具有自主發(fā)育能力的機(jī)器人感知與認(rèn)知方法研究
發(fā)布時(shí)間：2019-05-24 16:30
【摘要】：集裝箱裝卸自動(dòng)化是運(yùn)輸集裝箱化的必然要求,在當(dāng)前集裝箱裝卸作業(yè)中,扭鎖的裝卸仍然由人工來完成,這不僅增加了勞動(dòng)強(qiáng)度,降低了生產(chǎn)效率,還嚴(yán)重威脅到工人的人身安全,亟需以機(jī)器人為核心的自動(dòng)化技術(shù)來取代人工操作。本文以海港集裝箱扭鎖的自動(dòng)化安裝為研究背景,根據(jù)扭鎖安裝需求搭建模擬平臺(tái),主要解決扭鎖的認(rèn)知識(shí)別與抓取位姿估計(jì)問題。由于扭鎖種類繁多,且隨著需求不斷改進(jìn)更新,抓取任務(wù)不斷有新的挑戰(zhàn),機(jī)器人的認(rèn)知系統(tǒng)需要在線實(shí)時(shí)地更新、存儲(chǔ)新的特征,否則無法準(zhǔn)確識(shí)別新類別物體。而傳統(tǒng)機(jī)器人的認(rèn)知系統(tǒng)存在任務(wù)確定、離線學(xué)習(xí)、實(shí)時(shí)性差及自適應(yīng)性差等問題,無法完成非特定任務(wù)。為了解決從工作場(chǎng)景中識(shí)別并準(zhǔn)確抓取指定物體的問題,針對(duì)傳統(tǒng)機(jī)器人認(rèn)知系統(tǒng)存在的局限性,從認(rèn)知機(jī)器人的研究思路出發(fā),模擬人類學(xué)習(xí)方式、智能表現(xiàn)形式以及人腦智能信息處理機(jī)制,建立本文的機(jī)器人認(rèn)知系統(tǒng),使機(jī)器人通過在線學(xué)習(xí),將累積的知識(shí)和經(jīng)驗(yàn)動(dòng)態(tài)有組織地存儲(chǔ)到記憶系統(tǒng)中,在執(zhí)行任務(wù)時(shí)回調(diào)以往的經(jīng)驗(yàn)知識(shí)做出準(zhǔn)確的識(shí)別,進(jìn)而獲取準(zhǔn)確的位姿估計(jì)�；谧灾靼l(fā)育范式將扭鎖抓取機(jī)器人的認(rèn)知系統(tǒng)分為:感知發(fā)育、認(rèn)知發(fā)育以及任務(wù)執(zhí)行三大模塊,從三個(gè)功能模塊展開,本文主要的研究工作如下:(1)傳感器數(shù)據(jù)預(yù)處理,本文提出了基于分域策略的聯(lián)合雙邊濾波預(yù)處理方法,解決了Kinect傳感器采集的深度圖像存在漏洞、不對(duì)齊以及噪聲等問題。根據(jù)Kinect傳感器三種誤差來源的區(qū)域特性,對(duì)深度圖像進(jìn)行分區(qū)域?yàn)V波處理。根據(jù)深度圖像和彩色圖像的結(jié)構(gòu)相關(guān)性,對(duì)深度像素進(jìn)行分類,將漏點(diǎn)及不對(duì)齊像素歸類為不可信任區(qū)域,其余像素歸為可信任區(qū)域。融合彩色圖像信息,采用聯(lián)合濾波方法對(duì)深度圖像進(jìn)行引導(dǎo)濾波,針對(duì)可信任區(qū)域像素采用聯(lián)合三邊濾波方法;針對(duì)不可信任區(qū)域中邊緣像素采用Sigmoid-方向高斯的聯(lián)合雙邊濾波方法,非邊緣像素采用Sigmoid-顏色相似的聯(lián)合雙邊濾波方法。其中,基于增強(qiáng)學(xué)習(xí)中的獎(jiǎng)懲原則,使用Sigmoid函數(shù)為不可信任區(qū)域像素動(dòng)態(tài)產(chǎn)生置信度空域權(quán)重,賦予濾波鄰域內(nèi)與中心點(diǎn)屬性相同的可信任信息較高權(quán)重;使用方向高斯濾波函數(shù)為邊緣像素產(chǎn)生顏色權(quán)重,賦予濾波鄰域內(nèi)與邊緣方向一致的像素較高權(quán)重,保留邊界方向性;基于可信度勢(shì)場(chǎng)理念選取濾波方向,確保濾波鄰域內(nèi)含有更多有效的與待濾波點(diǎn)屬性相同的可信任信息,通過以上策略手段來保證濾波后深度信息的合理性和準(zhǔn)確性。最后通過對(duì)比實(shí)驗(yàn)從視覺度量、降噪性能及運(yùn)行時(shí)間上,有力地證明了本文濾波方法的優(yōu)越性能。(2)本文提出了在線自適應(yīng)增量PCA學(xué)習(xí)方法,解決了感知發(fā)育中特征提取與數(shù)據(jù)降維問題。該方法能夠在線自主地發(fā)現(xiàn)和選擇輸入數(shù)據(jù)的有效特征,更新優(yōu)化特征空間,發(fā)育出適合機(jī)器人內(nèi)部表達(dá)的模型。針對(duì)PCA學(xué)習(xí)方法對(duì)樣本數(shù)量及多樣性依賴程度高、缺乏自適應(yīng)性、不能在線增量更新、可擴(kuò)展性差等問題;增量PCA方法隨著樣本輸入,特征維度、計(jì)算量和存儲(chǔ)量都隨之增加等問題。本文算法在增量PCA的基礎(chǔ)上進(jìn)行改進(jìn),基于新樣本與已有特征空間重建樣本之間的差異程度監(jiān)測(cè)新類別輸入,控制特征空間增量地更新;基于類內(nèi)距離比較,自適應(yīng)地更新類內(nèi)距離閾值,優(yōu)化特征空間向量。實(shí)驗(yàn)表明該算法在少量訓(xùn)練樣本的情況下,能夠在線地學(xué)習(xí)、更新與優(yōu)化、累積新特征,將高維輸入信號(hào)合理降維,增強(qiáng)了視覺系統(tǒng)的感知和識(shí)別能力。(3)本文借鑒人腦記憶系統(tǒng)中前額葉、海馬以及海馬前額葉回路的信息處理機(jī)制,提出了三層的基于增量式神經(jīng)網(wǎng)絡(luò)的認(rèn)知發(fā)育模型,能夠在線對(duì)所學(xué)的知識(shí)和經(jīng)驗(yàn)實(shí)時(shí)有效地存儲(chǔ)、累積、整合以及回調(diào),解決傳統(tǒng)數(shù)據(jù)庫存儲(chǔ)知識(shí)的固定性、封閉性等問題,更好地適應(yīng)未知的動(dòng)態(tài)環(huán)境。認(rèn)知發(fā)育網(wǎng)絡(luò)中有監(jiān)督學(xué)習(xí)和無監(jiān)督學(xué)習(xí)方式可同時(shí)并存,隨著與外界不斷的交互,中間層神經(jīng)元同時(shí)接受外界通過效應(yīng)層傳遞的自上而下的監(jiān)督指導(dǎo)信號(hào)和來自輸入自底向上的響應(yīng)信號(hào),使用Hebbian學(xué)習(xí)規(guī)則來模擬神經(jīng)元學(xué)習(xí)響應(yīng)過程,采用Top-K競(jìng)爭(zhēng)機(jī)制模擬神經(jīng)元的側(cè)抑制效應(yīng),引入遺忘平均函數(shù)產(chǎn)生權(quán)重模擬人類接受新知識(shí)的速度,通過以上策略模擬大腦皮層理解、記憶情況。認(rèn)知發(fā)育神經(jīng)網(wǎng)絡(luò)在第四章感知發(fā)育模塊基礎(chǔ)上,基于重建誤差控制神經(jīng)網(wǎng)絡(luò)節(jié)點(diǎn)的增加,基于熟悉相似度控制被激活神經(jīng)元的權(quán)重更新。通過實(shí)驗(yàn)表明,認(rèn)識(shí)發(fā)育網(wǎng)絡(luò)可以將學(xué)習(xí)的結(jié)果以“知識(shí)”的形式有組織地、動(dòng)態(tài)地存儲(chǔ)到記憶系統(tǒng)中,取代傳統(tǒng)數(shù)據(jù)庫,提高了扭鎖的準(zhǔn)確識(shí)別率。(4)扭鎖抓取位姿估計(jì),本文根據(jù)扭鎖安裝需求搭建抓取平臺(tái),經(jīng)認(rèn)知分析后獲取扭鎖正確類別及其正反面信息,與相應(yīng)類型的標(biāo)準(zhǔn)位姿做比對(duì),將位姿估計(jì)問題簡(jiǎn)化為兩個(gè)點(diǎn)云集匹配問題,采用迭代最近點(diǎn)(ICP)算法估算可抓取點(diǎn)的位置和姿態(tài),為下一步抓取規(guī)劃提供數(shù)據(jù)支持。通過實(shí)驗(yàn),證明了該方法的可行性。最后,總結(jié)全文所做的工作,提出今后進(jìn)一步需要研究的問題。
[Abstract]:The container loading and unloading automation is an inevitable requirement for the transportation of the container. In the present container loading and unloading operation, the loading and unloading of the twist lock is still carried out manually, which not only increases the labor intensity, reduces the production efficiency, but also seriously threatens the personal safety of the workers, Robotic-based automation technology is needed to replace manual operations. The paper takes the automatic installation of the twist lock of the harbor container as the research background, and sets up the simulation platform according to the installation requirements of the twist lock, and mainly solves the problem of the cognition recognition and the grasping pose estimation of the twist lock. Due to the wide variety of twist locks, and with the continuous improvement of the demand, the grasping task has new challenges, and the robot's cognitive system needs to be updated online in real time, and new features can be stored, otherwise, the new category object cannot be accurately identified. The cognitive system of the traditional robot has the problems of task determination, off-line learning, poor real-time performance and poor self-adaptability. In order to solve the problem of identifying and accurately capturing the specified object from the work scene, aiming at the limitation of the traditional robot cognitive system, the human learning method, the intelligent expression form and the human brain intelligent information processing mechanism are simulated from the research thinking of the cognitive robot, In this paper, the robot cognitive system is established, which enables the robot to dynamically organize the accumulated knowledge and experience into the memory system through on-line learning, and to make an accurate identification of the past experience knowledge when executing the task, so as to obtain the accurate pose estimation. The cognitive system of the twist-lock grasping robot is divided into three modules: the sense development, the cognitive development and the task execution based on the independent development paradigm, and the main research work in this paper is as follows: (1) the sensor data is pre-processed, In this paper, a combined double-side filtering pre-processing method based on the split-domain strategy is proposed, and the problems such as the vulnerability, the misalignment and the noise of the depth image acquired by the Kinect sensor are solved. According to the region characteristics of three error sources of the Kinect sensor, the depth image is divided into region filtering processing. According to the structure correlation of the depth image and the color image, the depth pixel is classified, and the missing point and the non-aligned pixel are classified as the non-trusted area, and the remaining pixels are classified as a trusted area. the method comprises the following steps of: fusing the color image information, carrying out direct filtering on the depth image by using a joint filtering method, adopting a combined trilateral filtering method for the trusted area pixels, and adopting a joint bilateral filtering method of the sigmoid-direction gauss aiming at the edge pixels in the non-trusted area, The non-edge pixels adopt a joint double-sided filtering method similar to the Simoid-color. The method comprises the following steps of: dynamically generating a confidence spatial weight for a non-trusted area pixel by using a Simoid function based on the reward and punishment principle in the enhanced learning, and giving a high weight of the trusted information which is the same as the center point attribute in the filter neighborhood; and generating a color weight for the edge pixel by using the directional Gaussian filter function, and the filtering direction is selected based on the concept of the reliability potential field to ensure that more effective trust information is contained in the filter neighborhood which is the same as that of the point to be filtered, And the rationality and the accuracy of the depth information after filtering are ensured through the above strategy means. Finally, the superiority of the filtering method in this paper is proved by the contrast experiment from the visual measurement, the noise reduction performance and the running time. (2) In this paper, an on-line self-adaptive incremental PCA learning method is proposed to solve the problem of feature extraction and data reduction in sensing development. The method can automatically discover and select the effective characteristics of the input data, update the optimized feature space, and develop a model suitable for the internal expression of the robot. The method of PCA learning has the problems of high sample number and diversity, lack of self-adaptability, no on-line incremental updating, poor scalability, etc. The increment PCA method increases with the sample input, the feature dimension, the calculation quantity and the storage amount. the algorithm is improved on the basis of the increment PCA, the new category input is monitored based on the difference between the new sample and the existing feature space reconstruction sample, the control feature space is updated incrementally, the intra-class distance threshold is adaptively updated based on the intra-class distance comparison, The feature space vector is optimized. The experiment shows that the algorithm can study, update and optimize on-line, accumulate new features in a small amount of training samples, reduce the dimension of the high-dimension input signal, and enhance the perception and recognition ability of the vision system. (3) Based on the information processing mechanism of the prefrontal lobe, the hippocampus and the frontal lobe of the hippocampus of the brain memory system, a three-layer cognitive development model based on the incremental neural network is proposed, which can effectively store and accumulate the learned knowledge and experience in real time. And the problem that the traditional database storage knowledge is fixed, closed and the like is solved, and the unknown dynamic environment is better adapted. in that cognitive development network, the supervised learning and the non-supervised learning method can coexist at the same time, and as the interaction with the external environment, the middle-layer neuron receives the top-down supervision guidance signal transmitted by the outside through the effect layer and the response signal from the input self-bottom, Using the Hebbian learning rule to simulate the learning response of the neuron, the side effect of the neuron was simulated by the Top-K competition mechanism, and the forgetting average function was introduced to generate the weight to simulate the speed of the human being's new knowledge. The above strategy was used to simulate the understanding and memory of the cerebral cortex. The cognitive development neural network, based on the fourth-sense development module, controls the increase of the neural network node based on the reconstruction error, and controls the weight update of the activated neuron based on the familiar similarity. The experiment shows that the cognitive development network can be organized and dynamically stored in the memory system in the form of "knowledge", instead of the traditional database, the accurate recognition rate of the twist lock is improved. (4) the position and position estimation of the twist lock is constructed, a grasping platform is built according to the installation requirement of the twist lock, the correct category of the twist lock and the positive and negative information of the twist lock are acquired through the cognitive analysis, and the pose estimation problem is simplified into two point cloud matching problems, An iterative recent point (ICP) algorithm is used to estimate the position and attitude of the grab points and provide data support for next-step grab planning. The feasibility of this method is proved by the experiment. Finally, the paper sums up the work done in the whole text, and puts forward some problems that need to be studied in the future.
【學(xué)位授予單位】：吉林大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP391.41;TP242
本文編號(hào)：2485013
[Abstract]:The container loading and unloading automation is an inevitable requirement for the transportation of the container. In the present container loading and unloading operation, the loading and unloading of the twist lock is still carried out manually, which not only increases the labor intensity, reduces the production efficiency, but also seriously threatens the personal safety of the workers, Robotic-based automation technology is needed to replace manual operations. The paper takes the automatic installation of the twist lock of the harbor container as the research background, and sets up the simulation platform according to the installation requirements of the twist lock, and mainly solves the problem of the cognition recognition and the grasping pose estimation of the twist lock. Due to the wide variety of twist locks, and with the continuous improvement of the demand, the grasping task has new challenges, and the robot's cognitive system needs to be updated online in real time, and new features can be stored, otherwise, the new category object cannot be accurately identified. The cognitive system of the traditional robot has the problems of task determination, off-line learning, poor real-time performance and poor self-adaptability. In order to solve the problem of identifying and accurately capturing the specified object from the work scene, aiming at the limitation of the traditional robot cognitive system, the human learning method, the intelligent expression form and the human brain intelligent information processing mechanism are simulated from the research thinking of the cognitive robot, In this paper, the robot cognitive system is established, which enables the robot to dynamically organize the accumulated knowledge and experience into the memory system through on-line learning, and to make an accurate identification of the past experience knowledge when executing the task, so as to obtain the accurate pose estimation. The cognitive system of the twist-lock grasping robot is divided into three modules: the sense development, the cognitive development and the task execution based on the independent development paradigm, and the main research work in this paper is as follows: (1) the sensor data is pre-processed, In this paper, a combined double-side filtering pre-processing method based on the split-domain strategy is proposed, and the problems such as the vulnerability, the misalignment and the noise of the depth image acquired by the Kinect sensor are solved. According to the region characteristics of three error sources of the Kinect sensor, the depth image is divided into region filtering processing. According to the structure correlation of the depth image and the color image, the depth pixel is classified, and the missing point and the non-aligned pixel are classified as the non-trusted area, and the remaining pixels are classified as a trusted area. the method comprises the following steps of: fusing the color image information, carrying out direct filtering on the depth image by using a joint filtering method, adopting a combined trilateral filtering method for the trusted area pixels, and adopting a joint bilateral filtering method of the sigmoid-direction gauss aiming at the edge pixels in the non-trusted area, The non-edge pixels adopt a joint double-sided filtering method similar to the Simoid-color. The method comprises the following steps of: dynamically generating a confidence spatial weight for a non-trusted area pixel by using a Simoid function based on the reward and punishment principle in the enhanced learning, and giving a high weight of the trusted information which is the same as the center point attribute in the filter neighborhood; and generating a color weight for the edge pixel by using the directional Gaussian filter function, and the filtering direction is selected based on the concept of the reliability potential field to ensure that more effective trust information is contained in the filter neighborhood which is the same as that of the point to be filtered, And the rationality and the accuracy of the depth information after filtering are ensured through the above strategy means. Finally, the superiority of the filtering method in this paper is proved by the contrast experiment from the visual measurement, the noise reduction performance and the running time. (2) In this paper, an on-line self-adaptive incremental PCA learning method is proposed to solve the problem of feature extraction and data reduction in sensing development. The method can automatically discover and select the effective characteristics of the input data, update the optimized feature space, and develop a model suitable for the internal expression of the robot. The method of PCA learning has the problems of high sample number and diversity, lack of self-adaptability, no on-line incremental updating, poor scalability, etc. The increment PCA method increases with the sample input, the feature dimension, the calculation quantity and the storage amount. the algorithm is improved on the basis of the increment PCA, the new category input is monitored based on the difference between the new sample and the existing feature space reconstruction sample, the control feature space is updated incrementally, the intra-class distance threshold is adaptively updated based on the intra-class distance comparison, The feature space vector is optimized. The experiment shows that the algorithm can study, update and optimize on-line, accumulate new features in a small amount of training samples, reduce the dimension of the high-dimension input signal, and enhance the perception and recognition ability of the vision system. (3) Based on the information processing mechanism of the prefrontal lobe, the hippocampus and the frontal lobe of the hippocampus of the brain memory system, a three-layer cognitive development model based on the incremental neural network is proposed, which can effectively store and accumulate the learned knowledge and experience in real time. And the problem that the traditional database storage knowledge is fixed, closed and the like is solved, and the unknown dynamic environment is better adapted. in that cognitive development network, the supervised learning and the non-supervised learning method can coexist at the same time, and as the interaction with the external environment, the middle-layer neuron receives the top-down supervision guidance signal transmitted by the outside through the effect layer and the response signal from the input self-bottom, Using the Hebbian learning rule to simulate the learning response of the neuron, the side effect of the neuron was simulated by the Top-K competition mechanism, and the forgetting average function was introduced to generate the weight to simulate the speed of the human being's new knowledge. The above strategy was used to simulate the understanding and memory of the cerebral cortex. The cognitive development neural network, based on the fourth-sense development module, controls the increase of the neural network node based on the reconstruction error, and controls the weight update of the activated neuron based on the familiar similarity. The experiment shows that the cognitive development network can be organized and dynamically stored in the memory system in the form of "knowledge", instead of the traditional database, the accurate recognition rate of the twist lock is improved. (4) the position and position estimation of the twist lock is constructed, a grasping platform is built according to the installation requirement of the twist lock, the correct category of the twist lock and the positive and negative information of the twist lock are acquired through the cognitive analysis, and the pose estimation problem is simplified into two point cloud matching problems, An iterative recent point (ICP) algorithm is used to estimate the position and attitude of the grab points and provide data support for next-step grab planning. The feasibility of this method is proved by the experiment. Finally, the paper sums up the work done in the whole text, and puts forward some problems that need to be studied in the future.
【學(xué)位授予單位】：吉林大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP391.41;TP242
【參考文獻(xiàn)】
相關(guān)期刊論文 前10條
1 任紅格;向迎帆;李福進(jìn);;基于內(nèi)在動(dòng)機(jī)的智能機(jī)器人自主發(fā)育算法[J];計(jì)算機(jī)應(yīng)用;2015年09期
2 胡啟祥;瞿心昱;;內(nèi)部動(dòng)機(jī)驅(qū)動(dòng)的機(jī)器人未知環(huán)境在線自主學(xué)習(xí)[J];計(jì)算機(jī)工程與應(yīng)用;2014年04期
3 黃誠(chéng);沈昱明;劉華平;李格格;;基于增量PCA的目標(biāo)跟蹤算法[J];江南大學(xué)學(xué)報(bào)(自然科學(xué)版);2013年06期
4 陳曉明;蔣樂天;應(yīng)忍冬;;基于Kinect深度信息的實(shí)時(shí)三維重建和濾波算法研究[J];計(jì)算機(jī)應(yīng)用研究;2013年04期
5 陳洋;張道輝;趙新剛;韓建達(dá);;基于IHDR自主學(xué)習(xí)框架的無人機(jī)3維路徑規(guī)劃[J];機(jī)器人;2012年05期
6 瞿心昱;姚明海;顧勤龍;;視覺陌生度驅(qū)動(dòng)的增量自主式視覺學(xué)習(xí)算法[J];中國(guó)圖象圖形學(xué)報(bào);2012年06期
7 王作為;張汝波;;自主發(fā)育智能機(jī)器人體系結(jié)構(gòu)研究[J];計(jì)算機(jī)應(yīng)用與軟件;2011年11期
8 姚明海;瞿心昱;;基于自適應(yīng)子空間在線PCA的手勢(shì)識(shí)別[J];模式識(shí)別與人工智能;2011年02期
9 王功伍;蔡景霞;;海馬-前額葉神經(jīng)回路與工作記憶[J];動(dòng)物學(xué)研究;2010年01期
10 紀(jì)昂;姚丹;郭躍飛;;一種基于葉分量分析的帶有監(jiān)督信息的在線學(xué)習(xí)方法[J];計(jì)算機(jī)應(yīng)用與軟件;2009年08期
，本文編號(hào)：2485013
本文鏈接：http://www.sikaile.net/shoufeilunwen/xxkjbs/2485013.html
最近更新
教材專著
