Sensor resetting localization for poorly modelled mobile robots

Lenser, Scott; Veloso, Manuela

doi:10.1109/robot.2000.844766

Cited by 199 publications

(129 citation statements)

References 3 publications

Supporting

Mentioning

125

Contrasting

Unclassified

Order By: Relevance

“…Sensor Resetting Localization (SRL) [22] evaluates this mismatch by sampling directly from the observation likelihood function P (l|s), and adds the newly generated samples to the particle filter proportional to the mismatch. KLD Resampling [10] similarly evaluates the mismatch by sampling from the odometry model P (l i |l i−1 , u i ) and computing the overlap with the belief Bel(l).…”

Section: Sensing Mismatch Based Selectionmentioning

confidence: 99%

Multi-sensor Mobile Robot Localization for Diverse Environments

Biswas

Veloso

2014

RoboCup 2013: Robot World Cup XVII

View full text Add to dashboard Cite

Abstract. Mobile robot localization with different sensors and algorithms is a widely studied problem, and there have been many approaches proposed, with considerable degrees of success. However, every sensor and algorithm has limitations, due to which we believe no single localization algorithm can be "perfect," or universally applicable to all situations. Laser rangefinders are commonly used for localization, and state-of-theart algorithms are capable of achieving sub-centimeter accuracy in environments with features observable by laser rangefinders. Unfortunately, in large scale environments, there are bound to be areas devoid of features visible by a laser rangefinder, like open atria or corridors with glass walls. In such situations, the error in localization estimates using laser rangefinders could grow in an unbounded manner. Localization algorithms that use depth cameras, like the Microsoft Kinect sensor, have similar characteristics. WiFi signal strength based algorithms, on the other hand, are applicable anywhere there is dense WiFi coverage, and have bounded errors. Although the minimum error of WiFi based localization may be greater than that of laser rangefinder or depth camera based localization, the maximum error of WiFi based localization is bounded and less than that of the other algorithms. Hence, in our work, we analyze the strengths of localization using all three sensors -using a laser rangefinder, a depth camera, and using WiFi. We identify sensors that are most accurate at localization for different locations on the map. The mobile robot could then, for example, rely on WiFi localization more in open areas or areas with glass walls, and laser rangefinder and depth camera based localization in corridor and office environments.

show abstract

Section: Sensing Mismatch Based Selectionmentioning

confidence: 99%

Multi-sensor Mobile Robot Localization for Diverse Environments

Biswas

Veloso

2014

RoboCup 2013: Robot World Cup XVII

View full text Add to dashboard Cite

show abstract

“…These techniques can first be classified according to the type of sensor used for localization and the underlying representation of the map m. For example, several authors used features extracted from camera images to calculate the likelihood of observations. Typical features are average grey values [3], color values [10], color transitions [12], feature histograms [22], or specific objects in the environment of the robot [11], [17], [18]. Additionally, several likelihood models for Monte-Carlo localization with proximity sensors have been introduced [2], [8], [19], [20].…”

Section: Related Workmentioning

confidence: 99%

“…In the limit, i.e., for a perfect sensor, the number of required particles becomes infinite. To deal with this problem, Lenser and Veloso [11] and Thrun et al [21] introduced techniques to directly sample from the observation model and in this way ensure that there is a critical mass of samples at the important parts of the state space. Unfortunately, this approach depends on the ability to sample from observations, which can often only be done in an approximate, inaccurate way.…”

Section: Related Workmentioning

confidence: 99%

Improved likelihood models for probabilistic localization based on range scans

Pfaff

Plagemann

Burgard

2007

2007 IEEE/RSJ International Conference on Intelligent Robots and Systems

View full text Add to dashboard Cite

Abstract-Range sensors are popular for localization since they directly measure the geometry of the local environment. Another distinct benefit is their typically high accuracy and spatial resolution. It is a well-known problem, however, that the high precision of these sensors leads to practical problems in probabilistic localization approaches such as Monte Carlo localization (MCL), because the likelihood function becomes extremely peaked if no means of regularization are applied. In practice, one therefore artificially smoothes the likelihood function or only integrates a small fraction of the measurements. In this paper we present a more fundamental and robust approach, that provides a smooth likelihood model for entire range scans. Additionally, it is location-dependent. In practical experiments we compare our approach to previous methods and demonstrate that it leads to a more robust localization.

show abstract

“…Odometry is used to decompose information that is dependent: although both the players process and the ball process require the current pose of the robot, they can run in parallel to the self-localization process, because the odometry can be used to estimate the spatial offset since the last absolute localization. This allows running the ball modeling with a high priority, resulting in a fast update rate, while the self-localization can run as a background process to perform a computationally expensive probabilistic method as, e.g., the one described in [4] Currently, it is not known which process layout will be the more successful one. The Darmstadt Dribbling Dackels are using a third approach that is a compromise between the two discussed here, and all three will compete against each other at the German Open.…”

Section: Process-layoutsmentioning

confidence: 99%

An Architecture for a National RoboCup Team

Röfer

2003

RoboCup 2002: Robot Soccer World Cup VI

View full text Add to dashboard Cite

Abstract. This paper describes the architecture used by the GermanTeam 2002 in the Sony Legged Robot League. It focuses on the special needs of a national team, i.e. a "team of teams" from different universities in one country that compete against each other in national contests, but that will jointly line up at the international RoboCup championship. In addition, the tools developed by the GermanTeam will be presented, e.g. the first 3-D simulation used in the Sony Legged Robot League.

show abstract

Sensor resetting localization for poorly modelled mobile robots

Cited by 199 publications

References 3 publications

Multi-sensor Mobile Robot Localization for Diverse Environments

Multi-sensor Mobile Robot Localization for Diverse Environments

Improved likelihood models for probabilistic localization based on range scans

An Architecture for a National RoboCup Team

Contact Info

Product

Resources

About