基于增量式雙向主成分分析的機(jī)器人感知學(xué)習(xí)方法研究

王肖鋒; 張明路; 劉軍

doi:10.11999/JEIT170561

基于增量式雙向主成分分析的機(jī)器人感知學(xué)習(xí)方法研究

doi: 10.11999/JEIT170561 cstr: 32379.14.JEIT170561

2.
(河北工業(yè)大學(xué)機(jī)械工程學(xué)院天津 300130)

基金項(xiàng)目:

國(guó)家自然科學(xué)基金(61503119, 61473113)，天津市自然科學(xué)基金(15JCYBJC19800, 16JCZDJC30400)，天津市智能制造科技重大專項(xiàng)(15ZXZNGX00090)

計(jì)量
- 文章訪問(wèn)數(shù): 1360
- HTML全文瀏覽量: 166
- PDF下載量: 226
- 被引次數(shù): 0
出版歷程
- 收稿日期: 2017-06-09
- 修回日期: 2017-10-13
- 刊出日期: 2018-03-19

Robot Perceptual Learning Method Based on Incremental Bidirectional Principal Component Analysis

2.
(School of Mechanical Engineering, Hebei University of Technology, Tianjin 300130, China)

Funds:

The National Natural Science Foundation of China (61503119, 61473113), The Tianjin Natural Science Foundation (15JCYBJC19800, 16JCZDJC30400), The Tianjin Intelligent Manufacturing and Technology Key Project (15ZXZNGX00090)

摘要

摘要: 針對(duì)直觀協(xié)方差無(wú)關(guān)增量式主成分分析算法(CCIPCA)需要滿足零均值高斯分布的問(wèn)題，該文提出含均值差向量更新的泛化CCIPCA算法(GCCIPCA)，拓展了算法的適用范圍。其次，針對(duì)機(jī)器人感知學(xué)習(xí)存在的在線增量計(jì)算及有效數(shù)據(jù)降維等問(wèn)題，將GCCIPCA的增量思想引入到現(xiàn)有的雙向主成分分析算法(BDPCA)，提出基于增量式BDPCA(IBDPCA)的機(jī)器人感知學(xué)習(xí)方法。該方法直接針對(duì)圖像矩陣行列方向的類散度矩陣進(jìn)行迭代估計(jì)，具有一定的泛化能力和快速的增量學(xué)習(xí)能力，提高了實(shí)時(shí)處理速度。最后，以機(jī)器人待抓取物塊作為感知對(duì)象進(jìn)行實(shí)驗(yàn)，結(jié)果表明所提算法能夠滿足機(jī)器人感知學(xué)習(xí)的實(shí)時(shí)處理需求，相比現(xiàn)有的增量式主成分分析算法，在收斂率、分類識(shí)別率、計(jì)算時(shí)間及所需內(nèi)存等性能方面均得到顯著提升。
- 機(jī)器人感知學(xué)習(xí) /
- 增量學(xué)習(xí) /
- 數(shù)據(jù)降維 /
- 直觀協(xié)方差無(wú)關(guān)增量式主成分分析 /
- 雙向主成分分析
Abstract: Existing Candid Covariance-free Incremental PCA (CCIPCA) has the limitation of the stable image inherent covariance, and a Generalized CCIPCA (GCCIPCA) with an appended term of the mean difference vector is presented. It can be considered that the CCIPCA is only a special case of the GCCIPCA and can extend the scope of the algorithm. Then, the incremental learning of the proposed GCCIPCA is innovated to the existing Bi-Directional PCA (BDPCA), and the called Incremental BDPCA (IBDPCA) is used for the robot perceptual learning and it can be used to incrementally compute the principal components without estimating the similar scatter matrixes in the row and column directions, which can build up the real-time processing speed greatly. Finally, the blocks grasped by the robot are used as the perceptual objects, and the experimental results demonstrate that the proposed algorithm works well, and the convergence rate, the classification recognition rate, the computation time and the required memory are improved significantly.
- Robot perceptual learning /
- Incremental learning /
- Dimension reduction /
- Candid Covariance-free Incremental PCA (CCIPCA) /
- Bi-Directional PCA (BDPCA)

HTML全文

參考文獻(xiàn)(25)

GATSOULIS Y and MCGINNITY T M. Intrinsically motivated learning systems based on biologically-inspired novelty detection[J]. Robotics and Autonomous Systems, 2015, 68: 12-20. doi: 10.1016/j.robot.2015.02.006.

WENG J Y, MCCLELLAND J, PENTLAND A, et al. Artificial intelligence-autonomous mental development by robots and animals[J]. Science, 2001, 291(5504): 599-600. doi: 10.1126/science.291.5504.599.

JI ZP and WENG J Y. A developmental wherewhat network for concurrent and interactive visual attention and recognition[J]. Robotics and Autonomous Systems, 2015, 71: 35-48. doi: 10.1016/j.robot.2015.03.004.

SIGAUD O and DRONIOU A. Towards deep developmental learning[J]. IEEE Transactions on Cognitive and Developmental Systems, 2016, 8(2): 99-114. doi: 10.1109/ TAMD.2015.2496248.

YAN H, ANG M H and POO A N. A Survey on perception methods for human-robot interaction in social robots[J]. International Journal of Social Robotics, 2014, 6(1): 85-119. doi: 10.1007/s12369-013-0199-6.

LI Lingjun, LIU Shigang, PENG Yali, et al. Overview of principal component analysis algorithm[J]. Optik- International Journal for Light and Electron Optics, 2016, 127(9): 3935-3944. doi: 10.1016/j.ijleo.2016.01.033.

PARK G and KONNO A. Imitation learning framework based on principal component analysis[J]. Advanced Robotics, 2015, 29(9): 639-656. doi: 10.1080/01691864.2015.1007084.

WENG J Y, ZHANG Y L, and HWANG W S. Candid covariance-free incremental principal component analysis[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003, 25(8): 1034-1040. doi: 10.1109/TPAMI. 2003.1217609.

謝自強(qiáng), 葛為民, 王肖鋒, 等. 發(fā)展型機(jī)器人實(shí)時(shí)特征提取方法研究[J]. 機(jī)器人, 2017, 39(2): 189-196. doi: 10.13973/ j.cnki.robot.2017.0189.

XIE Ziqiang, GE Weimin, WANG Xiaofeng, et al. Real time feature extraction method of developmental robot[J]. Robot, 2017, 39(2): 189-196. doi: 10.13973/j.cnki.robot.2017.0189.

BAI H L and CHEN M. CCIPCA-OPCSC: An online method for detecting shared congestion paths[J]. Computer Networks, 2012, 56(1): 399-411. doi: 10.1016/j.comnet.2011.09.016.

WANG J. Generalized 2-D principal component analysis by Lp-Norm for image analysis[J]. IEEE Transactions on Cybernetics, 2016, 46(3): 792-803. doi: 10.1109/TCYB.2015. 2416274.

曹明明, 干宗良, 崔子冠, 等. 基于2D-PCA特征描述的非負(fù)權(quán)重鄰域嵌入人臉超分辨率重建算法[J]. 電子與信息學(xué)報(bào), 2015, 37(4): 777-783. doi: 10.11999/JEIT140739.

CAO Mingming, GAN Zongliang, CUI Ziguan, et al. Novel neighbor embedding face hallucination based on non-negative weights and 2D-PCA feature[J]. Journal of Electronics Information Technology, 2015, 37(4): 777-783. doi: 10.11999/ JEIT140739.

YANG Wankou, SUN Changyin, and RICANEK K. Sequential row-column 2DPCA for face recognition[J]. Neural Computing and Applications, 2012, 21(7): 1729-1735. doi: 10.1007/s00521-011-0676-5.

XU F, GU G, KONG X, et al. Object tracking based on two- dimensional PCA[J]. Optical Review, 2016, 23(2): 231-243. doi: 10.1007/s10043-015-0178-2.

HUANG J, MA Y, MEI X G, et al. A hybrid spatial-spectral denoising method for infrared hyperspectral images using 2DPCA[J]. Infrared Physics Technology, 2016, 79: 68-73. doi: 10.1016/j.infrared.2016.09.009.

PEI J, HUANG Y, HUO W, et al. SAR imagery feature extraction using 2DPCA-based two-dimensional neighborhood virtual points discriminant embedding[J]. IEEE Journal of Selected Topics in Applied Earth Observations Remote Sensing, 2017, 9(6): 2206-2214. doi: 10.1109/JSTARS.2016.2555938.

ZUO Wangmeng, ZHANG David, and WANG Kuanquan. Bidirectional PCA with assembled matrix distance metric for image recognition[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B(Cybernetics), 2006, 36(4): 863-872. doi: 10.1109/TSMCB.2006.872274.

SUN Yanfeng, CHEN Shangyou, and YIN Baocai. Color face recognition based on quaternion matrix representation[J]. Pattern Recognition Letters, 2011, 32(4): 597-605. doi: 10.1016/j.patrec.2010.11.004.

YANG Wankou, SUN Changyin, ZHANG Lei, et al. Laplacian bidirectional PCA for face recognition[J]. Neurocomputing, 2010, 74(1/3): 487-493. doi: 10.1016/j. neucom.2010.08.020.

NGUYEN T H B and KIM H. Novel and efficient pedestrian detection using bidirectional PCA[J]. Pattern Recognition, 2013, 46(8): 2220-2227. doi: 10.1016/j.patcog.2013.01.007.

REN Chuanxian and DAI Daoqing. Incremental learning of bidirectional principal components for face recognition[J]. Pattern Recognition, 2010, 43(1): 318-330. doi: 10.1016/ j.patcog.2009.05.020.

余映, 王斌, 張立明. 一種面向數(shù)據(jù)學(xué)習(xí)的快速PCA算法[J]. 模式識(shí)別與人工智能, 2009, 22(4): 567-573. doi: 10.16451/ j.cnki.issn1003-6059.2009.04.003.

YU Ying, WANG Bin, and ZHANG Liming. A fast data- oriented algorithm for principal component analysis[J]. Pattern Recognition and Artificial Intelligence, 2009, 22(4): 567-573. doi: 10.16451/j.cnki.issn1003-6059.2009.04.003.

相關(guān)文章

施引文獻(xiàn)

資源附件(0)

訪問(wèn)統(tǒng)計(jì)