工業(yè)機(jī)器人外文翻譯--對移動(dòng)機(jī)械臂在其環(huán)境中使用有效質(zhì)量和阻尼控制所產(chǎn)生的脈沖接觸力進(jìn)行控制（原文）

1、INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 11, No. 5, pp. 697-704 OCTOBER 2010 / 697DOI: 10.1007/s12541-010-0082-4 1. Introduction A mobile manipulator that consists of a mobile base and

2、a robotic manipulator provides numerous advantages over a fixed-base manipulator. The most important advantage of the mobile manipulator is its flexible workspace. Also, the mechanical configuration of the robotic manipu

3、lator mounted on the mobile base results in a redundant system that is advantageous for dexterous manipulation.26 While these features enable various applications, control of the mobile manipulator is challenging due to

4、its intrinsic structure. First, the redundant DOFs of the mobile manipulator complicate manipulation control. Second, the overall dynamics of the mobile manipulator is far more complex compared to fixed-base manipulators

5、. Third, many applications involve dynamic interaction of the manipulator with its environment. Mobile manipulation in human populated environments brings the issue of safety to the forefront. There have been a number of

6、 studies on mobile manipulation. Yamamoto proposed kinematic/dynamic manipulability to resolve the redundancy in motion planning and analysis for mobile manipulator.1 Khatib2,3 introduced potential field in controlling p

7、osition of redundant manipulators and providing joint limit constraints. Also, he proposed a novel approach to utilize the intrinsic dynamics characteristics of mobile manipulator in operational space.2,3 Additionally, h

8、e analyzed the reduced inertia effect of redundant robotic arm with macro-micro structure. In other studies on mobile manipulation, ZMP (Zero Moment Point) was proposed as a measure for mobile manipulator systems perform

9、ing heavy-load transferring tasks.4,5 There have been studies on the stable motion control of a mobile manipulator under unknown external force application from the environment.6,7 Many of the studies mentioned above ada

10、pted the idea of conventional fixed-based mechanisms to mobile manipulators. There are several issues to consider in mobile manipulator control. First, the mobile base part of the mobile manipulator is heavier than the u

11、pper robotic manipulator part, which leads slower dynamic response of the mobile base compared with that of the manipulator part. The overall dynamics of the mobile manipulator is somewhat analogous to that of macro-micr

12、o manipulators. The macro part is much heavier than the micro part.2,3 Second, the Control of Impulsive Contact Force between Mobile Manipulator and Environment Using Effective Mass and Damping Controls Sungchul Kang1,

13、 Kiyoshi Komoriya2, Kazuhito Yokoi2, Tetsuo Koutoku2, Byungchan Kim1 and Shinsuk Park3,#1 Center for Cognitive Robotics Research, KIST, Hawolgok-dong, Seongbuk-gu, Seoul, South Korea, 136-791 2 Intelligence Systems Rese

14、arch Institute, AIST, Tsukuba, Ibaraki, Japan, 305-8568 3 School of Mechanical Engineering, Korea University, Anam-dong 5-ga, Seongbuk-gu, Seoul, South Korea, 136-713# Corresponding Author / E-mail: drsspark@korea.ac.kr,

15、 TEL: +82-2-3290-3868, FAX: +82-2-926-9290KEYWORDS: Mobile manipulation, Effective inertia, Kinematic Redundancy, Null space motion Recently, mobile manipulators are being widely employed for various service robots in hu

16、man environments. Safety is the most important requirement for the operation of mobile robot in a human-populated environment. Indeed, safe human-machine interaction is one of grand challenges in robotics research. Thi

17、s paper proposes a novel control method to reduce impulsive compact force between a mobile manipulator and its environment by using optimized manipulator inertia and damping-based motion control. To find the optimized

18、configuration through null space motion, the combined potential function method is proposed considering both the minimum effective mass and joint limit constraints. The results of this study show that the inertia optim

19、ization along with a damping controller significantly reduces the impulsive force upon collision and the contact force after collision. Manuscript received: January 6, 2009 / Accepted: July 26, 2010© KSPE and Sprin

20、ger 2010 INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 11, No. 5 OCTOBER 2010 / 699Table 1 D-H parameters and link mass of M3 mobile manipulator joint 1 i α + (rad) 1 i a + (m) 1 i θ + (r

21、ad) 1 i d + (m) Joint type Mass (kg) 1 / 2 π ?0 0 0 P 90.01 2 / 2 π ?0 / 2 π 0 P 3 0 0 π 7 R 4 / 2 π ?0 0 0 R 7.81 5 / 2 π0 0 0 R 3.41 6 / 2 π ?0 0 0.45 R 1.18 7 / 2 π0 0 0 R 7.45 8 /

22、 2 π ?0 0 0.5 R 2.64 9 / 2 π0 0 0 R 1.46 10 0 0 0 0.08 R 0.24 P : Prismatic, R : Revolute 2.1 Effective Mass of Mobile Manipulator For a mobile manipulator, the inertial properties in joint space are descr

23、ibed by kinetic energy matrix A(q), which is a matrix function of joint variables q. When the dynamic response or impact force at the end-effector is of interest, the inertial properties can be evaluated at the end-effec

24、tor, which is represented by kinetic energy matrix A(q) in Cartesian space. Kinetic energy matrix A(q) is calculated from kinetic energy matrix A(q) using Jacobian matrix J(q).18 1 1 ( ) ( ( ) ( ) ( )) T ? ? = Λ q J q A

25、q J q(1) The inertia perceived at the end-effector in an arbitrary direction u is represented by effective mass Mu(q) as follows: 1 1 ( ) ( ) T M ? = u q u Λ q u (2) Effective mass Mu(q) represents directional inertia

26、 property of the manipulator at the end-effector. With a redundant manipulator such as M3 mobile robot, effective mass Mu(q) can be manipulated by changing the configuration of the manipulator, even when the position and

27、 orientation of the end-effector is fixed in Cartesian space. This feature allows us to control the impulsive force between the end-effector and the environment for a given pose (position and orientation) of the end-effe

28、ctor: If effective mass Mu(q) is reduced, the impulsive force upon collision can also be reduced since the force is mainly dependent on the inertia and velocity difference.19,20 2.2 Change of Effective Mass by Null Spac

29、e Motion For a given pose (position and orientation) of the end-effector, the configuration of a redundant manipulator can be changed by generating null space motion. The following equation depicts inverse kinematics of

30、a redundant manipulator.21 # # ( ) [ ( ) ( )] = + ? q J q v I J q J q φ ?(3) Here, J#(q) is pseudo-inverse of Jacobian matrix J(q), and φ represents an arbitrary joint motion. The matrix [I - J#(q)J(q)] defines the null

31、space associated with J#(q), and the vector [I - J#(q)J(q)]φ corresponds to the zero variation of the pose of the end-effector. By setting the velocity of the end-effector zero (v = 0), the pose of the end-effector is fi

32、xed in Cartesian space. For a redundant mobile manipulator, effective mass Mu(q) at a given pose of the end-effector can be changed by producing the null space motion of the mobile base. In case of M3 mobile robot, the

33、null space motion can be produced by moving the mobile base in X-direction (φ = (k,0,0,..,0)T) or Y-direction (φ = (0,k,0,..,0)T) of the coordinate system shown in Figs. 1 and 2. Figure 3 illustrates the changes of effec

34、tive mass in X-direction (Mx(q)), Y-direction (My(q)) and Z-direction (Mz(q)) as the position of the mobile base changes in XY-plane. As can be seen in the figure, effective mass Mu(q) is a simple function of X and Y for

35、 the given range of null space motion, while kinetic energy matrix in Cartesian space Λ(q) is a highly coupled function of joint variables q. 2.3 Minimization of Effective Inertia From the plots of the inertia property o

36、f M3 mobile robot for the given range of null space motion (Fig. 3), we can easily find the configuration that minimizes the effective mass in a specific direction by using optimization techniques. This can be formulated

37、 as an optimization problem with equality constraints. In the optimization problem, the objective function to be minimized is a scalar function Mu(q) that is effective mass in u-direction, and the constraint function is

38、the forward kinematics equation of the mobile manipulator g(q)-χ = 0, where χ is a given pose of the end-effector. Thus this problem becomes: Minimize Mu(q) subject to g(q)- χ = 0 (4) Similarly, the maximization problem

39、 yields to: Maximize Mu(q) subject to g(q)- χ = 0 (5) For a given end-effector pose, joint angle vector qm of the configuration that minimizes (or maximizes) effective mass Mu(q) can be obtained by using constrained opt

40、imization functions22 provided in MATLAB. When the end-effector of M3 mobile manipulator is kept horizontal to the ground with the height of 1.32 m, the minimum and maximum effective masses are calculated to be 11.75 kg

41、and 15.28 kg, respectively. The mobile manipulator can move from the current configuration to the minimum inertia configuration with joint trajectories calculated from the inverse kinematics equation of Eq. (3) and resol

42、ved motion rate control. Null space vector φ of Eq. (3) is often written as the gradient of an objective function to be minimized. In this study the objective function is chosen as to move the mobile manipulator to the m

43、inimum inertia configurations. The objective function has the form of potential function as follows: 2 ( )Nm i mi i V q q = ? ∑(6) Here, N is the number of joints, qi is the current joint angle of i-th joint, and qmi is