12. Final prototype of the iCub head. Height from base to top is 11cm.Electronics are not shown in the picture. Right: prototype with face mounted.V. TESTSSeveral tests were made to evaluate the funcional use ofthe designed head. Every joint was tested, in the worst casescenario, to verify that velocity and acceleration speci cationwhere met. Several other demos/tests were created to verify thecoordination between the sensors and actuators in the system.Object tracking - An object moving in front of the systemis considered to be successfully tracked if it is near the centerof both images. Since this is a 3 dof task and the robothas 6 dof, some extra criteria should be used to control alldof. The eye's 3 dof are directly controlled with a visualservoing mechanism, having the image positions of the objectas feedback. The neck is then controlled in order to maintainthe eyes as far as possible for their joint limits. This choiceis motivated by biological behavior. The rationale for havingthe eyes in a comfortable position is that they will be readilyavailable to track the object in any direction even if it movesvery fast.Balancing - The inertial sensor provides a reliable measureof the orientation of the head and also the angular velocities.This information can be used to keep the head always in aupright position. The inclination information controls directly the  rst 2 dof of the neck. The angular velocity is used tocreate an arti cial vestibular re ex, that in presence of a fastmotion keeps the eyes looking in the same direction.Sound localization - Using information about the time-difference, sound energy level and spectral power at somespecial frequencies, it is possible to localize sound sourceswith the iCub head. This algorithm is very reliable for thehorizontal plane and can be used either with a closed or anopen loop controller. Please refer to http://vislab.isr.ist.utl.ptfor videos of these demos.VI. CONCLUSIONS/FUTURE WORKWe presented the design for a robot head of a small sizehumanoid robot. This humanoid robot - the iCub - is meant tobe used as a research tool for human cognition and publiclydistributed worldwide. The iCub neck has 3 dof and the eyes 3dof, the more complex mechanism for similarly sized robots,where the eyes are usually  xed.The speci cations were derived from human anatomical andbehavioral data. As some of these data are not available (andeven less data can be found for children), extra assumptionshad to be made. We referred to typical tasks the robot shoulddo and the ratio between neck and eyes velocities/accelerationin humans. The  nal kinematic has three dof for the neck.For the eyes, three degrees of freedom (independent vergence,common tilt) were considered. Other movements like cyclo-torsion can be dealt with at the camera level.A  rst solution for the neck was a spring mechanism mim-icking the human anatomy. It has good kinematic capabilitiesbut low repeatability and precision, due to the spring.A very small 3 dof parallel actuator was proposed as oursecond design. It is very compact and delivers high torque.Due to extremely reduced size, a home-made linear actuatorwas designed. It meets all the speci cations, except the desiredrange of motion. Although this could be a very attractivesolution if we could afford using a 20% bigger neck, self-interference of the mechanical parts precluded its use at thecurrent stage.The design that met all the speci cations was a serialmechanism. It has a clutch system to overdrive protection.All motors are similar and the assembly is modular.The lightweight eye system has three dofs, consisting ofindependent eye pan and a common tilt. The head is equippedwith additional sensors, like an inertial sensor for the vestibularsystem, kinesthetic information from encoders, absolute posi-tion sensors in the neck and embedded controllers.
The overallweight is below 1.5Kg (all motors included) and the size isthat of the head of a 2 year old child. A  rst prototype of theiCub face was designed and built.The head has been mounted and tracking experiments weredone to assess the performance of the mechanism, that is quiteencouraging. As future work other experiments will be done,including the use of the microphones and the inertial signals.A re-formulation of the parallel mechanism may be conductedto increase the range of motion.The iCub  nal mechanical design will be freely availableto researchers worldwide, and released under General PublicLicense (GPL) - www.robotcub.org.VII. ACKNOWLEDGMENTSWork partially supported by EU Project IST-2004-004370ROBOTCUB and by the FCT Programa Operacional So-ciedade de Informac ¸ ˜ ao (POSI) in the frame of QCA III.REFERENCES[1] Giulio Sandini, Giorgio Metta, and David Vernon. Robotcub: An openframework for research in embodied cognition. International Journal ofHumanoid Robotics, 8(2), November 2004.[2] Alvin R. Tilley. The Measure of Man and Woman: Human Factors inDesign. Henry Dreyfuss, 2001.[3] L. Geppert. Qrio, the robot that could. IEEE Spectrum, 41(5):34–37,May 2004.[4] M. Hirose, Y. Haikawa, T. Takenaka, and K. Hirai. Developmentof humanoid robot ASIMO. In Workshop on Exploration towardsHumanoid Robot Applications at IROS, Hawaii, USA, 2001.[5] F. Yamasaki, T. Miyashita, T. Matsui, and H. Kitano. PINO thehumanoid: A basic architecture. In Proc. of The Fourth InternationalWorkshop on RoboCup, Melbourne, Australia, August 2000.[6] Fujitsu. Humanoid Robot HOAP-2. http://www.automation.fujitsu.com,2003.[7] Henry Gray. Anatomy of the Human Body. Bartleby.com,http://www.bartleby.com/107/, online edition, May 2000.[8] Loyola University Chicago. Master muscle list. Technical report, StritchSchool of Medicine, 1998.[9] ExRx.net. Muscle body map. http://www.exrx.net/Lists/MMale.html,1999.[10] Wikipedia. List of muscles of the human body. http://en.wikipedia.org/,July 2005.[11] Wikipedia. List of bones of the human skeleton. http://en.wikipedia.org/,July 2005.[12] Miguel Silva. Human Motion Analysis Using Multibody Dynamics andOptimization Tools. PhD thesis, Instituto Superior T´ ecnico, Lisboa,Portugal, 2003.[13] D. Laananen. Computer simulation of an aircraft seat and occupant in acrash environment. Program som-la/som-ta user manual, US Departmentof Transportation, Federal Aviation Administration, 1999.[14] Vladimir Zatsiorsky. Kinematics of Human Motion. Human KineticsEurope Ltd, 1997.[15] Zangemeister and Stark. Active head rotations and eye-head coordina-tion. Ann N Y Acad Sci, 374:540–59, 1981.[16] Julius Panero and Martin Zelnik. Human Dimension and Interior Space:A Source Book of Design Reference Standards,. Watson-Guptill Pubs,1979.[17] Raffaele Di Gregorio. Kinematics of the 3-UPU wrist. Mechanism andMachine Theory, 28:253–263, 2003.[18] G¨ ursel Alici and Bijan Shirinzadeh. Topology optimisation and singular-ity analysis of a 3-SPS parallel manipulator with a passive constrainingspherical joint. Mechanism and Machine Theory, 39:215–235, 2004.[19] Raffaele Di Gregorio. Statics and singularity loci of the 3-UPU wrist.IEEE Transactions on Robotics, 20(4), 2004.[20] Faulhaber. DC-Motors. www.faulhaber-group.com, 2005.[21] Ptgrey. Firewire cameras. www.ptgrey.com, 2005. 机器人iCub头部设计英文文献和中文翻译(4):http://www.youerw.com/fanyi/lunwen_34027.html