于仁颇黎@机器人

看到一段来自异乡的车祸视频，觉得非常有警示意义.开车是为了到达，不要因为什么借口，比如领导让你上班快一点；也不要让路上美女的吸引让你分心；不要显示自己的车技如何好，再好你就去做试车手，或者找个专门的车道跑一下；也不要随便在大路上的肆无忌惮的秀小两口恩爱；再有多少急的事,你也不要超速，安全第一，工作没有了可以再找，生命没有了，工作还有什么意义。

于仁颇黎注：本文是前一篇（自动化工程师们最明智的选择）英文原文，供大家参考,如发现一些专业词汇的翻译问题，请指正。

Helping Designers Make the Right Choices for Automation

By: David Mollert

I recentlylistened to a talk radio program about manufacturing. As the conversationwent on, one caller’s statement stuck in my mind: ‘‘While automation has playeda large part in increasing productivity, we have not gotten the help fromrobotics that we had hoped for.’‘ That took me by surprise, since I’veworked as a designer of all-electric, articulated robot systems for over tenyears, and know that robots have played a major role in improving manufacturingefficiency. I just assumed they were talking about some other typeof hard automation besides robots. Then I got to thinking that there arestill designers so comfortable with hard automation that they have not yetconsidered articulated robots.

Robots havematured from their birth in specific industries with specific tasks to becomingversatile mechanisms that are ideal for straightforward pick-and-placeapplications, as well as challenging applications that can utilize the uniquecapabilities inherently built into robotics. After working with automationequipment for 20 years, I feel it’s important to provide insights into mytransition from hard tooling to robotics so that others can understand thesignificant differences between the two. The purpose of this article isto touch on several features that have made today’s robot a vital tool for anyapplication.

TrueFlexibility:

The termflexibility means a variety of things when discussing robots. Let mefirst discuss flexibility in movement. With six-axis robots available,movement is virtually unrestricted. The designer spends less time on howthe parts are moved and more time on the tooling at the end of the robot thatpicks the parts. This flexibility allows the tooling to be designed withan eye toward multiple tasks. For example, picking boxes and pallets or assemblingtwo different parts and then setting them on an exit conveyor. The ideais having the robot do most of the work. In situations when theend-of-arm-tool cannot accommodate all of the different shapes or sizes of theparts, tool changers are added to allow the robot to pneumatically changeend-of-arm tools. This type of flexibility in movement is very usefulduring the building of a robotic cell. Hard tooling does not lend itselfto minor positional changes as well as robots do. These changes made ‘‘onthe floor’‘ often help with the overall productivity of the cell.

Flexibility inmounting. Floor, ceiling or rail-mount robots offer the designer anoption with most applications that does not require additional mountingstructures. This speeds up the engineering needed to develop mounts, aswell as the outside fabrication requirements.

Flexibility inyour long-term investment. Traditionally, the thought of reusing hardtooling would be unheard of, but robots can be re-deployed to accommodatechanges in products or procedures. When reusing robots, only the toolingand programming need modifications. They eliminate the question ofcompatibility when attempting to blend a variety of hard tooling products fromdifferent component manufactures together in one assembly. Because robotsoffer multiple axes and are self-contained, there is no need for a structuralframework to mount the various components of hard tooling. They alsogreatly reduce the time needed for hard wiring of the system. For mostapplications, power is only required for the robot and air if needed for the end-of-arm-tool. Another advantage of re-deploying robots to new applications is that it breedscontinuity throughout the plant. When reusing robots there is no learningcurve or additional spare part requirements, and only one point of contact forits electrical and mechanical components.

When I firststarted designing with robots, I had a tendency to limit their flexibility bythinking of only a single task, similar to hard tooling. I now look atthe overall system and incorporate the robot to do as many tasks aspossible. The key point is that robot flexibility allows the designermore options without having to deal with the compromises of hard tooling.

Programming:

Along withadvances in the drives and the mechanical unit, a robot’s programming languageis straightforward if you are accustomed to reading ladder logic. Eachline represents a separate robot command. The command lines that move therobot have four components. These components tell the robot were to go,how fast, how to get there, and whether to use all of the axes in unison orindividually. The development of these programs start with the hand held‘‘teach pendant’‘ which is used to physically drive the robot to a desiredpoint where the four variables can be selected and the point recorded. Itis a point-by-point process after that. These programs can become ascomplicated as the process demands, but even then the basic structure of thelanguage stays the same. This type of straightforward programming goes along way in removing the stigma of complicated controls and allows for a shortlearning curve for any individual.

In addition,FANUC Robotics offers a simulation program to set up a virtual cell on acomputer. Once the robot, tooling and other peripheral equipment areselected, the user can construct the program off-line. The softwareprovides the ability to create and watch the process and adjust locations andspeeds in order to refine the system’s cycle time. This program can thenbe loaded into a robot on the floor, and after verifying the positional points,it’s ready to run.

Limitations:

Robots havedefinitely made a positive impact on manufacturing, but there are a few keypoints to remember when designing with robots. The first point is thesize of the control cabinet. With a footprint of 24 by 30 inches, itconsumes more floor space than many smaller robots. Because of its size,designers must consider the controller during initial discussions of the systemor cell space requirements.

The second pointhas to do with safety considerations. Because the available travel of asix-axis robot resembles a sphere, when working with a specific application itis advisable to limit the travel to only where the robot needs to go. These limits must be accomplished with physical stops in order to adhere to theRobotic Industries Association’s safety requirements. Software limitscannot replace the physical stops. Once the maximum travel has beenestablished, guarding needs to be erected to prevent access by personnel. Includingphysical stops in the design helps to minimize the amount of floor space therobotic system consumes.

The last item ismore of a caution when designing robot-mounting bases. With the highspeeds of each axis, it is easy to underestimate the rigidity required of thebase, even with smaller robots. An adequately sized base insures that therobot will be on solid ground and not quiver when stopping, and can help withthe accuracy of the process.

作者: Ben Nagler 翻译:于仁颇黎

为了经济保持长足的增长。每一个厂商必须不停的提高产品的质量。并同时降低产品的价格。提高产品的性价比，客户要的是高可靠性。高质量的产品，而要达到这个目的的唯一方法就是提高生产的生动化程度，

基于机器人技术的自动化系统是一种柔性的自动化系统。而不是以前所谓的”固定的”或者”硬性的”的 自动化系统，硬性的自动化系统只有当生产流程永远是固定时才能起到它的作用，但事实上。那种在很长时期内生产同一样产品的时代已经不存在了，当前产品的生 命周期是很短的。一年已经很长了，即使在一年的生命周期中。生产线也可能要经过不少的微小的修正。从而使产品更好的能在市场上占有一定的份额。因此，厂家 需要一种可以及时快速的通过编程来实现客户加工要求的设备，基于机器人技术的自动化系统就能很好的达到这个目的，并且可以在一条自动化生产线上生产多种产 品。

与其他的投资目的相同，投资于工业机器人不仅能解决短期的问题，同时对提高产品的性价比，提高产品的市场竞争力，也 有很大的帮助。（你可以问问你自己：工业机器人在你的车间里可以做什么样的新的工作？）你的投资回报计算依赖于是如何将工业机器人与其他相比较的。如果你 要把工业机器人与工人相比，那么你要考虑的不仅是工人的单位时间工作量，还应该包括工人所消耗的的其他资源，这些包括工人的安全保险，休假时间，及其他一 些给工人的奖励，还有招聘与培训新工人的花费。在那些危险性比较高，或者重复性的工作中，机器人能更好的完成任务。

不要小看这么一点潜在节省的劳动力，在很多情况下，潜在劳动力的节省在收入中占有非常重要的意义 .消除资 源浪费和重复性的劳动，特别是对于后续的加工来说,重复性劳动的减少可以提高后续加工的精度.保证第一道工序的加工精度,并在以后的工序中继续保持,可以 大大减少资源与时间的浪费,降低产品的成本,提高产能.(记住:每一个在加工过程中余下的废料都会使用成本上升).而使用机器人,这种浪费可以降到最低.

使用工业机器人,使生产自动化程度大大提高,由于其在重复生产中不会因降低加工精度,因此将极大的提高企业 的收益.现代的企业管理人员都意识到不仅要关注在正常工作时间下的收益,而且还要关注单位设备运行时间所能带来的效益.当然,在某些特殊的工作中,比如焊 接,机器人的速度可能超不过熟练的工人,但是它可以重复的保持一个精度,大批量的产品都符合同一个质量要求.即使机器人连续不停的运转,它的精度还是一样 的,看看吧,使用工业机器人自动化系统工程,将会如何提高公司的收益.

选择标准

选择一个适合的机器人,你首先要关注的是机械手臂,然后是控制系统,机器人的生产厂商(ABB,KUKA,Staubli等世界级的机器人生产商)还有EOAT(End-of-arm tooling).经验认为,避免买到不合适的机器人的方法就是在选择的过程中,多角度的考虑,根据自己车间及企业的要求考虑所要的机器人的类型.

首先来关注机械手.你必须考虑它的工作行程范围,额定负荷,最大负荷,自由旋转轴数,后续工作的适应能力. 现在的机器人手臂都很坚固耐用,生产中应用的重型机器人手臂的大修时间通常是50,000小时.1992年以后生产的机器人手臂,其大修时间已经可以达到 70,000小时,甚至更多.机器人手臂寿命长,升级方便,只要升级一种新的控制器,得到的精度就和你刚买来一样.

接下来是控制器.控制器是机器人的大脑.你必须充分审视控制器的功能,就好像你在招聘员工的时候,审查他的能力一样.你必须关注,控制器是不是易于升级,是否有编程工具并且可以车间内的其它设备通信,是否有提供数据或报告输出功能以方便管理,还有是否可以增加视觉系统.

与其他所有主要的投资一样,你还应该细细的考虑到机器人供应商的技术服务,包括培训与应用支持.

最后,要考虑的是机器人就是EOAT(end-of-arm tooling).第三方的开发可以让你的机器人手臂适应多种功能,完成一些特殊的任务.最常见的就是各种不同尺寸的机械夹持装置.这些装置可以通过电 机、液压或气压驱动。除了这种机械夹持装置外，还有用来抓取表面平的大型零件的真空吸盘，去除毛刺的工具，甚至还有用来拾取销子的筒夹和拾取带孔的零件的 心轴状的拾取装置。EOAT有许多不同的系列可以选择，可以直接从机器人手臂生产商那里得到，或者从第三方厂家那里去采购，当然也可以按照车间的加工要求 定制。

就像公园的雕像一样，工业机器也必须安装后才能使用，根据车间所用加工要求的不同，机器人手臂可以安装在地板上，天花板上，墙上。安装的时候一定要符合工业机器人安全规范 GB 11291-89。可能在安装的时候，还想给工业机器人装上视觉监控系统，还要机器人能输出相关报告以方便生产。每项工作有一个学习过程，如果你的公司对自动化改造有丰富的经验，这些问题就可以自己来解决。要不然，你也可以选择定制的系统。

原文链接: Your first Robot?
附录:

中华人民共和国机械电子工业部1988–0–8–10批准1990–02–01实施

1　主题内容及适用范围
本标准规定了设计、制造、安装、维护、使用工业机器人(以下简称机器人)及机器
人系统的安全准则。
本标准适用于机器人及机器人系统。
2　引用标准
GB 2893　安全色
GB 2894　安全标志
GB 4053.3　固定式工业防护栏杆
GB 4064　电气设备安全设计导则
GB 4776　电气安全名词术语
GB 5083　生产设备安全卫生设计总则
GB 6527.2　安全色使用导则
3　术语
3.1　限位装置
用停止机器人全部运动来限定其工作范围的装置。
3.2　限定工作范围
指由限位装置所限定的机器人工作空间。
3.3　危险区域
人或障碍物进入其中会发生危险的机器人周围区域。其确定方法见附录A(补充件)。
3.4　报警装置
一种发光或发声装置，用于警告潜在人身或其他不安全因素。
3.5　现场安全传感装置
一种用来探测人或物体对现场侵优的安全装置(例如光帘、压敏地板垫、接近式探
测器以及视觉安全系统等)。
3.6　安全防护设备
具有安全防护功能的设备或装置(例如安全防护栏杆、紧急停机装置、防止越程装
置、报警装置等)。
3.7　安全防护措施
为实现安全防护所采取的手段、方法。
3.8　用户
购买或租用机器人、对使用机器人直接负责的单位或部门。
4　基本要求
4.1　机器人的安全防护，应包括其自身的安全防护功能和使用、管理中的安全防护
措施两方面。
4.2　在机器人的工作区，要设置阻止人进入危险区域的安全防护栏杆，且当人误入
危险区域时，机器应具有报警、停机功能。
因示教、检修、故障处理等不得不进入危险区域时，应采取相应的安全防护措施。
4.3　与安全防护有关的全部设备及措施，必须通过验证以确保安全可靠。
4.4　在机器人系统应用过程中，发现新的危险因素时，应及时采取相应的安全措施。
5　安全设计
安全设计应符合GB 5083及GB 4064的有关要求。
5.1　基本要求
5.1.1　为防止机器人误动作产生的危险性，应具有异常状态检测、显示、报警、紧
急停机等安全防护功能。
5.1.2　为防止其他人员误入危险区，应具有异常检测、显示、报警、紧急停机等安
全防护功能。
5.1.3　附属设备发生故障时，应具有异常检测、显示、报警、紧急停机等安全防护
功能。
5.1.4　由于机器人系统的故障而停机时，应具有显示、通知外部的功能。
5.1.5　为确保在危险区进行示教作业的安全性，机器人应具有安全动作速度的功能，
机器人慢速运动的最大值不超过0.25m/s。
5.1.6　在特殊环境中使用的机器人，应具有适应环境的安全防护功能。
5.1.7　控制装置面板上各操作键应设动作方向标记，与操作机各轴动作方向标记相
一致。
5.1.8　控制装置与操作机上要有醒目的安全标志。
5.1.9　机器人系统、动力装置等都要分别设置可靠的接地端子。
5.2　动力源
5.2.1　动力切断装置(开关、离合器、液压及气压控制阀等)应与其他装置分开，
单独设置。
5.2.2　动力切断装置受振动等影响时，不应产生松动、自动闭合或断开。
5.2.3　电压、液压、气压等发生异常变化或因周围的电磁干扰出现异常时，为防止
危险发生，应具有检测、报警、显示及紧急停机等安全防护装置。
5.2.4　用液压、气压等驱动的机器人，应设置易于安全释放驱动器内残压的卸荷机
构。
5.3　紧急停机功能
5.3.1　每台机器人都应具有紧急停机功能。一旦发出紧急停机指令，机器人的运动
应立即停止。
5.3.2　机器人控制装置应设置紧急停机按钮开关。
5.3.3　紧急停机按钮开关采用红色，其形状应有别于一般开关。除手持式控制板外，
紧急停机按钮开关一般应采用蘑菇头形。
5.3.4　紧急停机装置应设置在易于操作的部位。
5.3.5　在紧急停机装置的回路内，应采取可靠性措施。
5.3.6　紧急停机后，不应自动复位；恢复机器人工作，应按规定程序重新启动。
5.4　控制装置
5.4.1　基本要求
5.4.1.1　操作板上按钮、旋钮等的功能或作用的标志要醒目。
5.4.1.2　紧急停机按钮开关应设置在操作板上易于操作的部位。
5.4.1.3　在备有两个以上操作板的场合，应在每个操作板上设置紧急停机按钮开关。
5.4.1.4　通过选择开关来设定机器人的示教方式时，能自动设定机器人的安全动
作速度。
5.4.1.5　启动开关应具有预防无意识启动的结构。
5.4.1.6　在操作面板上应设置正常工作指示灯和故障指示灯。
5.4.2　固定式操作板
固定式操作板应符合下列要求：
a. 在固定式操作板上设置的状态选择开关，除手动外，应不能用其他方式选择或
变更状态；
b. 在固定式操作板上设置带保护装置的启动天关，以免无意识启动机器人。
5.4.3　手持式操作板
手持式操作板应符合下列要求：
a. 在手持式操作板上进行操作时，除紧急停机开关时，固定式操作板上相应的控
制功能应中断；
b. 在手持式操作板上设置的示教按钮开关，应具有松开自动停机的功能；
c. 连接手持式操作板的电缆，应具有可靠的绝缘性、耐磨性和强度。
5.4.4　输入输出端子
为实现安全防护功能，在机器人及其附属设备的控制装置上，要设置紧急停机的
信号输入及输出端子。输入、输出端子要有符号标记。
5.4.5　防干扰
控制装置的电子线路应采取抗电磁干扰的措施，以防外界及电网的电磁干扰使机
器人产生误动作。
5.5　操作机
5.5.1　防止越程的功能
机器人运动关节的始、终点应设有机械式限位装置，能使在额定负载或最大速度
下运动的机器人停机，具有防越程功能。
5.5.2　吊环和吊钩
5.5.2.1　为能进行安全搬动，机器人上应装有吊环、吊钩，其安装位置要考虑机
器人的构造、重心，以免发生异常倾斜，造成损伤。
5.5.2.2　为防止运输过程中振动、倾斜等引起的自然运动，应设有牢靠固定活动
部分的装置。
5.5.3　夹持器
夹持器应符合下列要求：
a. 除去作业必要的部分外，不能有锯齿状或锐利的边缘、突起等危险部分；
b. 即使在紧急停机情况下，夹持器也能牢固夹紧被夹持物。
5.5.4　电动机器人每个轴上均应设有可靠的制动装置。
6　使用时的安全防护
6.1　基本要求
6.1.1　根据机器人的使用条件明确危险区域，同时设置安全防护栏杆及显示机器人
工作状态的装置，防止其他人员进入危险区域。
6.1.2　固定式操作板不可设置在危险区域内。当危险区域有操作人员时，不允许在
自动状态下操作。
6.1.3　用户应对每一项与机器人系统有关的操作建立安全防护措施，并制订安全操
作规程。
6.1.4　应确定限定工作范围并给出显著标志。
6.2　安全防护栏杆
6.2.1　安全防护栏杆应具有足够的强度，且不易移动、撤除和跨越。
6.2.2　安全防护栏杆表面应光滑，不得有锯齿状或锐利的边缘、突起等危险部分。
6.2.3　安全防护栏杆如设栅门，在栅门打开时，机器人应紧急停机。
6.2.4　在安全防护栏杆近旁应定位设置警告牌，以防人误入危险区域。警告牌的设
置见GB 2894中2.2条规定。
6.3　平台
为便于操作维修需设平台时，应符合下列要求：
a. 平台应具有足够的面积和强度；
b. 平台的周围应设栏杆或挡板。
6.4　照明
为保证操作人员的作业和安全，除应对作业现场整体照明外，必要时应进行局部
照明。照度见GB 5083中2.8的规定。
6.5　操作人员的安全防护
6.5.1　操作人员应经过培训，取得安全操作的资格，并具有判断和处理事故的能力。
6.2.2　操作人员应配备相应的防护用品。
6.6　示教人员的安全防护
6.6.1　示教人员应熟悉示教程序和机器人安全操作规程。
6.6.2　在示教机器人之前，示教人员应目视检查机器人及操作的工作范围，以肯定
不存在引起事故的条件。手持式控制板须经功能试验才能用于操作。
6.6.3　在示教时，只有示教人员才允许进入限定工作范围。
6.6.4　在进入限定工作范围之前，示教人员应佩戴防护用品，并肯定所有必要的防
护设备都在位且功能正常。
6.6.5　当选定示教方式后：
a. 在限定工作范围内的示教人员，应单独控制机器人系统；
b. 机器人不响应任何其他导致运动的信号；
c. 在限定工作范围内，其他设备的运动也应置于示教人员的单独控制之下；
d. 机器人系统所有的紧急停机装置，应保持良好的功能。
6.6.6　在启动自动操作方式之前，示教人员应离开限定工作范围。
6.7　维修人员的安全防护
6.7.1　机器人及机器人系统的维修人员应经过安全操作的培训。
6.7.2　应预防机器人偶然的误动作对维修人员的伤害。
6.7.2.1　在给机器人通电、进入工作范围执行维修任务之前，应完成下列工作；
a. 先目检机器人系统是否存在导致误动作的因素；
b. 对控制装置应进行功能试验，以保证其正常操作。
6.7.2.2　当接通电源工作时，在限定工作范围内执行维修任务的人员，应规定专
人控制机器人及机器人系统。
7　安装的安全要求
制造厂家在安装使用说明书中应提出下列各项安装的安全要求：
7.1　机器人安装应有足够的空间，场地要平整，地基要结实、无油污，防止人员滑
跌。
7.2　液压、气压及电压的许可变动范围。
7.3　安全装置的种类、性能、设置方法以及设置时有关安全的注意事项。
7.4　搬运的方法以及搬动时有关安全的注意事项。
7.5　自动运行时(包括启动时及故障发生时)有关安全的注意事项。
7.6　检查、维修、调整、清扫及加油时的作业方法和有关安全注意事项。
7.7　常规检查、抽检及定期检查的项目、方法、判断依据及实施时间。
7.8　机器人系统的安装应尽量减少对建筑物、公用设施、其他机器及设备的干扰，
并考虑到特殊环境(如易燃物、腐蚀、潮湿、灰尘、温度、电磁干扰等)下，保护机器
人能正常工作。
7.9　机器人系统应可靠接地，接地电阻应小于1Ω。
8　试运转
8.1　当安全防护装置不在规定位置而需要进行试运转时，应采取可靠的临时的防护
措施。
8.2　在试运转期间，如无防护装置及措施，人不能进入限定工作范围。
8.3　最初启动过程应至少包括的内容。
8.3.1　通电前应检查下列内容：
a. 机械安装及其稳定性；
b. 电路连接；
c. 通信连接；
d. 外围设备和系统；
e. 限定工作范围的限位装置。
8.3.2　通电后要证实下列内容：
a. 每个轴都运动，且符合设计要求；
b. 紧急停机装置功能正常；
c. 电源切断功能正常；
d. 程序执行符合要求；
e. 互锁功能正常；
f. 安全防护功能正常；
g. 在慢速下机器人能正常运行，并具有作业能力；
h. 在自动状态下，机器人能按工作速度正常运行操作。
8.4　在控制装置的硬件或系统软件更改、机器人关键部件维修之后，通电前应检查
硬件系统和机器人的功能。
9　培训
用户须保证：
a. 对编程、示教、操作、维修人员进行培训，培训包括学习本标准及所用机器人
的安全防护规程；
b. 安装和使用特殊用途的机器人，要进行专门的培训。

附　录　A
危险区域的确定方法
(补充件)
A1　危险区域由限定工作范围及周围隔离带构成。限定工作范围见本标准3.2条。隔
离带的宽度为1–1.5m。

Wall-climbing autonomous vehicle wins robotics competition at international conference
 added by chinarobot

A wall-climbing, book-sized autonomous vehicle made by a Duke University team drove up a challenging vertical course to win first prize in an international competition Sept. 22-24 in Madrid.
The student competition was part of the seventh annual International Conference on Climbing and Walking Robots.

Jason Janet, an adjunct professor in Duke's electrical and computer engineering department and faculty advisor on the robotics project, said the Madrid competition shows the growing importance of climbing robots.

"Robots that climb walls and cross ceilings can go where humans can't," Janet said. "They can do security and safety jobs like looking for bombs or finding cracks in a support beam or the wing of a jumbo jet."

The Duke team's leader was Brian Burney, a staff member at Duke's Pratt School of Engineering and graduate student at North Carolina State University. The other team members were Pratt School undergraduates Kevin Parker, Andrew Meyerson and Julien Finlay.

"Our robot Wallter was the only one that could start flat on the floor and climb the wall on its own, go over a barrier across the wall or stop itself after crossing the finish line," Burney said.

Added Meyerson, "As the smallest, fastest and most novel robot, Wallter was one of the most popular exhibits. I was interviewed for Spanish national television for a story about the conference featuring the Duke robot."

According to Burney, the Duke vehicle set itself apart when it rolled to the foot of a metallic wall, reared up on its hind wheels, and used a "tornado in a cup" to hug the wall and start its ascent.

The "tornado" is generated by a patented device from Vortex HC, LLC of Morrisville, N.C., said Janet, who is vice president of development at the company. The device uses air currents swirling in a cylinder, about the size of an upside-down tuna can, to exert suction on a wall or ceiling. An impeller in the cylinder spins like a propeller but recirculates captive air rather than sucking air in one end and blasting it out the other.

"It's a tornado in a cup, but no ordinary tornado," Janet said. "Two vortexes swirl simultaneously, one in a spiral and the other in a toroidal path, like a donut. The forces generated hold the vehicle to the wall and yet allow free movement because the cup never touches the surface."

Parker said the Madrid competition required performing five tasks: starting on the metal competition wall and climbing as high as possible; climbing after the addition of randomly placed obstacles; crossing a barrier placed on the wall; starting from the floor and then climbing; and stopping after crossing the finish line.

"We faced stiff competition from German and Italian teams," Parker said. "The robot from the University of Catania was amazingly good at detecting and avoiding all the obstacles. Our robot brushed against a couple of obstacles, but it was the only one that completed all five tasks."

Janet said the Duke team combined the "tornado in a cup" technology with an original control system. "A human operates Vortex's commercial robots by remote control," Janet said. "The students added sensors and wrote software that enables their robot to operate on its own."

Parker said they added ultrasonic and infrared sensors across the front and programmed a tiny computer, called a microcontroller, to navigate based on information from the sensors. Ultrasonic sensors detect objects by bouncing sonar-like sound waves off them. Infrared sensors, used in television remote controls, detect light outside the range of human vision.

Burney provided an initial basic design for the Duke vehicle, Janet said. Meyerson and Parker, both biomedical engineering students, focused on writing software and incorporating the sensors.

When tests showed the centimeter-high barrier broke the hold of the Vortex technology, Janet called in Finlay to solve the problem of crossing the barrier without falling off the wall. Finlay is a mechanical engineering student and a veteran of the team that produced Duke's prize-winning autonomous underwater vehicle Charybdis.

Finlay said he tried to design a solution that would work with or without the metal wall at the competition.

"We tried adding treads," Finlay said. "We tried a wheelie bar to keep the rear end of the robot flat against the wall and prevent the front from lifting up. Unfortunately, the results were disappointing. Time was running out so we had to add magnets and take advantage of the metal."

According to Finlay, the magnets were successfully tested only one day before the team flew to Spain.

In Madrid, Meyerson and Parker had to adapt the robot's software for the competition wall. "The traction was different from what we were used to," Meyerson said.

With software tuned and magnets added, Wallter crossed the centimeter barrier without difficulty in practice runs. However, in the first competition runs, Wallter slipped down the wall when attempting to cross."

There were 15 minutes of pure terror and panic," Parker said. "We didn't know what was wrong." Burney said, "We finally realized we had the brackets for the magnets on wrong.

The magnets were upside down, and the magnetism was too weak that way."With the magnets positioned correctly, Wallter negotiated the barrier, reached the top of the wall, and won the first prize of about $250.

The team left Madrid triumphant but exhausted from coping with the competition while keeping Spanish hours without the siestas, said Meyerson. "Restaurants don't open for dinner until nine and a meal takes hours," Meyerson said. "Everyone stays out until four a.m. and that's without even trying to go clubbing."

Janet said Duke's future robotics efforts include teaming with a group from Carnegie Mellon University for the DARPA (Defense Advanced Research Projects Agency) Grand Challenge to design a full-sized autonomous land vehicle and continuing the development of autonomous underwater vehicles.

In addition, computer science professor Ronald Parr and graduate student Austin Eliazar are developing software that enables a mobile robot to map its surroundings as it moves and simultaneously locate itself on the map. Such "simultaneous localization and mapping" is a longstanding challenge in robotics research.

The Duke wall-climbing robot was funded by a grant from the Lord Foundation.Janet said the Vortex technology was developed by Vortex HC on a grant from the DARPA Microsystems Technology Office.

Source: Duke University