Odometric Information Research Articles

An Interactive Visual Navigation Method Using a Hand-drawn-Route-Map in an Unknown Dynamic Environment

In this paper, a visual navigation method based on Hand-drawn/sketched Route Map is proposed. This navigation method consists in considering the advance Hand-drawn-Route-Map and the effective image matching algorithm. At first, some key boot points are picked up from the hand-drawn route according to the principle of lesser deviation, so that the original route is divided into several segments. Then, in the process of the mobile robot's running along every segment, real-time image information from robot camera is matched with the corresponding one from advance Hand-drawn-Route-Map. In order to speed up image processing, a kind of prediction estimation method is proposed to find the most potential image. SURF algorithm is used to detect image features, where matching points can be found in term of KD-Tree. The projection transform matrix between the reference image and the real-time one is solved by RANSAC algorithm, in order to know the location of the reference image in the real-time one. With reference to the odometer and real-time image information, the robot can be roughly localized. And then, go on next segment, up to the last segment. At last, through a series of experiments, the advantage and efficiency of the new method in navigating and avoiding dynamic obstacles are testified with the imprecise route map.

A Visual Navigation Method Using a Hand-Drawn-Route-Map in Dynamic Environments

A visual navigation method of mobile robot based on hand-drawn route map and route is proposed.At first,some key boot points are picked up from the running path according to the principle of small deviation,so that the original path is divided into several segments.Then,during the mobile robot running along every segment,real-time image information from robot camera is matched with the corresponding one in the prior hand-drawn-route-map.In order to speeded up image processing,a kind of prediction estimation method is proposed to find the most potential image in the current field of vision. SURF(speed up robust feature) algorithm is used to detect image features.Matching points can be found rapidly in terms of KD-Tree.The projection transform matrix H between the referenced image and the real-time one is solved by RANSAC (RANdom SAmple Consensus) algorithm,in order to know the location of the referenced image in real-time one.With reference to odometer and real-time image information,the robot can be roughly localized.And then,for the next segment, the robot's running direction is computed according to its current rough localization until the last segment.At last,through a series of experiments,the advantage and efficiency of the new method in navigation and real-time dynamic obstacle avoidance are testified with the imprecise real map and route.

Anisotropic encoding of three-dimensional space by place cells and grid cells

The subjective sense of space may result in part from the combined activity of place cells, in the hippocampus, and grid cells in posterior cortical regions such as entorhinal cortex and pre/parasubiculum. In horizontal planar environments, place cells provide focal positional information while grid cells supply odometric (distance-measuring) information. How these cells operate in three dimensions is unknown, even though the real world is three–dimensional. The present study explored this issue in rats exploring two different kinds of apparatus, a climbing wall (the “pegboard”) and a helix. Place and grid cell firing fields had normal horizontal characteristics but were elongated vertically, with grid fields forming stripes. It appears that grid cell odometry (and by implication path integration) is impaired/absent in the vertical domain, at least when the animal itself remains horizontal. These findings suggest that the mammalian encoding of three-dimensional space is anisotropic.

Visual discrimination, sequential learning and memory retrieval in the Australian desert ant Melophorus bagoti

Bees, wasps and ants--so-called central-place foragers--need potent homing strategies to return to their nest. Path integration and view-based landmark guidance are the key strategies for the ants' navigation. For instance, they memorise different views in a sequence (sequential memory) but also have a step counter that informs them about the covered distance during each foraging trip (odometer). The sequential memory and the odometer information can act as contextual cues during travel for retrieving the appropriate stored view. When and which cue is used at different stages and lengths of the foraging trips is still unknown. In this study, we examined how the Australian desert ant Melophorus bagoti uses sequential memory and odometric information to retrieve visual memories. Using a set-up made out of channels and two-choice boxes (Y-mazes), we demonstrate first that M. bagoti foragers are able to learn and discriminate a variety of visual stimuli in a sequence of views along the inbound trip back to the nest. We then forced the homing ants to encounter a fixed sequence of two visual patterns during their inbound trips. By manipulating the position and distance of the visual stimuli and decision boxes, we could set the two contextual cues (sequential memory and odometer) into conflict. After the short 4-m outbound distance, a preference for odometric information as a contextual cue was found, but after the long 8-m outbound distance, ants relied primarily on their sequential memory retrieval. Odometer precision deteriorates with increasing travel distance, and accordingly, our findings imply that desert ants may be relying on the most reliable contextual cue for retrieving visual memories.

Design and Evaluation of a Yield Monitoring System for Pistachios

Information on the yield of individual trees within an orchard can be used to improve management practices. To achieve this goal, an automated yield monitoring system has been developed that is capable of correctly identifying trees in a commercial orchard and measuring yields of pistachio trees with small errors. Two commercial pistachio catch-frame harvesters were retrofitted and operated at approximately 75% of commercial harvesting field capacity. The system consists of a weighing bucket that measures individual tree yields in the transition between harvested trees. When the subsequent tree is shaken, nuts are weighed, dumped into the harvester's internal conveyor, and an embedded computer automatically records tree identification and yield. A DGPS-based automated row identification system was developed and tested throughout the 2006-2008 seasons. Inside dense orchard canopies, where a DGPS signal cannot be received with reliable accuracy, an algorithm was developed using real-time sensed odometric information to localize trees. System performance was tested in a commercial orchard by measuring yield and location of more than 70,000 trees over a six-year period. Tree identification and yield data were collected and mapped. Calibrations and validations of the weighing system resulted in weighing prediction standard errors of approximately 0.9 kg, while DGPS row identification was 100% accurate in over 480 estimates, and individual tree identification was highly accurate over 70,000 observations. The ability to measure individual tree yield provides previously unattainable information on tree productivity, facilitates precision field management, and enables the investigation of the mechanisms underlying fruit production and alternate bearing. The data recorded by the use of this system have been used in experimental studies to improve the productivity and management of pistachio orchards, investigate alternate bearing behavior, and provide the basis for developing efficient nutrient management protocols.

Consistent triangulation for mobile robot localization using discontinuous angular measurements

Localization is a fundamental operation for the navigation of mobile robots. The standard localization algorithms fuse external measurements of the environment with the odometric evolution of the robot pose to obtain its optimal estimation. In this work, we present a different approach to determine the pose using angular measurements discontinuously obtained in time. The presented method is based on an Extended Kalman Filter (EKF) with a state-vector composed of the external angular measurements. This algorithm keeps track of the angles between actual measurements from robot odometric information. This continuous angular estimation allows the consistent use of the triangulation methods to determine the robot pose at any time during its motion. The article reports experimental results that show the localization accuracy obtained by means of the presented approach. These results are compared to the ones obtained applying the EKF algorithm with the standard pose state-vector. For the experiments, an omnidirectional robotic platform with omnidirectional wheels is used.

Laser Pose Estimation and Tracking Using Fuzzy Extended Information Filtering for an Autonomous Mobile Robot

This paper presents methodologies and techniques for posture estimation and tracking of an autonomous mobile robot (AMR) using a laser scanner with at least three retro-reflectors. A three-point laser triangulation method is presented to find an initial posture of the robot and then a fuzzy extended information filtering (FEIF) method is used to improve the accuracy of the robot's posture estimation. With the odometric information from the driving wheels, a FEIF-based posture tracking algorithm is proposed to continuously keep trace of the robot's posture at slow speeds. Simulation and experimental results are conducted to show the efficacy and usefulness of the proposed methods.

Real-Time Tree Localization in Orchards

Tree localization is essential in numerous horticultural studies. In this study, a technique was developed to localize trees planted upon a uniform grid, such as in orchard and timber tracts manually planted at a specific tree spacing. Tree localization information is necessary for the development of yield, scout, or other management maps. Tree crops introduce new challenges with creating yield maps, as well as any management map, because accurate positional data cannot be obtained by the use of the Global Positioning System (GPS) when a clear view of the sky does not exist under the tree canopy. External to a canopy, a GPS can yield accurate positional data whereas underneath the canopy it cannot do so reliably. In this study, tree localization inside the canopy is obtained from combined information of partially known a priori tree plot characteristics and real-time sensed odometric information. The PR&S (periodic resetting and snap) method, developed partly based on the interpoint distance method, can be implemented to effectively localize every tree inside a given orchard. An encoder-based system and a DGPS-based system concurrently collected data in an open canopy orchard and in a dense canopy orchard to measure each system's accuracies in situ. The accuracy of each system was analyzed using a normalized root mean square error (RMSEn) and an absolute percentage error (APE). The encoder-based system in association with the PR&S method successfully localized all trees in both orchard configurations whereas the DGPS-based system did not produce enough accuracy in the dense canopy orchard.

Posture estimation and tracking of an autonomous mobile robot using a laser scanner with retro‐reflectors

This paper develops methodologies and techniques for posture estimation and tracking of an autonomous mobile robot (AMR) using a laser scanner with at least three retro‐reflectors. A novel three‐point triangulation method using the laser scanner is presented to find an initial posture of the robot and then an extended information filtering (EIF) method is used to improve the accuracy of the robot's posture estimation. The sensitivity to measurement errors with respect to different reflector arrangements is investigated as well. With the odometric information from the driving wheels, an EIF‐based posture tracking algorithm is proposed to continuously keep track of the robot's posture at slow speeds. Simulation and experimental results are compared to show the efficacy and usefulness of the proposed method. The proposed method can be applied to several mobile robots navigating over flat terrain in the areas of manufacturing factories, offices, hospitals, homes and etc.

Integrating two-dimensional paths: do desert ants process distance information in the absence of celestial compass cues?

When performing foraging trips desert ants of the genus Cataglyphis continuously process and update a ;home vector' that enables them to return to their nest on the shortest route. This capacity of path integration requires two types of information: (i) information about the travelling directions, and (ii) odometric information about the distances travelled in a particular direction. We have investigated how these two necessary pieces of information interact within the path integration processor. The specific question is: how do the ants process distance information if there is no simultaneous input from the sky compass available. Ants were trained to forage in a ;Z'-shaped channel system, the three segments of which joined at right angles. Individual animals were transferred from the feeder to a test field where their homing paths could be observed. In the crucial tests the middle segment of the maze was covered by orange Perspex that did not transmit the UV part of the spectrum, and thus precluded the perception of polarization patterns. Changes of the ant's processing of odometric information within this channel segment directly translate into a change in homing direction on the test field. The results indicate that the odometric information about travelling distance is largely ignored for path integration if there is no simultaneous input from the sky-view-based compass. They further show that idiothetic information cannot adequately substitute for the polarization compass to infer travelling directions.

Ant navigation en route to the goal: signature routes facilitate way-finding of Gigantiops destructor

We investigated in laboratory conditions how foragers of the tropical ant Gigantiops destructor develop individually distinctive landmark routes. Way-finding along a familiar route involved the recognition of at least two locations, nest and feeding site, and the representation of spatial relations between these places. Familiar visual landmarks were important both at the beginning and at the end of the foraging journey. A motor routine guided the ants at the start of their foraging path towards the first landmarks, which they learnt to pass consistently on the same side, before taking the next direction. At the last stage of the route, landmark recognition allowed them to pinpoint their preferred feeding site without using distant cues or odometric information. By contrast, ants en route to the goal were not systematically guided by a stereotyped sequence of snapshots recalled at each corresponding stage of the route. Each ant slalomed in an idiosyncratic distinctive way around different midway landmarks from a foraging excursion to the next, which induced a variability of the path shapes in their intermediate parts. By reducing the number of landmark recognition-triggered responses, this economical visuomotor strategy may be helpful in the Amazonian forest where many prominent landmarks are alike.

Encoding spatial information in the waggle dance

Apis mellifera bees execute waggle dances to recruit other bees to desirable food sources. Several components of the waggle dance are correlated with the direction of and the distance to food. Moreover, recruits use the spatial information encoded in the dance to locate the signalled food. However, although dance communication has been studied extensively, little is known about how the dancers combine the compass (direction) and the odometric (distance) information they acquire during the foraging flight. In the present study, we analysed the encoding of spatial information in the waggle dance by manipulating the navigational information provided to dancing bees. To this end, we took advantage of the bees' visually driven odometer. We found that the waggle dance basically encodes information on the distance gauged during the outbound (hive-to-food) flight. However, it does not necessarily refer to a global vector based on path integration of the outbound flight. Whenever the direction connecting the subjective food location and the hive does not match the direction of the global vector, dancers refer to a direction close to that of the shortcut connecting the actual food location and the hive. Moreover, in our experiments, this direction was close to that of the inbound (food-to-hive) flight, indicating that landmark-based information is computed during the inbound flight and that it may strongly affect the encoding of directional information in the waggle dance. Moreover, we found that the bees' experience of the terrain modulates the encoding of spatial information in the waggle dance, suggesting that interactions between path integration and visual landmarks are computed in the context of dance communication.

Odometry based pose determination and errors measurement for a mobile robot with two steerable drive wheels

A trigonometric method is proposed to calculate the pose (position and orientation) of a mobile robot, possessing a locomotion platform with a unique combination of two steerable and driven wheels and two caster wheels. The model of the locomotion platform is derived to deliver timely and accurate odometer information from measurements of drive wheel revolutions and steering angles. Non-systematic errors, mainly caused by wheel slippage, are detected by a new computation method based on the ratios between the two drive wheels' incremental distances. Real-world experiments validate the proposed model and the algorithm's ability to detect non-systematic errors.

Natural landmark-based autonomous vehicle navigation

This article describes a natural landmark navigation algorithm for autonomous vehicles operating in relatively unstructured environments. The algorithm employs points of maximum curvature, extracted from laser scan data, as point landmarks in an extended Kalman filter. A curvature scale space algorithm is developed to locate points of maximum curvature. The location of these points is invariant to view-point and sensor resolution. They can therefore be used as stable and reliable landmarks in a localization algorithm. This information is then fused with odometric information to provide localization information for an outdoor vehicle. The method described is invariant to the size and orientation of the range images under consideration (with respect to rotation and translation), is robust to noise, and can reliably detect and localize naturally occurring landmarks in the operating environment. The method developed is generally applicable to a range of unstructured environments and may be used with different types of sensors. The method is demonstrated as part of a navigation system for an outdoor vehicle in an unmodified operating domain.

Corridor navigation and wall-following stable control for sonar-based mobile robots

In this paper, a mobile robot control law for corridor navigation and wall-following, based on sonar and odometric sensorial information is proposed. The control law allows for stable navigation avoiding actuator saturation. The posture information of the robot travelling through the corridor is estimated by using odometric and sonar sensing. The control system is theoretically proved to be asymptotically stable. Obstacle avoidance capability is added to the control system as a perturbation signal. A state variables estimation structure is proposed that fuses the sonar and odometric information. Experimental results are presented to show the performance of the proposed control system.

Learning Geometrically-Constrained Hidden Markov Models for Robot Navigation: Bridging the Topological-Geometrical Gap

Hidden Markov models (HMMs) and partially observable Markov decision processes (POMDPs) provide useful tools for modeling dynamical systems. They are particularly useful for representing the topology of environments such as road networks and office buildings, which are typical for robot navigation and planning. The work presented here describes a formal framework for incorporating readily available odometric information and geometrical constraints into both the models and the algorithm that learns them. By taking advantage of such information, learning HMMs/POMDPs can be made to generate better solutions and require fewer iterations, while being robust in the face of data reduction. Experimental results, obtained from both simulated and real robot data, demonstrate the effectiveness of the approach.

An affective mobile robot educator with a full-time job

Sage is a robot that has been installed at the Carnegie Museum of Natural History as a full-time autonomous member of the staff. Its goal is to provide educational content to museum visitors in order to augment their museum experience. This paper discusses all aspects of the related research and development. The functional obstacle avoidance system, which departs from the conventional occupancy grid-based approaches, is described. Sage's topological navigation system, using only color vision and odometric information, is also described. Long-term statistics provide a quantitative measure of performance over a nine month trial period. The process by which Sage's educational content and personality were created and evaluated in collaboration with the museum's Divisions of Education and Exhibits is explained. Finally, the ability of Sage to conduct automatic long-term parameter adjustment is presented.

The segmentation of cursive handwriting: an approach based on off-line recovery of the motor-temporal information

This paper presents a segmentation method that partly mimics the cognitive-behavioral process used by human subjects to recover motor-temporal information from the image of a handwritten word. The approach does not exploit any thinning or skeletonization procedure, but rather a different type of information is manipulated concerning the curvature function of the word contour. In this way, it is possible to detect the parts of the image where the original odometric information is lost or ambiguous (such as, for example, at an intersection of the handwritten lines) and interpret them to finally recover a part of the original temporal information. The algorithm scans the word, following the natural course of the line, and attempts to reproduce the same movement as executed by the writer during the generation of the word. It segments the cursive trace where the contour shows the slow-down of the original movement (corresponding to the maximum curvature points of the curve). At the end of the scanning process, a temporal sequence of motor strokes is obtained which plausibly composed the original intended movement.

Using a Mobile Robot to Test a Model of the Rat Hippocampus

A model of how internal and external sensory information contribute to the firing of place cells in the rat hippocampus, and of how these cells contribute to the rat's spatial behavior, is tested on a miniature mobile robot. The experiments show that crude visual, odometric and short-range proximity information provided by sensors on the robot is sufficient to enable the formation of a robust spatial code within rectangular environments. They further show that the model of navigation can accurately return the robot to an unmarked goal location. Since the rat's perceptual systems are probably similarly crude, these results support our intuition that the model of hippocampal function is reasonable in not demanding too much of its inputs. The combined robotic and neuronal simulation can be used to make predictions regarding both electrophysiological and behavioral experiments. Finally, the model is applied to the neural basis of navigation in humans.

Correction of Odometric Errors in Mobile Robot Localization Using Inertial Sensors

A basic requirement for an autonomous mobile robot is its capability to elaborate the sensor measures to localize itself with respect to a coordinate system. The low cost positioning system here proposed corrects the partial odometric information, and the consequent long-term drift, exploiting the measures coming from an optical fiber gyroscope. The fusion algorithm is based on a suitably defined extended Kalman filter.