smooth planned trajectory

    Good day everyone,
    I'm working on a project which involves generate a path along the surface of any give 3D geometry,I use Matlab to read a from an external stl file to get the surface point and its related normal vector(small triangle mesh and the normal vector),by writing a script processing these datas,I could come up with a series of point containing Cartesian coordinates and the orientation(X,Y,Z,A,B,C), I used SLIN point C_Dis to run these points,supposing that I could get a smooth path,but the robot either says approximation not possible or stutter.
    anyone can teach me how to make the motion smooth,I even used spline block,the stutter disappeared but I got extremly slow speed. is there any rules that we could follow in order to generate the data for points so that we could get a smooth motion?? I used Matlab spline function to generate interpolation point hoping to help the transition in a smooth way,but it doesn't seems like to improve.
    following is my code:
    &ACCESS RVO
    &REL 8
    &PARAM EDITMASK = *
    &PARAM TEMPLATE = C:\KRC\Roboter\Template\vorgabe
    &PARAM DISKPATH = KRC:\R1\1
    DEF shoes( )

    GLOBAL INTERRUPT DECL 3 WHEN $STOPMESS==TRUE DO IR_STOPM ( )
    INTERRUPT ON 3

    BAS(#INITMOV, 0)
    $OUT[4] = False
    $OUT[2] = False
    $OUT[3] = False
    BAS(#VEL_PTP, 1)
    BAS(#ACC_PTP, 0.5)
    slowspeed=0.01
    fastspeed=0.33

    midspeed=0.27

    $VEL.CP=fastspeed
    BAS(#BASE, 1)
    BAS(#TOOL, 1)
    $VEL_AXIS[1]=10
    $VEL_AXIS[2]=10
    $VEL_AXIS[3]=10
    $VEL_AXIS[4]=10
    $VEL_AXIS[5]=10
    $VEL_AXIS[6]=10
    $ACC_AXIS[1]=10
    $ACC_AXIS[2]=10
    $ACC_AXIS[3]=10
    $ACC_AXIS[4]=10
    $ACC_AXIS[5]=10
    $ACC_AXIS[6]=10
    $APO.CDIS=0.1


    ;FOLD SPTP P12 Vel=50 % PDAT4 Tool[1]:Tool1 Base[1]:Base1 ;%{PE}
    ;FOLD Parameters ;%{h}
    ;Params IlfProvider=kukaroboter.basistech.inlineforms.movement.spline; Kuka.IsGlobalPoint=False; Kuka.PointName=P12; Kuka.BlendingEnabled=False; Kuka.MoveDataPtpName=PDAT4; Kuka.VelocityPtp=50; Kuka.VelocityFieldEnabled=True; Kuka.ColDetectFieldEnabled=True; Kuka.CurrentCDSetIndex=0; Kuka.MovementParameterFieldEnabled=True; IlfCommand=SPTP
    ;ENDFOLD
    SPTP XP12 WITH $VEL_AXIS[1] = SVEL_JOINT(50.0), $TOOL = STOOL2(FP12), $BASE = SBASE(FP12.BASE_NO), $IPO_MODE = SIPO_MODE(FP12.IPO_FRAME), $LOAD = SLOAD(FP12.TOOL_NO), $ACC_AXIS[1] = SACC_JOINT(PPDAT4), $APO = SAPO_PTP(PPDAT4), $GEAR_JERK[1] = SGEAR_JERK(PPDAT4), $COLLMON_TOL_PRO[1] = USE_CM_PRO_VALUES(0)
    ;send tape out signal
    PULSE($OUT[3], TRUE, 1.0)
    WAIT FOR ( $IN[10] )
    $OUT[3] = False


    ;ENDFOLD
    $ADVANCE=5

    SLIN {X 400.00,Y -125.13,Z 217.66,A -53.62,B 53.74,C 168.74} C_Dis
    SLIN {X 400.00,Y -124.51,Z 217.32,A -53.60,B 53.76,C 168.77} C_Dis
    SLIN {X 400.00,Y -123.90,Z 216.99,A -53.58,B 53.79,C 168.80} C_Dis
    SLIN {X 400.00,Y -123.29,Z 216.65,A -53.55,B 53.81,C 168.83} C_Dis
    SLIN {X 400.00,Y -122.68,Z 216.31,A -53.53,B 53.83,C 168.85} C_Dis
    SLIN {X 400.00,Y -122.06,Z 215.97,A -53.51,B 53.85,C 168.88} C_Dis
    SLIN {X 400.00,Y -121.45,Z 215.63,A -53.48,B 53.88,C 168.91} C_Dis
    SLIN {X 400.00,Y -120.84,Z 215.30,A -53.46,B 53.90,C 168.94} C_Dis
    SLIN {X 400.00,Y -120.22,Z 214.96,A -53.44,B 53.92,C 168.97} C_Dis
    SLIN {X 400.00,Y -119.61,Z 214.62,A -53.42,B 53.94,C 169.00} C_Dis
    SLIN {X 400.00,Y -119.00,Z 214.28,A -53.39,B 53.97,C 169.02} C_Dis
    SLIN {X 400.00,Y -118.38,Z 213.95,A -53.37,B 53.99,C 169.05} C_Dis
    SLIN {X 400.00,Y -117.77,Z 213.61,A -53.35,B 54.01,C 169.08} C_Dis
    SLIN {X 400.00,Y -117.16,Z 213.27,A -53.32,B 54.03,C 169.11} C_Dis
    SLIN {X 400.00,Y -116.54,Z 212.93,A -53.30,B 54.06,C 169.14} C_Dis
    SLIN {X 400.00,Y -115.93,Z 212.59,A -53.28,B 54.08,C 169.17} C_Dis
    SLIN {X 400.00,Y -115.32,Z 212.26,A -53.25,B 54.10,C 169.19} C_Dis
    SLIN {X 400.00,Y -114.71,Z 211.92,A -53.23,B 54.12,C 169.22} C_Dis
    SLIN {X 400.00,Y -114.09,Z 211.58,A -53.21,B 54.15,C 169.25} C_Dis
    SLIN {X 400.00,Y -113.48,Z 211.24,A -53.19,B 54.17,C 169.28} C_Dis
    SLIN {X 400.00,Y -112.87,Z 210.91,A -53.16,B 54.19,C 169.31} C_Dis
    SLIN {X 400.00,Y -112.25,Z 210.57,A -53.14,B 54.21,C 169.34} C_Dis
    SLIN {X 400.00,Y -111.64,Z 210.23,A -53.12,B 54.24,C 169.37} C_Dis
    SLIN {X 400.00,Y -111.03,Z 209.89,A -53.09,B 54.26,C 169.39} C_Dis
    SLIN {X 400.00,Y -110.41,Z 209.55,A -53.07,B 54.28,C 169.42} C_Dis
    SLIN {X 400.00,Y -109.80,Z 209.22,A -53.05,B 54.30,C 169.45} C_Dis
    SLIN {X 400.00,Y -109.19,Z 208.88,A -53.02,B 54.32,C 169.48} C_Dis
    SLIN {X 400.00,Y -108.57,Z 208.54,A -53.00,B 54.35,C 169.51} C_Dis
    SLIN {X 400.00,Y -108.34,Z 208.41,A -52.99,B 54.36,C 169.52} C_Dis
    SLIN {X 400.00,Y -107.72,Z 208.08,A -52.97,B 54.38,C 169.55} C_Dis
    SLIN {X 400.00,Y -107.11,Z 207.75,A -52.94,B 54.40,C 169.58} C_Dis
    SLIN {X 400.00,Y -106.49,Z 207.42,A -52.92,B 54.42,C 169.61} C_Dis
    SLIN {X 400.00,Y -105.88,Z 207.08,A -52.90,B 54.44,C 169.64} C_Dis
    SLIN {X 400.00,Y -105.26,Z 206.75,A -52.87,B 54.47,C 169.67} C_Dis
    SLIN {X 400.00,Y -104.64,Z 206.42,A -52.85,B 54.49,C 169.70} C_Dis
    SLIN {X 400.00,Y -104.03,Z 206.09,A -52.82,B 54.51,C 169.72} C_Dis
    SLIN {X 400.00,Y -103.41,Z 205.75,A -52.80,B 54.53,C 169.75} C_Dis
    SLIN {X 400.00,Y -102.80,Z 205.42,A -52.78,B 54.56,C 169.78} C_Dis
    SLIN {X 400.00,Y -102.18,Z 205.09,A -52.75,B 54.58,C 169.81} C_Dis
    SLIN {X 400.00,Y -101.56,Z 204.76,A -52.73,B 54.60,C 169.84} C_Dis
    SLIN {X 400.00,Y -100.95,Z 204.42,A -52.71,B 54.62,C 169.87} C_Dis
    SLIN {X 400.00,Y -100.33,Z 204.09,A -52.68,B 54.65,C 169.90} C_Dis
    SLIN {X 400.00,Y -99.72,Z 203.76,A -52.66,B 54.67,C 169.93} C_Dis

  • massula June 29, 2024 at 3:53 AM

    Approved the thread.

    Your points are very close together (less than one millimeter). Robot controllers are not like CNC controllers which can handle close points at high speeds. Did you try to space the interpolated points further apart?

    Quote from geir_hokarob

    Your points are very close together (less than one millimeter). Robot controllers are not like CNC controllers which can handle close points at high speeds. Did you try to space the interpolated points further apart?

    yes,I tried,it doesn't make big differences,unless I made the points very far away from each other,but this makes the robot running away too much from the planned trajectory.

    SLINs, by definition, need to plot their motions in more detail than LINs. An SLIN path must be "twice differentiable", which basically means that the position, velocity, and acceleration (as a function of time) cannot have any step transitions.

    LIN motions are more forgiving, in that the robot is only restrained to stay within the approximation radius of each motion, and is much freer to approximate everything else. Which leads to an easier, but less accurate, path.

    In this case, I don't thin hermann is suggesting that LINs would necessarily solve your problem, but that they might be an experiment worth trying.

Advertising from our partners