SYSTEMS AND METHODS FOR INCREMENTALLY UPDATING A POSE OF A MOBILE DEVICE CALCULATED BY VISUAL SIMULTANEOUS LOCALIZATION AND MAPPING TECHNIQUES
First Claim
1. A method of updating a map for autonomous localization and mapping for a device, the method comprising:
- receiving an indication that a landmark has been observed, where the landmark is associated with a landmark pose, and where the map comprises one or more landmarks;
predicting a first device pose based on a change in pose from a prior update to device pose, wherein the change in pose is computed at least partly from dead reckoning data;
predicting a second device pose based on a visual measurement relative to the observed landmark and based on a stored estimate of landmark pose;
predicting a new pose for the device based on at least one of the first device pose and the second device pose; and
using a statistical filter to update the corresponding landmark pose estimate based at least partly on the new device pose.
7 Assignments
0 Petitions
Accused Products
Abstract
The invention is related to methods and apparatus that use a visual sensor and dead reckoning sensors to process Simultaneous Localization and Mapping (SLAM). These techniques can be used in robot navigation. Advantageously, such visual techniques can be used to autonomously generate and update a map. Unlike with laser rangefinders, the visual techniques are economically practical in a wide range of applications and can be used in relatively dynamic environments such as environments in which people move. One embodiment further advantageously uses multiple particles to maintain multiple hypotheses with respect to localization and mapping. Further advantageously, one embodiment maintains the particles in a relatively computationally-efficient manner, thereby permitting the SLAM processes to be performed in software using relatively inexpensive microprocessor-based computer systems.
97 Citations
27 Claims
-
1. A method of updating a map for autonomous localization and mapping for a device, the method comprising:
-
receiving an indication that a landmark has been observed, where the landmark is associated with a landmark pose, and where the map comprises one or more landmarks;
predicting a first device pose based on a change in pose from a prior update to device pose, wherein the change in pose is computed at least partly from dead reckoning data;
predicting a second device pose based on a visual measurement relative to the observed landmark and based on a stored estimate of landmark pose;
predicting a new pose for the device based on at least one of the first device pose and the second device pose; and
using a statistical filter to update the corresponding landmark pose estimate based at least partly on the new device pose. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10)
-
-
11. A circuit for updating a map for autonomous localization and mapping for a device, the circuit comprising:
-
a circuit configured to receive an indication that a landmark has been observed, where the landmark is associated with a landmark pose, and where the map comprises one or more landmarks;
a circuit configured to predict a first device pose based on a change in pose from a prior update to device pose, wherein the change in pose is computed at least partly from dead reckoning data;
a circuit configured to predict a second device pose based on a visual measurement relative to the observed landmark and based on a stored estimate of landmark pose;
a circuit configured to predict a new pose for the device based on at least one of the first device pose and the second device pose; and
a statistical filter that updates the corresponding landmark pose estimate based at least partly on the new device pose. - View Dependent Claims (12, 13, 14)
-
-
15. A computer program embodied in a tangible medium for updating a map for autonomous localization and mapping for a device, the computer program comprising:
-
a module with instructions configured to receive an indication that a landmark has been observed, where the landmark is associated with a landmark pose, and where the map comprises one or more landmarks;
a module with instructions configured to predict a first device pose based on a change in pose from a prior update to device pose, wherein the change in pose is computed at least partly from dead reckoning data;
a module with instructions configured to predict a second device pose based on a visual measurement relative to the observed landmark and based on a stored estimate of landmark pose;
a module with instructions configured to predict a new pose for the device based on at least one of the first device pose and the second device pose; and
a module with instructions configured to use a statistical filter to update the corresponding landmark pose estimate based at least partly on the new device pose. - View Dependent Claims (17, 18)
-
-
16. The computer program as defined in claim 115, wherein the device corresponds to a mobile robot, wherein the computer program is used by the mobile robot for navigation.
-
19. A circuit for updating a map for autonomous localization and mapping for a device, the circuit comprising:
-
a means for receiving an indication that a landmark has been observed, where the landmark is associated with a landmark pose, and where the map comprises one or more landmarks;
a means for predicting a first device pose based on a change in pose from a prior update to device pose, wherein the change in pose is computed at least partly from dead reckoning data;
a means for predicting a second device pose based on a visual measurement relative to the observed landmark and based on a stored estimate of landmark pose;
a means for predicting a new pose for the device based on at least one of the first device pose and the second device pose; and
a statistical filter configured to update the corresponding landmark pose estimate based at least partly on the new device pose. - View Dependent Claims (20)
-
-
21. A method of automatically updating a map for navigation with a Kalman filter, the method comprising:
-
providing the map, where the map comprises one or more landmarks, where a landmark includes a landmark pose estimate and an uncertainty measure, where a landmark pose estimate is initially based on a device pose estimate in a global reference frame corresponding to when the landmark was added to the map and where the landmark pose estimate is updated as the landmark is re-encountered;
predicting a landmark pose based at least in part on a reference device pose and one or more measurements of relative pose between the device pose and the landmark, wherein the one or more measurements of relative device pose are based at least in part on one or more visual measurements;
computing the uncertainty measure for error in the predicted landmark pose;
using the Kalman filter, wherein;
inputs to the Kalman filter comprise the predicted landmark pose, the uncertainty measure thereof, a previous estimate of landmark pose, and the uncertainty measure thereof; and
outputs of the Kalman filter comprise an updated estimate of landmark pose and the uncertainty measure thereof, and storing the updated landmark pose estimate and the updated uncertainty measure in the map such that the updated landmark pose estimate and the updated uncertainty measure can be updated again when the landmark is re-encountered. - View Dependent Claims (22, 23, 24, 25, 26, 27)
-
Specification