value of the MCSE procedure. ....

R. McKendall and M. Mintz. Data fusion techniques using robust statistics. In M. A. Abidi and R. C. Gonzalez, editors, Data Fusion in Robotics and Machine Intelligence, pages 211--244. Academic Press, 1992.

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