[Downloaded from www.aece.ro on Thursday, December 22, 2016 at 09:21:15 (UTC) by 14.139.181.229.

Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 16, Number 4, 2016

A Novel Keep Zero as Zero Polar Correlation
Technique for Mobile Robot Localization using
LIDAR
Rakesh Kumar SIDHARTHAN1, Ramkumar KANNAN1, Seshadhri SRINIVASAN2,
Marius Mircea BALAS3
1
Electric Vehicle Engineering and Robotics (EVER) Lab, School of Electrical and Electronics
Engineering, SASTRA Univeristy, Thanjavur 613401, Tamil Nadu, India
2
Department of Engineering, University of Sannio, Benevento, Italy-82100
3
Department of Automatics and Applied Informatics, Aurel Vlaicu University of Arad, Romania

1
Abstract—Sensor fusion based localization techniques often widely investigated. Exteroceptive sensors like Light
need accurate estimate of the fast and uncertain scene change incident detection and ranging (LIDAR) are used for
in environment. To determine the scene change from two environment scanning [7-8]. This is mainly due to their
consecutive LIDAR scans, this paper proposes a novel
good sensitivity and fast scan rate [1].
technique called ‘keep zero as zero’ polar correlation. As it
name implies any zero in the scan data is kept isolated from Feature extraction algorithms filter environment scan data
scene change estimation as it do not carry any information based on specific property (e.g., geometry [2, 5-6, 9]), to
about scene change. Unlike existing techniques, the proposed reduce the dimensionality of the map build using them.
methodology employs minimization of selective horizontal and Hence, accuracy of the map very much depends on the
vertically shifted sum of difference between the scans to feature selection criteria [10-11] and a wrong selection may
estimate scene change in terms of rotation and translation.
result in loss of accuracy in the map. To overcome this
Minimization of the proposed correlation function across the
specified search space can guarantee an accurate estimate of challenge, adaptive [11], and graphical techniques [12-13]
scene change without any ambiguity. The performance of the have been proposed for scan data selection. The feature
proposed method is tested experimentally on a mobile robot in maps thus obtained are associated using features available in
two modes depending on the scene change. In the first mode, common among them. Typically, association techniques in
scene change is detected using dynamic LIDAR, whereas static literature use geometric techniques [14-15] that are
LIDAR is used in the second mode. The proposed methodology
considered to be computationally cumbersome. Finally, the
is found to be more robust to environmental uncertainties with
a reliable level of localization accuracy. change in associated features is used to determine the
changes in the scene.
Index Terms—correlation, mobile robots, pattern matching, A survey of existing literature reveals that the feature
sensor fusion, simultaneous localization and mapping. extraction techniques used will result in loss of scan data
and the scene change information carried by them [11].
I. INTRODUCTION Association techniques with the reduced feature will result
Scene change estimation (SCE) is an important task in in multiple association possibilities and demands additional
sensor fusion based localization algorithms [1]. An accurate validation mechanism [16-18]. Therefore, existing
SCE is important for reliable localization and consists of the association techniques, requires more iterations and
following steps: environment scanning [2], feature increases complexity as the number of feature becomes
extraction [3-4], data association [5-6] and estimate changes large.
[7] in the associated feature. In SCE, errors introduced in To overcome the existing challenges, the proposed work
feature extraction and association mismatch deteriorates the develops a robust and computationally fast SCE algorithm
localization accuracy. The uncertainty and dynamic changes using a novel LIDAR based selective polar cross correlation
in environmental scenes due to appearance/disappearance of technique. This method in a novel way determines the
features from the robot’s vicinity compounds the SCE common features in the maps and employ cross correlation
complexity further. Therefore, reliable feature extraction techniques to estimate the scene change in them. The cross
and data-association techniques are required for accurate correlation technique [19] involves minimization of
localization. mismatch between the consecutive maps by shifting within a
Objective of this investigation is to propose a SCE bounded search space. Furthermore, by carefully studying
algorithm that is more robust to the uncertainties in the robot dynamics, the constraint in the search space can be
environmental features and computationally simple to be determined in advance and which in turn can enhance the
adapted in robots requiring run-time localization. The robustness of the proposed technique even amidst
problem of SCE based localization has attracted significant environmental changes.
research attention. The choice of sensors for various Main contributions of this investigation are: (i) new
environments and techniques for various steps in SCE are filtering technique called keep zero as zero (KZZ) to filter
the old/new features and to extract the common features
from the LIDAR scans, (ii) cross correlation based

Digital Object Identifier 10.4316/AECE.2016.04003
15
1582-7445 © 2016 AECE

illustrate the performance of the algorithm are presented in Section 5. Hence. The pose of a mobile robot ( Rk ) is defined by its center position ( xrt . rr . The LIDAR scan is a discrete set of Euclidean distance of III. Let rs1 (n) and rs 2 (n) be the two consecutive LIDAR scans Figure. 2. the angular resolution (  res ) of the LIDAR is given static mode. Section 4 describes the methodology rs 2 ( s )  f tr (rs1 ( s   ). Section 2 describes the problem formulation and presents the theory  res  ( max   min ) / N s (5) behind the sensor fusion based localization briefly. the predicted state has to be corrected using SCE between two successive LIDAR scans. 2016 at 09:21:15 (UTC) by 14. December 22. Using this pose change the robot can be localized by defining the robot pose at the given time instant‘t’ as in (2).229. The The relationship between the local scans of the LIDAR can formulation of the proposed KZZ correlation for SCE is be defined in terms of relative robot pose as in (6).aece..  )  (S  T)(S  T)t (7) this investigation are discussed in Section 6. t t r cos  s   rr cos   II. 3. 2) of the robot pose change ( R ) as in (1).ro on Thursday. The translation function ( f tr ) in polar coordinate is given by The experimental results on a wheeled mobile robot to (7) using the scan matrix ( S ) and translation matrix ( T ). the odometer for state Thus. Scene change in LIDAR scans acquired when the robot is at R1 and R2 poses respectively. The relative changes are described in calculated from its indices ( n ) as in (4) terms of rotation and translation between those common 16 . Number 4. 2016 minimization algorithm over the specific search space for rst ( n)  r ( n) | n  [0 N s ]. The angles of the scanned can be determined by observing the relative changes obstacles (  s ) are a set of monotonically increasing between the associated features available in common constants equal to the number of scanned data and between the scans.   R     1  s  (1) r  r  rs   xrt   rr cos t 1   xrt 1  Rt   yrt    rr sin t 1    yrt 1  (2) Figure 1. The scene change between the scans are described by the relative rotation (  s ) and translation ( rs ) which is numerically equal and opposite (see Fig. PROPOSED SCE METHODOLOGY the obstacles from the robot center as in (3) over the entire The scene change between the successive LIDAR scans span of scan area (  min  max  ). n   t  1. (3) data association (iii) experimental set ups to validate the proposed technique to localize the mobile robot with sensor  s  n res   min (4) placed onboard as well as the sensor placed externally in a Where. the past scan is rotated using the estimated rotation [20] to align it with the present scan. 1. the robot wheel displacement is measured by the change between the consecutive scans.139. the proposed work aims to estimate the relative prediction and LIDAR for state correction.181. Where. [Downloaded from www.] Advances in Electrical and Computer Engineering Volume 16. presented in Section 3. Proposed SCE in sensor fusion based localization framework t     t 1  This investigation proposes a SCE technique that uses KZZ horizontal cross correlation (KZZ-HXC) to determine the relative rotation (  s ) between the present and past LIDAR scan. Then the relative translation ( rs ) between these oriented scans is estimated using KZZ vertical cross correlation (KZZ-HXC) and finally using this estimated scene change the change in robot pose for state correction is calculated. Finally the conclusion and the future prospects of f tr ( rs . Next. in (5) The rest of the paper is organized as follows. PROBLEM FORMULATION S and T   r sin  s   rr sin   Typically. odometer and the past robot pose measurements are used to predict the current pose of the robot as in Fig. 2.  ) (6) to evaluate the performance of the proposed SCE technique. Redistribution subject to AECE license or copyright. This predicted robot pose is subjected to error due to wheel slip that is time-varying and complex to model. yrt ) in a 2 dimensional space and its heading ( t ). In state rotation (  ) and translation ( rr ) of the robot using scene prediction. sensor fusion based localization [1] techniques employ two sensors namely. rr .

k  0   n 0  (ii) association and (iii) estimation of changes in the common features between successive scans. Consequently. Departing from the correlation shifted scan will be zero padded to compensate for the techniques found in literature [21-23].   rs1 (n)  rs 2 (n  k ) . kmin  arg min k [ H c (k )] (10) The new features thus obtained do not have any common feature with past scan. line segments etc has been reported [3.e. 2. Hence. KZZ is a zero in the past scan and non-zero in the current scan indicates a new feature. determined by minimizing KZZ horizontal cross correlation 25] to be used for scene change estimation.229. The key idea of minimal value of H c as in (10) over the search space. 4.ro on Thursday. the old features are  s   res  kmin (11) absent in the current scan. the magnitude of vertical shift (scan translation) required to match the scans is used to estimate the robot translation in Figure 4. scan and a zero in the current scan indicate an old feature. The horizontal shift will be uniform whereas the A bound in shifting sample size (k) to calculate the vertical shift uses a complex polar transform as in (6) using correlation coefficient and usage of all the common features the robot heading (  ). Number 4. to estimate the changes in the common features. the proposed SCE methodology uses two KZZ-HXC. Finally.   rs1 (n)  rs 2 (n  k ) . k   (8) termed as ‘Keep Zero as Zero (KZZ)’ polar cross  Ns  correlation.aece. the identification step of horizontal shift ( kmin ) to be compensated to match the involves the detection of common features among two scans is determined. using the calculated coefficient ( H c ) as a LIDAR scans have old. k  0  H c (k )   Nn s k k  (9) A. Two successive Finally. In the association step. common features among the two scans are identified by eliminating these new and old features that are not common in successive scans. In the SCE operation. SCE Operation SCE is a three step process involving (i) identification. the bounded as the relative rotation between the LIDAR scans proposed KZZ-HXC can able to estimate the relative are limited due to high scan rate of LIDAR. A small set of features like The rotation angle between two LIDAR scans can be corners. rather than a small set of features ensures global minima of B. new and common features as shown deviation metric for scan match (association). the amount of shift required to associate the common features is computed by minimizing the cross- correlation coefficient. the range data will be discarded (kept aside) throughout the correlation operation. The estimation respectively. Hence the proposed technique is K  k | k  [  N s N s ].e. a negative sign indicates that the scan is rotated types of KZZ cross correlation technique namely (i) KZZ. However. The scan matching is indicated by a successive scans using the proposed KZZ. The degree of association between common features is determined by calculating the cross correlation coefficient. the zero in horizontal cross correlation coefficient ( H c ) using (9). the proposed method length of the stationary scan ( rs1 ). This often coefficient ( H c ) over the discrete search space ( K ) as in results in complex and multiple iterations to associate and to (8). Horizontal cross correlation coefficient (Hc) KZZ-VXC.139. the relative rotation (  s ) is calculated using (11). The proposed correlation technique uses the polar coordinate [24] system wherein the horizontal axis represents the scan angle in degrees and the vertical axis represents the magnitude of obstacle range in meters. In the proposed KZZ-HXC.181. anticlockwise (left shift) whereas a positive sign indicates a HXC for rotation and (ii) KZZ-VXC for translation clockwise rotation (right shift) as indicated in Fig 3. In the KZZ approach. Illustration of KZZ-HXC features after retranslating the present scan. 9. the identified common features between the successive scans are associated using cross- correlation technique. whereas a non-zero in the past Finally. the magnitude of horizontal shift (scan rotation) required to match or align the two consecutive scans is used to estimate the robot rotation. 11. Redistribution subject to AECE license or copyright. Similarly. the search space is determine scene change between those features. December 22. the magnitude in Fig. In the proposed rotation without any ambiguity within the bounded search 17 . these padded uses shifted sum of difference to calculate the correlation zeros (new features) are not used in calculating the KZZ coefficients with the exception on zero values i. [Downloaded from www. 2016 at 09:21:15 (UTC) by 14.] Advances in Electrical and Computer Engineering Volume 16. KZZ Horizontal cross correlation (KZZ-HXC) H c as illustrated in Fig. It involves shifted matching of common features in present LIDAR scan with the static scan (i. Unlike conventional technique. A minimum cross-correlation coefficient indicates an accurate matching between common Figure 3. 2016 features. past scan). edges. Thus.

Illustration of KZZ-VXC used to determine the change in robot pose and to localize the robot. ls . The match between the scans is vertical space between the scans where as the horizontal indicated by minima in the vertical correlation coefficient space are distorted because of discontinuous features in the scans. the present scan data are vertically shifting (see coefficient converges smoothly to minima which give the Fig.5686m ) which is maximum The correlation coefficient is calculated by the absolute allowable deviation of the estimated robot pose from its difference between the present and past scan for a particular actual pose. the rs 1 ( n )  0 localization performance can be evaluated by benchmarking rs 2 ( n )  0 a threshold for MAE ( Al  0.229. PERFORMANCE EVALUATION The proposed SCE technique is used to estimate the scene change between the successive LIDAR scans which can be Figure 5.  ) (14) localization accuracy of the proposed SCE. Finally. Vertical cross correlation coefficient (Vc) IV. the past scan and brings the MAE to a minimal value. vertical cross correlation can be estimated using the proposed KZZ-VXC. 2016 space. This is due to existence of clear transform as in (7). retranslate the present scan such that it exactly aligns with both these regions don’t participate in calculating Vc . non-zero region ( rs1 (n)  0 ^ rs 2 (n)  0 ) in present scan and rs 2T  ftr (rs 2 ( s   ). The occurrence of new feature can be inferred by a KZZ-VXC) by the proposed SCE methodology as in (17). Hence.181. Once the scans the given local scans with respect to the applied translations are aligned the deviations between them will be due to the is illustrated in Fig. Hence. This NS indicates a MAE between scans can able to validate the Vc (ls )  n 0 rs1 (n)  ftr (rs 2 . 6. magnitude of translation. The present scan is retranslated using the estimated corresponds to presence of common features between both scene change estimated (  s in KZZ-HXC and rs in scans. a magnitude of translation The threshold ( Al ) is calculated by Euclidean distance of 18 . Thus. the performance of the SCE technique is Unlike KZZ-HXC. For the given parameters. [Downloaded from www. the range vector ( rs ) of the LIDAR evaluated in terms of its localization accuracy using mean scans is in continuous space. l   (13) Figure 6. In this technique. Redistribution subject to AECE license or copyright. This relative change in translation ( rs ) correlation coefficient ( H c ). ( Vc ). The coefficient is calculated only past scan ( rs1 ) and the retranslated present scan ( rs 2T ) as in non-zero region ( rs1 (n)  0 ^ rs 2 (n)  0 ) which in (16). hence the proposed KZZ-VXC average error (MAE) [26] as performance metric and its is evaluated by using a user defined discrete translation complexity in terms of its computation time for the given search space ( L ) as in (13).] Advances in Electrical and Computer Engineering Volume 16. the MAE is calculated as arithmetic mean of the be calculated using (14) for each element ( ls ) over this user absolute error between the common features identified in the defined search space ( L ). The proposed KZZ-VXC is illustrated MAE  NS n 1 rs1 (n)  rs 2T (n) (17) rs1 ( n )  0 in Fig.139.aece. However. 5) to match the past scan data using the complex relative change in scans. The span ( Sl ) and resolution specifications. 2016 at 09:21:15 (UTC) by 14.  ) (16) a lost feature by non-zero region ( rs1 (n)  0 ^ rs 2 (n)  0 ) in NS 1 past scan. Unlike the horizontal cross translation alone. December 22. for which the correlation coefficient ( Vc ) is minimal will C. rs 2 (n)  rs 2 (n  kmin ) (12) L  ls  lRl | l  [  Sl Sl ]. KZZ Vertical cross correlation (KZZ-VXC) result in the estimated robot translation as in (15).ro on Thursday. These zeros are discarded in KZZ-VXC as constrained in (14). ( Rl ) of the search space are defined by the user based on The search space specifications for KZZ-HXC and KZZ- the speed and size of the robot platform respectively. the translation for which the vertical correlation coefficient is minimal indicates the robot rs  arg min lr [Vc (lr )] (15) translation. VXC technique are determined by the robot dimension and The KZZ based vertical correlation coefficient ( Vc ) can its navigation velocity as shown in Table I. 7 and from which it is observed that the first part of rs 2 ( n )  0 KZZ infers about the addition of new features to the scan Accurate scene change estimation can effectively and last part indicates the loss of old features. The The present scan is rotated as in (12) using the estimated variation of vertical cross correlation coefficient ( Vc ) for robot rotation in order to align both scans. rr . Number 4.

implementation.139. 8. PARAMETERS FOR EXPERIMENTAL SETUP In both modes.2sec/ scan scans (‘s’) having MAE lesser than the threshold indicates LIDAR the efficiency of the proposed technique. Number 4. From the information that evaluated in terms of its computational time which can comes out of this scene change the robot pose can be illustrate the efficiency of the technique for real time estimated using the proposed SCE technique.47 m  0. 2016 the robot’s top view (plan) dimension (Ref Table I) which in The algorithms are written on MATLAB® and executed in this particular experimental set up is a rectangular surface.229. The number of LIDAR Scan Rate 0. The KZZ-VXC Resolution local scans are the maps of the environment with robot Search Space Range 0. The computational time taken by the proposed SCE is the sum of time taken to acquire the LIDAR scans and time taken for KZZ-HXC and KZZ-VXC algorithms.2s.327  0. The duty observed if the MAE is above the threshold as shown in Fig.05m CPU and sends the scan data along with the time stamp. The robot can be the robot using the estimated scene change between the operated and the performance of the proposed SCE can be present and past LIDAR scans.25m per scan position as reference. (Hokuyo URG.32m robot navigates. Computation time for proposed SCE technique Figure 7. Mobile robot with LIDAR sensor in each LIDAR scans as shown in Fig. Redistribution subject to AECE license or copyright.01m / s estimate the localization accuracy. This illustrates the proposed SCE to be good enough for deployments requiring The performance of the proposed SCE technique is run-time localization. the proposed SCE is used to estimate the KZZ-HXC scene change between the local scans acquired from the Search Space Range 60o per scan LIDAR fitted on top of the mobile robot (see Fig. During experiments different (Coroware® – CoroBot-2WD) and Hokuyo® URG-04LX. the proposed SCE is used to estimate the Component Parameter Value scene change between the consecutive LIDAR scans as the Mobile Robot Dimension 0.352o 04LX-UG01) A. The LIDAR position varies dynamically as the robot navigates creating a scene change The performance of the proposed SCE technique is between successive local scans. the robot’s central processing unit (CPU) which has direct The MAE below this threshold indicates that the rectangular access to the robot hardware.181. December 22. The experiments conducted with the given specifications (Ref Table. ratio for individual wheels can be send from MATLAB® 7. I) clearly shows that the average computational time for the KZZ-HXC to be 0. illustrated by its ability to retranslate the local scans using the estimated scene change. Illustration of threshold for localization accuracy Two modes of experiments namely (i) Dynamic Scanner (DS) mode and (ii) Static Scanner (SS) mode are conducted.ro on Thursday. 10). EXPERIMENTAL RESULTS change.aece. Figure 8. by the robot movement in an unknown environment and the 19 . Only a small and acceptable variance in the computational time has been observed due to variations in common features present Figure 9. This can be used to validate the proposed SCE to localize environment using Phidget® APIs. Experimental Setup-1 DS Mode Scan Area 30o to 270o Search Space Index 371 In DS mode. Angular Resolution 0. [Downloaded from www. 9). TABLE I. This gives the total computation time to be 0. With the estimated scene V.238s along with the LIDAR scan time of 0. the present scan is transformed accurately to match The performance of the proposed SCE is evaluated from the past scan as illustrated in both polar and Cartesian experiments conducted using a two wheeled mobile robot coordinates (see Fig. conditions were tested considering the scenarios envisaged UG01type LIDAR with the specifications shown in Table I.036s. 2016 at 09:21:15 (UTC) by 14. visualized from a remote computer connected to the robot’s CPU via remote connectivity using wireless network.021s and for KZZ-VXC is 0. MAE along with its threshold is used to (CoroBot-2WD) Set Speed 0. The Search Space Span 50 LIDAR communicates serially via USB port of the robot’s Search Space 0. The robot can be driven by surface of the estimated robot pose is in overlap with that of manipulating the duty ratio of the Phidget® motor controller the actual robot pose and wider deviations have been to control the robot wheel speed and direction.] Advances in Electrical and Computer Engineering Volume 16.

Redistribution subject to AECE license or copyright. 2016 at 09:21:15 (UTC) by 14. results illustrates that the MAE lies predominately well (i) Scenario-1: rotation only (ii) Scenario-2: translation only (a) Raw LIDAR scans (b) Retranslated scans (c) MAE in robot localization (iii) Scenario-3: both rotation and translation Figure 11. December 22. Scan matching in polar and rectangular coordinates proposed SCE is employed to localize the robot and to below the threshold ( Al ) (see Fig.229. 2 and 3) as shown in robot pose using scans acquired from a static LIDAR kept in Fig. Number 4. Three scenarios namely (1) demonstrates the localization accuracy of the proposed SCE.The free space robot as shown in Fig.aece. 12. 2016 Figure 10.] Advances in Electrical and Computer Engineering Volume 16. these scenarios are retranslated scans (see Fig. Localization and mapping using proposed SCE technique in DS mode to the LIDAR scan data. The proposed algorithm is executed in a MATLAB® navigate. Experimental Setup-2 SS Mode estimating the robot pose.139. rotation only (2) translation only and (3) both rotation and Acceptable level of deviations has been observed in the translation of robot are studied. A test arena is designed to evaluate the performance of the proposed SCE technique in B. 20 . 11b) and the original map of the exhaustive to capture all possible robot movements. Each region can be characterized by its depth of free a known position which observes the changing pose of the space ( rFS ) over the LIDAR angle index ( n ). It is static and can be found out from the initial environment on the laptop computer which can have access LIDAR scans. [Downloaded from www. 13.181.ro on Thursday. the proposed SCE is used to estimate the of free space segmented as regions (1. The arena consists of three ranges In SS mode. 11c) which indirectly retranslate the local scans. The environment can be generated. the static LIDAR is connected to a laptop computer through the USB is the obstacles free region where the robot can possibly port. Unlike DS mode.

The accuracy of the estimated pose change is evaluated by calculating the MAE between past and present robot scan data retranslated using the estimated pose change. Hence the performance of the proposed SCE degrades in this mode compared with the DS mode.229. the static LIDAR scan data which captures robot pose needs to be identified and isolated from the static environment data. In order to realize this isolation. can reduce the computational time and also suffers loss of information. Test arena for SS mode thresholds (see Fig. December 22. 13. Finally. The MAE (see Fig 15b) A novel KZZ polar correlation technique for SCE was calculated for each LIDAR scans can be used to evaluate the proposed in this paper. Figure  r (n)  rFS (n) rst (n)  0r (n)|| else  (18) 15 illustrates the performance of the proposed SCE VI.ro on Thursday. 14) which methodology is able to maintain the robot on track. Figure 12. 2016 at 09:21:15 (UTC) by 14. Redistribution subject to AECE license or copyright. the proposed SCE technique has been employed to this robot scan data ( rst ) to determine the robot pose change.aece. CONCLUSION technique to track the mobile robot navigating across all the three regions of the test arena.139. the mobile robot is fitted with the identification marker block (cubic block) which in turn helps to identify the robot in the LIDAR scan data.] Advances in Electrical and Computer Engineering Volume 16. It consists of a modified 21 . green shades) derived from free Figure 14. The data associated with the obstacles in the environment will be selectively filtered out (set to zero) from the robot pose data using appropriate Figure 13. Number 4. Tracking performance in SS mode It is observed that the robot scan data is minimal in tracking accuracy and it is observed that the proposed contrast with the environment scan data (see Fig. Isolation of robot scan data and its retranslation in SS mode Figure 15.181. [Downloaded from www. Experimentation set up for SS mode To localize the robot. 2016 space ( rFS ) as in (18).

Nebel. no. 2016 at 09:21:15 (UTC) by 14. Automation. R. and Systems. Z.229. 62-89.181. 6. pp. Quanjun Song.1109/ICMA." IEEE Transactions on Industrial Informatics. Shi.” International Journal of 22 . E. doi: 10. 2014.2013.2013.. International Journal (i.2245694 [9] G. Ivan Petrović. pp. H. [Downloaded from www. M. Petrou. F. D. P. Posner. "Correlation-based scan matching using ultrasonic sensors for EKF localization".1177/027836491456223 localization accuracy. the robot scene change in the scans Automation." Advances in Electrical and Computer variations for localization is the future prospect of this Engineering. 2015. 2013.2005. Sun. pp.976019 track the robot in SS mode. B. Altermatt. 10. pp. pp. “Root mean square error (RMSE) or mean 10. 651-667 2001. 2004. IEEE transaction on Robotics and robot. S. 72.2014. the proposed Journal of Advanced Robotic Systems. [14] R. Weigel. 2010. pp.” Sensors. past scan and retranslated present scan). Zeng Wen.1109/TII. no.1545181 Combining the proposed technique with sensor fusion [20] D. pp. vol. The performance of the proposed SCE is evaluated no.2013. "A fast method for the alignment of the techniques employing extended Kalman filter or its displacement of voxel data.1016/j. D. pp. R.4316/AECE. Yuan. “A vertical and floor line-based monocular SLAM system for corridor environments. 1243-1251. 1999. D. 8. doi: robot in run-time. vol. Shuai Li. doi: 10.1851m per scan was observed 10. 10.] Advances in Electrical and Computer Engineering Volume 16. pp. pp. Jiang Da. T. 2014. Ning. no. pp.”. N. A. Suh. Zalik. M. vol. Neira. Sung-In Choi. robust literature.1.robot. 17. pp. "Critical rays scan match SLAM". method in 3d slam and the effect of sampling strategies". A threshold on of Robotics Research.1080/01691864.690201 [1] T. no. A. doi: and fast feature-based scan matching for an online operation". doi: 10.10. vol. 1495-1519. Gutmann. Bailey. Novak. 28.009 during a combined rotation and translation movement of [16] J.2013. vol. IEEE SMC '99 [3] W. 137-157. 4. 2013. 1001-1008. vol. [19] A. Wang. Mirkhani. L. no. “Change detection and deformation analysis using static and mobile laser scanning. 2014.04. Wang. Ahmad Sharbafi. 2012.139. no. pp. “Polar-Cartesian hybrid pp. Wang. pp. Spelic.1109/IROS. 7. Forsati.e. 5. which uses [15] M. Choi. Y. Lindenbergh.1678144 Danube Region – RAAD2010. vol. [24] Lei Zhang. 8). vol. Ulas. 14. 1439. 108-117. doi: [25] M. 13. pp. "A fast. L." 10. 3. no.3. pp. 2013.2012. 920-931. X.ro on Thursday.5524574 for self-localization in polygonal environments using laser range [23] K.03011 corner features based on a relative map”. and Cybernetics. 860-864.1007/s12555-012-0311-y for rotation estimation and its extension to vertical axis [10] C. Siegwart. Journal 10. 11. Bando.920721 absolute error (MAE)? – arguments against avoiding RMSE in the [6] M. B. Durrant-Whyte. pp.1109/TVT. 19th International Workshop on Robotics in Alpe-Adria- 10. Srećko Jurić-Kavelj. Zhang. "ICE matching. 1053-1058. doi: 10. pp. 2013. doi: 10. 441-462. 2016 conventional cross correlation technique called KZZ-HXC Control. Zeng. IEEE Transactions on Vehicular Technology. doi: 10. Journal technique can be configured by proper selection of its search of Intelligent and Robotic Systems: Theory and Applications. vol. Zhao. 2. W. 34. 2011. an average MAE of 0. J.2014. A. December 22. 61. 10. 1247–1250. IEEE Robotics and Automation pose tracking by correlation of laser range finder scans in Hough Magazine. doi: using fast normalized cross correlation.5194/gmd-7-1247-2014 of Experimental and Theoretical Artificial Intelligence.1080/0952813X. doi: 10. doi: 10. Furthermore. H.1999. no. Tardos. 11.2004. "Object extraction from architecture scenes International Journal of Control. Similarly. M. 2012. 2012. Pietrzyk. an advantage that promotes deployment of application to SLAM". P.12. doi: 10.4197m per [17] Yangming Li.812493 Advances in Electrical and Computer Engineering.1389536 Robotics. Newman. “SLAM with 10. 62. Moayedikia.924576 [7] R. doi: 10. H.5772/56964 technique used non-linearly shifted sum of selective [11] A. Taleghani. no. demonstrated the efficiency of the proposed SCE vol.1080/01691864.1109/MRA. Number 4. Robotics and environmental scene change acquired from a dynamic Autonomous Systems. 2015. 2. R. Hao. doi: LIDAR and a MAE of 0.1007/s12555-012-0172-4 Computer Engineering. J. 1. "A novel efficient algorithm for mobile robot localization". Chai. Ricardo Carelli. "A fast and robust feature-based scan-matching called KZZ-VXC that was used for translation estimation. 547-557. Choi. 9.339. Advanced Robotics. doi: 10.2012.73-78.03002 transforms: a novel 2d range scan registration algorithm”. L.1109/70. space.3390/s110100062 [12] E. ISSN: 1001-506X. Advanced vol.2012. pp. Fig. Temeltas. no. 2. Zhang.2009. 2013. I. vol. Intelligent Robots and Systems.1163/156855301750078720 Conference on Systems.13. and computational speed of the proposed cross correlation doi: 10. 2413-2420. "Fast scan was observed. 2004 IEEE/RSJ [5] S.2271506 methodology in various scenarios. J. 2005 IEEE/RSJ International Conference on the localization algorithm in applications. doi:10.4316/AECE. Ning. 65-74. Journal of Systems Engineering and Electronics. Zhang. "Simultaneous localization and [22] Davor Graovac. 13. 1. "A tracking system of multiple LiDAR sensors using scan point matching".3. 2006. projection in dominant direction of indoor environment". Q. Meng. Advanced Robotics. “Analysis of different difference between two LIDAR scan sequences which feature selection criteria based on a covariance convergence makes it an ideal tool for LIDAR based SCE. 2009 IEEE International Conference on To evaluate the proposed methodology in localizing the Mechatronics and Automation. Automation and Systems. X. Briechle. doi: domain. Esmaeili. vol. no. accurate and robust method 10. 2013. Kleeman.4316/AECE. 2015. vol. "Data association in stochastic mapping using the joint compatibility test”. 12. The accuracy perspective for a slam algorithm. Chung. no. 2010 proposed SCE was quite suitable to localize the robot in run. 7. 890–897. pp. vol.11-16.5245983 The results showed that. Tomatis. T.” Geosci. “Mobile robot mapping (SLAM): Part II". Martinelli. Z. two sets of experiments were performed.2010. in the DS mode. “Analysis of data association experiments showed that the computation time of the methods of SLAM”. 6. Yuta. U.1007/s12518-014-0151-y [8] S.1109/RAAD.aece.1007/s10846-012-9811-5 by computing the MAE between successive LIDAR scans [13] D. pp. R. doi: [2] J. 1122. Hai Liu. Redistribution subject to AECE license or copyright. vol.331. Shahri. no. F. pp. Fernando. 2005. through 3d local scanned data analysis. "Self-localization of a mobile robot finders". Tsardoulias. REFERENCES vol. Conference Proceedings. International Unlike conventional cross correlation. S. [4] X. doi: obtained from a static LIDAR using the proposed SCE to 10. Hanebeck. no. "Scan matching method using International Conference on Intelligent Robots and Systems (IROS). the [18] J. 2012. T. doi: 10. 2001. no. P. Model Dev. “Laser scan matching in polar coordinates with time (see. 3-4. Draxler.02007 investigation. Y." Advances in Electrical and no. 18. no. Soon-Yong Park. 27.720-725. 2009. "Model-free detection and tracking of dynamic objects with 2D LIDAR". doi: [26] T. 3. IROS. [21] M. ICMA 2009. 41-46.” Applied Geomatics. K. These experimental results clearly and robust data association using posterior based approximate joint compatibility test.1109/IROS. I. vol. "ICP based on polar point matching with application to graph-SLAM". 1039-1063. 26. Zhang." IEEE International 10. Man. Guo. pp. Tsubouchi. doi: "Automatic building extraction from terrestrial laser scanning data. 4. vol.1109/ICSMC. 7(2). vol. doi: the MAE value was used as a measure to evaluate the 10. no. vol.2006. 32. vol. Auat Cheein. Diosi.