This could mean abigger need of smaller and lighter robots with higher precisionand rigidity。 If electric drives and fuel cells will be commonlyused, then new applications will arise, probably with higherrequirements on robot accuracy and short cycle times。However, since it is still manufacturing of products in largevolumes, probably made by the same industrial organisationsthat make cars and car components today, the infra structure ofthe manufacturing system will probably be the same。As pointed out in Section 1, one part of the carmanufacturing that is still labour intensive is the final assembly。Using the robot technology of today for this application,problems arise with the geometric complexity and variability ofthe work objects, with the space needed for safety fences, withdifficult robot programming and with time consuming handlingof assembly failures。 Only few robot installations can beeconomically motivated, as, for example, for the gluing andmounting of car windows。 In this special case, the health risk isone reason for the introduction of robots。 The workingconditions are also the incitement for the introduction ofequipment to assist at heavy lifts。 Here, a development has beenrunning for some years to develop lifting tools from beingpassive mechanical systems to be powered and servocontrolled。 These intelligent assistant devices (IAD) areinteresting with respect to their safety requirements and theforce controlled interaction between the operator and the IAD(Colgate, Peshkin,&Klostermeyer, 2003)。However, this kind ofdevice does not give a real automation solution and the conceptcannot be developed towards a complete robotized final carassembly。 Instead, it could be expected that the automotiveindustrywill push for a newtype of assembly robot systems and anew type of infra structure for flexible automation。In order to find other industries that might drive future robotdevelopment, one possibility is to look at manufacturingsystems lacking the infrastructure needed for the efficient use ofrobots today。 One opportunity here is the need of automation insmall and medium size enterprises (SMErobot, 2005)。 What isneeded in this environment is low cost safe robot systems thatare easy to install, configure, calibrate, program and maintain。The application processes will be the same as those that robotsare used for today and the biggest challenge will be to developrobot technology that gives a much lower robot life cycle cost。Going further in this development direction therewill be a lot ofother important areas for the future use of industrial robots。Examples are flexible automation for disassembly (Steele,2004) and sorting during scrap handling, for cutting up of meat,for other types of food processing and handling (Fig。 6;Hamazawa, 1999), for consumer goods handling, for theassembly and processing of big components for airplanes(KUKA Robotics, 2005), bridges, buildings, ships, trains,trains, power stations, windmills, etc。, and for a wide spectrumof craftsmen tasks。 If economically feasible solutions forflexible robot automation can be found for these types ofapplications a huge new market for industrial robots will drivethe robot technology development in a new direction。Looking at driving forces from future new applicationprocesses for robotics, enhancement of the robot performancewill be very important。 It is, for example, well known that arm-type robots are cheaper to manufacture and install thanCartesian manipulators。 However, the bandwidth, the stiffnessand the accuracy of a Cartesian manipulator can be made muchhigher than for an articulated robot, but if the performance ofarticulated industrial robots is significantly increased theserobots could take over large market shares from the moreexpensive Cartesian manipulators。 Examples of applicationswhere arm-type industrial robots could then be used are highperformance laser cutting, plasma cutting assembly andmachining。
One example where both improved robot perfor-mance and a new flexible automation concept is needed isfettling of iron castings (Lauwers, Wallis, Haigh, & Sohald,2004)。 In this case, a third aspect is also important and that is thevery unhealthy foundry environment。 Having the healthproblems of the workers in mind, several other future drivingapplications for industrial robots can be found。 Examples of thiscan be found in slaughterhouses, cold stores, glazier work-shops, fisheries and garbage handling plants。4。 Scenarios about future industrial robot controldevelopmentEven if it is very difficult to predict the long-termdevelopment directions, some scenarios will be outlined to give an idea about how to make an extrapolation based on thedriving forces that can be observed for the use of robotics。 Sincethe automotive industry is the major force driving the roboticdevelopment today it could be relevant to start looking at thefuture of car manufacturing with a scenario based on the earlierproposed need of more accurate, rigid, fast and slender robotsfor such processes as laser welding, soldering, assembly,riveting and gluing to replace the large number of big heavyrobots used for spot welding。 Assuming the use of fibre opticsand improved Yttrium Aluminium Garnet (YAG) lasertechnology (Rooks, 2000) and lightweight soldering, rivetingand gluing equipment, the weight of the load carried around bythe robot will be radically reduced from that of a spot weldinggun。 This means that it is not far away to look at animplementation scenario for the development of lightweightslender robots with wrist- and upper arm concepts having muchlower mass using integrated actuator solutions, fibre compo-sites and other lightweight materials。 Even if a lightweight highspeed motor together with a lightweight high ratio speedreducer will probably be more expensive than the heavier wristdrive systems used today, the robot installations will probablybe cheaper since the main axes can be equipped with actuatorshaving lower power and since less massive frames can be usedfor the mounting of the robots。 In order to make a slenderlightweight robot stiff with respect to tool forces, the robotcontrol may need to provide a virtual stiffness, for whichsensors are needed in the arm structure of the robot。 Examplesof sensors are capacitive encoders (Fig。 7) with optionalvibration measurement electrodes to measure the angle and thevibrations of the output shafts of speed reducers, joint torquesensors (Pfeffer, Khatib, & Hake, 1989), accelerometers (deJager, 1994) and strain gauges (Feliu, Garcı ´a, & Somolinos,2001) to measure arm vibrations。 These distributed sensorscould then be used by the servo in a sensor fusion manner basedon the real time models running in the controller。 Important inthis scenario will of course be to have accurate dynamicmodels。 An industrial robot is a strongly coupled multivariablesystem with up to 50 mass-spring elements to be modelledtogether with the non-linearities originating from friction andlost motion in gears and other transmission components。 未来工业机器人控制的发展英文文献和中文翻译(5):http://www.youerw.com/fanyi/lunwen_62458.html