本文關鍵詞： 普適計算 工藝操作流程 活動建模 活動識別　出處：《大連理工大學》2012年碩士論文　論文類型：學位論文

【摘要】：隨著計算機技術和無線網(wǎng)絡的快速發(fā)展、傳感器與移動設備的廣泛應用,計算模式從傳統(tǒng)的桌面計算模式發(fā)展到普適計算模式,處于這種模式下的用戶,能夠透明地使用普適計算環(huán)境下的各種計算能力和服務資源,所以普適計算環(huán)境應該具備獲取和處理周圍環(huán)境及推斷與用戶活動相關的情境信息的能力。研究表明如果能對用戶的動作行為或者從事的活動進行準確地識別,根據(jù)識別的活動信息可以為用戶提供準確及時的服務信息,有助于用戶規(guī)范地操作和完成活動。因此活動識別已經(jīng)成為近年來普適計算領域中的一個重要課題。 在活動識別的過程中,根據(jù)獲取周圍環(huán)境的變化和用戶的動作行為去推斷用戶的活動是一項有挑戰(zhàn)性的工作。本文研究的活動識別是處于工藝操作流程中的規(guī)范活動,它是由若干復雜的操作組成、為了達到一定目的或者完成一定任務的復雜的有機活動實體,根據(jù)一段時間內(nèi)的情境變化,識別用戶操作的活動信息和用戶的異常動作,為用戶提供相應的提示和服務。 本文結合已有活動識別的相關研究,歸納和總結出活動識別的一般步驟。本文首先提出活動識別的系統(tǒng)框架：情境采集層、情境處理層和情境感知服務層。在情境處理層,包括活動建模和活動識別,其中采用情境推理的方法實現(xiàn)活動識別。為了形式化描述工藝操作流程中活動間相互作用及不同活動的情境信息,在對工藝操作流程結構和活動進行定義后,根據(jù)工藝操作流程中活動間的邏輯關系建立活動流程模型,同時使用情境模式和情境演化模式的概念來描述活動和活動不同的狀態(tài),以便于準確識別用戶的活動和操作活動的異常信息。在活動模型的基礎之上,我們提出活動狀態(tài)可以區(qū)分的規(guī)則,活動模型若滿足活動狀態(tài)可區(qū)分的規(guī)則,根據(jù)活動模型、已知的活動和活動狀態(tài)及其某一時刻情境取值,通過相應的識別算法識別當前情境對應的活動信息；對于不滿足活動狀態(tài)可區(qū)分規(guī)則的活動,根據(jù)活動模型和一段時間內(nèi)情境取值進行匹配,實現(xiàn)活動識別和檢測用戶在操作過程中的異常動作信息,從而可以幫助用戶更好地操作活動。 本文通過對具體應用的工藝操作流程進行活動識別,設計活動識別的原型系統(tǒng),利用計算機實現(xiàn)提出的活動識別算法和原型系統(tǒng),驗證提出的活動識別方法,有效地實現(xiàn)活動識別和監(jiān)測用戶異常操作。
[Abstract]:With the rapid development of computer technology and wireless network and the wide application of sensors and mobile devices, the computing mode has developed from the traditional desktop computing mode to the universal computing mode. Be able to use transparently the various computing capabilities and service resources in a pervasive computing environment, Therefore, the pervasive computing environment should have the ability to acquire and process the surrounding environment and infer the situational information related to the user's activities. According to the identified activity information, it can provide users with accurate and timely service information, and help users to operate and complete activities in a standardized manner. Therefore, activity recognition has become an important topic in the field of pervasive computing in recent years. In the process of activity identification, it is a challenging task to infer the user's activity according to the change of the surrounding environment and the user's action behavior. The activity identification studied in this paper is a normative activity in the process operation process. It is composed of a number of complex operations, in order to achieve a certain purpose or complete a certain task of the complex organic activity entities, according to a period of time, according to the situation changes, identify the user operation of the activity information and abnormal actions of the user. To provide the user with the appropriate tips and services. In this paper, the general steps of activity recognition are summarized and summarized. Firstly, the system framework of activity recognition is put forward: situation collection layer, situational processing layer and situation-aware service layer. It includes activity modeling and activity recognition, in which situational reasoning is used to realize activity recognition. After defining the structure and activity of process operation process, the activity process model is established according to the logical relationship between the activities in the process operation process, and the concept of situational pattern and situational evolution pattern is used to describe the different states of activity and activity. In order to accurately identify the abnormal information of user's activity and operation activity, on the basis of the activity model, we propose the rules that can be distinguished by the activity state. If the activity model satisfies the rule that the activity state can be distinguished, according to the activity model, The known activities and activity states and their situation values at a certain time are identified by the corresponding recognition algorithm. For those activities which do not satisfy the rules of activity state, the corresponding activity information is recognized by the corresponding recognition algorithm. According to the activity model and a period of time to match the situation to achieve activity identification and detection of abnormal action information in the process of operation, which can help users to better operate the activities. In this paper, a prototype system of activity recognition is designed, and the algorithm and prototype system of activity recognition are realized by computer, and the proposed method of activity recognition is verified. It can effectively realize the activity identification and monitor the abnormal operation of the user.
【學位授予單位】：大連理工大學
【學位級別】：碩士
【學位授予年份】：2012
【分類號】：TP338

【參考文獻】

相關期刊論文 前1條

1 徐光yP ,史元春 ,謝偉凱;普適計算[J];計算機學報;2003年09期

相關博士學位論文 前1條

1 謝偉凱;智能空間關鍵支撐技術的研究[D];清華大學;2003年

，

本文編號：1535610