2R-manipulator Geometric Modeling
Q1: Direct Geometric Model |
(θ1,θ2)→(x,y)(\theta_1,\theta_2) \rightarrow (x,y)(θ1?,θ2?)→(x,y)
OP→=OA→+AP→\overrightarrow{OP}=\overrightarrow{OA}+\overrightarrow{AP}OP=OA+AP
OP→=l1u→+l2v→\overrightarrow{OP}=l_1\overrightarrow{u}+l_2\overrightarrow{v}OP=l1?u+l2?v
Here,we know u?=[cosθ1sinθ1]\vec{u}=\begin{bmatrix} cos\theta_1 \\ sin\theta_1 \end{bmatrix}u=[cosθ1?sinθ1??],and v1?=[cos(θ1+θ2)sin(θ1+θ2)]\vec{v_1}=\begin{bmatrix} cos(\theta_1+\theta_2) \\ sin(\theta_1+\theta_2 )\end{bmatrix}v1??=[cos(θ1?+θ2?)sin(θ1?+θ2?)?],
For the direct geometric model,we have
Px=l1cosθ1+l2cos(θ1+θ2)P_x=l_1cos\theta_1+l_{2}cos(\theta_1+\theta_2) Px?=l1?cosθ1?+l2?cos(θ1?+θ2?)
Py=l1sinθ1+l2sin(θ1+θ2)P_y=l_1sin\theta_1+l_{2}sin(\theta_1+\theta_2) Py?=l1?sinθ1?+l2?sin(θ1?+θ2?)
Q2: Inverse Geometric Model |
(x,y)→(θ1,θ2)(x,y) \rightarrow (\theta_1,\theta_2)(x,y)→(θ1?,θ2?)
As we would like to consider about the inverse geometric model,
x2=l12cos2θ1+2l1l2cosθ1cos(θ1+θ2)+l22cos2(θ1+θ2)x^2=l_1^2cos^2\theta_1+2l_1l_2cos\theta_1cos(\theta_1+\theta_2)+l_2^2cos^2(\theta_1+\theta_2)x2=l12?cos2θ1?+2l1?l2?cosθ1?cos(θ1?+θ2?)+l22?cos2(θ1?+θ2?)
y2=l12sin2θ1+2l1l2sinθ1sin(θ1+θ2)+l22sin2(θ1+θ2)y^2=l_1^2sin^2\theta_1+2l_1l_2sin\theta_1sin(\theta_1+\theta_2)+l_2^2sin^2(\theta_1+\theta_2)y2=l12?sin2θ1?+2l1?l2?sinθ1?sin(θ1?+θ2?)+l22?sin2(θ1?+θ2?)
We add these two equations,we know that
x2+y2=l12+l22+2l1l2cosθ2x^2+y^2=l_1^2+l_2^2+2l_1l_2cos\theta_2x2+y2=l12?+l22?+2l1?l2?cosθ2?
So we know that
cosθ2=?l12?l22+x2+y22l1l2cos\theta_2=\frac{-l_1^2-l_2^2+x^2+y^2}{2l_1l_2}cosθ2?=2l1?l2??l12??l22?+x2+y2?
equation(R):θ2=+/?arccos(?l12?l22+x2+y22l1l2)equation (R):\theta_2=+/-arccos(\frac{-l_1^2-l_2^2+x^2+y^2}{2l_1l_2})equation(R):θ2?=+/?arccos(2l1?l2??l12??l22?+x2+y2?)
QUESTION 1:How about θ1(x.y)\theta_1(x.y)θ1?(x.y)?For inverse geometric,if we know the position of the end-effector(x,y)(x,y)(x,y),we could get 2 solution which could be observed in equation R.We have positive and negative θ2\theta_2θ2?,and each solution has its corresponding value of θ1\theta_1θ1?.So why there is no equation for θ1(x.y)\theta_1(x.y)θ1?(x.y). I still feel confused with it.
Feedback:knowing θ2\theta_2θ2?, you end up with a linear system of equations with two equations and two unknown (cos(θ1)(cos(\theta_1)(cos(θ1?) and sin(θ1))sin(\theta1))sin(θ1))
Then we would like to calculate the Kinematic Jacobian Matrix.
Q3: Kinematic Jacobian Matrix |
dPdt=d(OA→+AP→)dt=l1du→dt+l2dv→dt=l1Euθ1˙+l2Ev(θ1˙+θ2˙)\frac{\mathrm{d}P}{\mathrm{d}t}=\frac{d({\overrightarrow{OA}+\overrightarrow{AP}})}{dt}=l_1\frac{\mathrm{d}\overrightarrow{u}}{\mathrm{d}t}+l_2\frac{\mathrm{d}\overrightarrow{v}}{\mathrm{d}t}=l_1\mathbf{E}\mathbf{u}\dot{\theta_1}+l_2\mathbf{E}\mathbf{v}(\dot{\theta_1}+\dot{\theta_2})dtdP?=dtd(OA+AP)?=l1?dtdu?+l2?dtdv?=l1?Euθ1?˙?+l2?Ev(θ1?˙?+θ2?˙?)
Where E=[0?110]E=\begin{bmatrix} 0 &-1 \\ 1&0 \end{bmatrix}E=[01??10?]Then we get
P˙=(l1Eu+l2Ev)θ1˙+l2Evθ2˙\dot P =(l_1\mathbf{E}\mathbf{u}+l_2\mathbf{E}\mathbf{v})\dot{\theta_1}+l_2\mathbf{E}\mathbf{v}\dot{\theta_2}P˙=(l1?Eu+l2?Ev)θ1?˙?+l2?Evθ2?˙?
We use this kind of way(using vector).
P˙=E[R1R2][θ1˙θ2˙]\dot P=\mathbf{E}[\mathbf{R_1\quad R_2}]\begin{bmatrix} \dot{\theta_1} \\ \dot{\theta_2} \end{bmatrix}P˙=E[R1?R2?][θ1?˙?θ2?˙??]
When we talk about singularity it always equal to calculate ∣∣J∣∣||J||∣∣J∣∣ or det(J)=0det(J)=0det(J)=0.
E[R1R2]=[0?110][l1cosθ1+l2cos(θ1+θ2)l2cos(θ1+θ2)l1sinθ1+l2sin(θ1+θ2)l2sin(θ1+θ2)]\mathbf{E}[\mathbf{R_1\quad R_2}]=\begin{bmatrix} 0 &-1 \\ 1&0 \end{bmatrix}\begin{bmatrix} l_1cos\theta_1+l_{2}cos(\theta_1+\theta_2) &l_{2}cos(\theta_1+\theta_2)\\ l_1sin\theta_1+l_{2}sin(\theta_1+\theta_2) &l_{2}sin(\theta_1+\theta_2) \end{bmatrix}E[R1?R2?]=[01??10?][l1?cosθ1?+l2?cos(θ1?+θ2?)l1?sinθ1?+l2?sin(θ1?+θ2?)?l2?cos(θ1?+θ2?)l2?sin(θ1?+θ2?)?]
We have θ2=0\theta_2=0θ2?=0 to make the rank<2rank<2rank<2
QUESTION2:When passing close to this singularity(maybe a type of Elbow singularity),how can we analyze what will happen to the robot?
Feedback:the kinematic Jacobian matrix gives you the motions that are not possible
We have θ2=π\theta_2=\piθ2?=π to make the rank<2rank<2rank<2
I have found a video “Meca500”,it introdruces3tpyes of singularities(Note,the last one is about the situation which is passing close to shoulder singularity)
wrist singularity | Elbow singularity | Shoulder singularity |
---|---|---|