0121-連桿工藝及擴(kuò)孔夾具設(shè)計(jì)
0121-連桿工藝及擴(kuò)孔夾具設(shè)計(jì),連桿,工藝,擴(kuò)孔,夾具,設(shè)計(jì)
畢業(yè)設(shè)計(jì)(論文)外文翻譯
譯文題目:創(chuàng)造機(jī)器人:把基本機(jī)制創(chuàng)造成一個(gè)具有自我激勵,自我組織構(gòu)架的機(jī)器人
外文題目: Bringing up robot: F undamental mechanism For creating a self-motiv ated, self-organizing architecture
Bringing up robot: F undamental mechanisms for
creating a self-motiv ated, self-organizing architecture
In this paper we describe an intrinsic developmental algorithm designed to allow a mobile robot to incrementally progress through levels of increasingly sophisticated behavior. We believe that the core ingredients for such a developmental algorithm are abstractions, anticipations, and self-motivations. We propose a multi-level, cascaded discovery and control architecture that includes these core ingredients. Toward this proposal we explore two novel models: a governor for automatically regulating the training of a neural network; and a path-planning neural network driven by patterns of \mental states" which represent protogoals.
1 Introduction
Most intelligent robot control systems begin with the goal of creating a robot to carry out humanissued tasks. While these tasks v ary in difficulty , they must, by their very nature, involve abstract concepts. For example, typical tasks might be: go to a specific location, identify an object, or pick up an object. Attempting to directly achieve the goal of carrying out human commands creates basic assumptions about the architectural design of a robot. We call this philosophy task-oriented design.
Within the task-oriented design paradigm, there are two competing methodologies: top-down, and bottom-up. T op-down designers apply computational algorithms that can be carried out on the robots so as to accomplish a given task. There is a range of computational models employed in robotics: dead reckoning (e.g., using internal measures of space), sensor fusion, behavior fusion, and symbolic logic.
Bottom-up designers again usually take the task to be performed by the robot as a pre-specified assumption. However, the control architecture of the robot is designed in a bottom-up fashion.Examples include subsumption architectures, supervised learning schemes, and evolutionary computation.
We believe that a significant pitfall exists in both the top-down and bottom-up task-oriented robot design methodologies: inherent anthropomorphic bias. This bias refers to the design of prespecified robot tasks: traditional research in the design of intelligent robots has attempted to get robots to do the tasks a human can, and do it in a human-centered manner. Historically , this methodology started out by imitating the physical actions of a child playing with blocks. A task was decomposed into a planning problem, and then, with a robot equipped with an arm and a gripper, the robot was asked to manipulate speciˉc blocks. The inherent anthropomorphic bias existed by design, since the issue was to explore models of intelligent behavior. The pitfall in this approach is that the symbolic modeling of behavior is based on the capabilities of a human body and human concepts. Both capabilities may be inappropriate assumptions for the physical body and experiences of the robot.
Furthermore, even if we could build a robot with a human-like body and senses, it is not clear that we can jump straight to the abstract task at hand. Many control issues need to be solved in order to have a robotic system carry out even the simplest of tasks. After a half-century of continued research, the artificial intelligence and robotics communities are still far from developing any type of general purpose intelligent system.
Recently , a new approach called developmental robotics is being applied to the design of robot behaviors. In this approach, an artifact under the control of an intrinsic developmental algorithm discovers capabilities through autonomous real-time interactions with its environment using its own sensors and effectors. That is, given a physical robot or an artifact, behaviors (as well as mental capabilities) are gr own using a developmental algorithm. The kinds of behaviors and mental capabilities exhibited are not explicitly specified. The focus is mainly on the intrinsic developmental algorithm and the computational models that allow an artifact to grow.
A developmental approach to robotics is partly an attempt to eliminate anthropomorphic bias.By exploring the nature of development, the robot is essentially freed from the task of achieving a pre-specified goal. As long as the intrinsic developmental algorithm demonstrates growing behavior there is no need to pre-specify any particular task for the robot to perform. Indeed, it is the goal of developmental robotics to explore the range of tasks that can be learned (or grown) by a robot, given
a specific developmental algorithm and a control architecture. This paper outlines our approach to a developmental robotics program and two experiments toward an implementation.
2 Overview of a Developmental Robotics Paradigm
The ultimate goal of our developmental robotics program is to design a control architecture that could be installed within a robot so that when that robot is turned on for the first time, it initiates an ongoing, autonomous developmental process. This process should be unsupervised, unscheduled, and task-less, and the architecture should work equally well on any robot platform|a fixed robot arm, a wheeled robot, or a legged robot.
The intrinsic developmental process we are currently exploring contains three essential mechanisms: abstraction, anticipation, and self-motivation. In a realistic, dynamic environment, a robot is flooded with a constant stream of perceptual information. In order to use this information effectively for determining actions, a robot must have the ability to make abstractions so as to focus its attention on the most relevant features of the environment. Based on these abstractions, a robot
must be able to anticipate how the environment will change over time, so as to go beyond simple reflexive behavior to purposeful behavior. Most importantly , the entire process is driven by internal motivations to push the system toward further abstractions and more complex anticipations.
We believe that the developmental process should be employed in a hierarchical, bootstrapping manner, so as to result in the discovery of a range of increasingly sophisticated behaviors. That is, starting with a basic, built-in innate behavior, the robot exercises its sensors and motors, uses the mechanisms for abstraction and anticipation and discovers simple reflex behavior. A self-motivated control scheme employs these discoveries in order to supercede its innate behavior. This constitutes
the first stage of the bootstrapping process.
The same intrinsic developmental algorithm can be employed recursively in subsequent stages, using the knowledge discovered in previous stages. For example, a secondary stage may abstract sequences of behaviors and corresponding perceptual views. These behavior sequences, termed protoplans [12], can lead the robot through a series of views in the environment thus resulting in `interesting' places to visit. We will call these places protogoals. Here, the proto prefix implies a distinction between standard notions of plans and goals from the developmental ones used here.The same developmental process may be cascaded beyond this stage to result in discovery of actual goals and plans.
The control scheme that is responsible for driving the robot at each stage uses the discovered abstractions and anticipations while being pushed by internal motivations. At the lowest level, the motivational model indicates to the system how `comfortable' it is in the given environment. If it is too comfortable, it becomes bored, and takes measures to move the robot into more interesting areas. Conversely , if the environment is chaotic, it becomes over-excited and attempts to return to more stable and well known areas. These anthropomorphic terms will be described below in more technical terms.
創(chuàng)造機(jī)器人:把基本機(jī)制創(chuàng)造成一個(gè)具有自我激勵,自我組織構(gòu)架的機(jī)器人
在本文中,我們描述了一種具有內(nèi)在發(fā)展算法的機(jī)器人,這種機(jī)器人能通過日益復(fù)雜的行為來提高自己。我們認(rèn)為對于這樣一個(gè)發(fā)展的算法它的核心成分是抽象的,隨機(jī)應(yīng)變的和自我動機(jī)的。于是我們提出了一個(gè)多層次的,級聯(lián)的發(fā)現(xiàn),和包括那些核心成分的控制結(jié)構(gòu)。針對這樣一個(gè)提議我們討論了兩種新的模型:一個(gè)帶自動調(diào)節(jié)神經(jīng)網(wǎng)絡(luò)的調(diào)節(jié)器和一個(gè)代表著“protogoals”神經(jīng)網(wǎng)絡(luò)路徑規(guī)劃的模式驅(qū)動。
1介紹
大多數(shù)機(jī)器人的控制系統(tǒng)開始于創(chuàng)建一個(gè)機(jī)器人來執(zhí)行人們發(fā)出的目標(biāo)任務(wù)。而這些任務(wù)都很困難,它們有本身的性質(zhì),是一些抽象的概念。例如,典型的任務(wù)可能是:去一個(gè)特定的位置,識別或者選擇一個(gè)目標(biāo)。試圖直接實(shí)施人們的命令來創(chuàng)建關(guān)于機(jī)器人設(shè)計(jì)的基本假設(shè)。我們稱這種理念為面向任務(wù)設(shè)計(jì)。
在面向任務(wù)設(shè)計(jì)理念里,有兩套互相對立的設(shè)計(jì)方法:自頂向下設(shè)計(jì)和自底向上設(shè)計(jì)。自頂向上的設(shè)計(jì)算法可以讓機(jī)器人完成指定的任務(wù)。在機(jī)器人內(nèi)部運(yùn)用了一系列的計(jì)算模型:航位推算(例如,使用內(nèi)部空間的措施),傳感融合,行為融合,和符號邏輯。
自底而上的設(shè)計(jì)師往往讓機(jī)器人執(zhí)行一個(gè)預(yù)先指定好的假象任務(wù)。然而,對機(jī)器人的控制結(jié)構(gòu)體系則是一個(gè)自底而上的設(shè)計(jì)方案。
我們認(rèn)為在自頂而下和自底而上設(shè)計(jì)方案中都存在著一個(gè)重大的缺陷:內(nèi)在的擬人化的偏見。這種偏見是指對預(yù)先指定的機(jī)器人任務(wù)的設(shè)計(jì):智能機(jī)器人的傳統(tǒng)研究已試圖讓機(jī)器人以人為中心做人們可以做的任務(wù)。從歷史上看,這種方法開始于通過模仿孩子玩積木的肢體動作。目標(biāo)任務(wù)被分解為一個(gè)規(guī)劃方案,然后通過配備一個(gè)臂和夾持器,機(jī)器人被用來處理特定的物體。內(nèi)在的擬人化的偏見之所以能存在設(shè)計(jì)中,是因?yàn)檫@個(gè)問題是探討智能行為的模型。這種方法的缺點(diǎn)是行為符號建?;谌梭w和人的觀念能力之上。這兩個(gè)功能都將成為機(jī)器本身不恰當(dāng)?shù)募傧搿?
此外,即使我們能制造一個(gè)與人類身體和感官一樣的機(jī)器人,它也不會明確我們手上正要執(zhí)行的任務(wù)。為了能有一個(gè)好的機(jī)器人來完成簡單的任務(wù),許多問題都要解決。經(jīng)過一個(gè)半世紀(jì)的不斷研究,人工智能和機(jī)器人領(lǐng)域仍然遠(yuǎn)未發(fā)展任何類型的通用智能系統(tǒng)。
最近,一種新的方法稱為發(fā)展機(jī)器人的方法被應(yīng)用于機(jī)器人的行為設(shè)計(jì)中。在這種方法中,在一個(gè)具有固有發(fā)展能力算法的控制下,通過自主的實(shí)時(shí)互交,利用自身的傳感器和環(huán)境效應(yīng)來發(fā)展自己。就是說,給定一個(gè)物理或人造機(jī)器人,行為(以及精神能力)的生長使用發(fā)展算法。這種行為和心理能力表現(xiàn)沒有明確指定。重點(diǎn)是內(nèi)在發(fā)展算法和計(jì)算模型允許一個(gè)人造機(jī)器人發(fā)展。
機(jī)器人發(fā)展的一種方法是試圖清除擬人化的偏見。通過探索發(fā)展的本質(zhì),機(jī)器人基本上可以自由地完成一個(gè)預(yù)先指定的目標(biāo)任務(wù)。只要內(nèi)在發(fā)展的算法在成長就無需對機(jī)器人執(zhí)行任何特定的任務(wù)。事實(shí)上,是機(jī)器人目標(biāo)探索任務(wù)可以學(xué)到的范圍,給出了具體的算法和控制結(jié)構(gòu)的發(fā)展。本文概述了機(jī)器人發(fā)展的方法和兩個(gè)實(shí)驗(yàn)的實(shí)現(xiàn)。
2發(fā)展式機(jī)器人的描述
我們發(fā)展機(jī)器人的最終目標(biāo)是設(shè)計(jì)一個(gè)可安裝在機(jī)器人上的控制構(gòu)架以至于一打開機(jī)器人時(shí)就會啟動一個(gè)持續(xù)發(fā)展的過程。這個(gè)過程應(yīng)該是無監(jiān)督的,不定期的,任務(wù)少的,而且這種構(gòu)造同樣使用于任何一個(gè)固定的機(jī)器人手臂的機(jī)器人平臺,輪式機(jī)器人或足式機(jī)器人。
在內(nèi)在發(fā)展過程中,我們目前正在探索三個(gè)基本機(jī)制:抽象的,預(yù)期的,和自我激勵的。在一個(gè)現(xiàn)實(shí)的動態(tài)環(huán)境中,一個(gè)機(jī)器人充斥著一個(gè)恒定的感知信息流。為了有效地使用此信息來確定動作,機(jī)器人必須要有抽象的已關(guān)注環(huán)境的最相適應(yīng)的能力。在此基礎(chǔ)上,機(jī)器人必須能夠預(yù)測到環(huán)境隨著時(shí)間變化而變化從而超越簡單的反身行為或有目的的行為。最重要的是,整個(gè)過程是由內(nèi)在動機(jī)驅(qū)動的推進(jìn)系統(tǒng)進(jìn)一步的抽象和更復(fù)雜的預(yù)測。
我們認(rèn)為在發(fā)展過程中應(yīng)采用分層引導(dǎo)的方式,從而導(dǎo)致一系列的日益復(fù)雜的行為出現(xiàn)。那就是起始于一個(gè)基本的,內(nèi)在的先天行為,機(jī)器人的運(yùn)動傳感器和電機(jī)采用抽象的預(yù)測的和發(fā)現(xiàn)簡單的反身行為。該控制方案采用了這些方法目的是取代其先天行為。這是自發(fā)過程的第一階段。
相同的內(nèi)在發(fā)展算法可以被使用在后期階段利用所知道的發(fā)現(xiàn)以前的階段。例如,第二階段可能要記錄機(jī)器人的行為和相應(yīng)的感官。這些行為序列稱為protoplans,可以是機(jī)器人通過一系列有趣的地方從而形成對環(huán)境的感知,我們稱這些地方為protogoals。在這里,原始前綴意味著從這里使用的發(fā)展計(jì)劃與目標(biāo)的標(biāo)準(zhǔn)概念之間的區(qū)別。同樣的發(fā)展過程可以被超越這個(gè)階段發(fā)現(xiàn)實(shí)際目標(biāo)和計(jì)劃。
這是負(fù)責(zé)在每個(gè)階段驅(qū)動機(jī)器人的控制方案,利用發(fā)現(xiàn)抽象的和預(yù)期的目標(biāo)而推測到內(nèi)部動機(jī)。在最低水平模型表示系統(tǒng)是在給定的環(huán)境中如何適應(yīng)。如果它太舒適了,會變的無聊,并采取措施使機(jī)器人到達(dá)更有趣的地方。相反,如果環(huán)境是混沌的,它將變得過度興奮并試圖返回到更穩(wěn)定和熟悉的地方,這些擬人化的條款將在下面描述的更多。
7
收藏