# Lecture 8 ## movie segment Automatic Parrallel parking ## Kinematic Singularty: * Well i didn't understood his explaination of singularity using that.. * rank of matrix decreases if there is singularity. * here we will get deteminate as (-------)qdot and we simply need to equate it to zero. * Why determinate of J is independent of frame ?:in second slide below if we take determinant both side , (othogonal matrix has determinant equal to one.) we get second statment below. * ![](https://i.imgur.com/X7uxKYf.png) * When Theta2 approches zero or pi or n*pi joint approches singularity. * ![](https://i.imgur.com/KCGiyGV.png) * At singularity i.e, when Theta2 equal to zero we get below given J1 matrix. * What Theta1 they have assumed to get that final J1? ( In first slide below) : It is J0 but still doubt not resolved)[27:28](https://youtu.be/XrNdB4k5kUk?t=1648) * We are looking at the displacement by $\delta x \,and\, \delta y$ when we make small displacment $\delta \theta$ * The two column are dependent and therfore matrix has rank = 1. * How $\delta \theta$ equal to zero mathematically(Didn't understood it well)? * When there is small $\delta x \,and\, \delta y$ their is no displacment along x axis therefore $\delta x$ = 0 and almost all displacment is along tangent direction and therefore almost all is along y direction.(This all is in frame 1) * ![](https://i.imgur.com/l0pzGCs.png)This is actually J0 and also in formula above it. * Now if we want to find the $\delta \theta$ required for small displacement $\Delta X$ we get inverse of jacobian.( Here is it a inverse kinematics ?) (second slide) * ![](https://i.imgur.com/izzqfcx.png) * Now we get appoximation of Delta_q1 and Delta_q2. * i didn't understood the graph (So basically Delta_q1 is change in angle q1 then why the change in angle be infinite as angle gets smaller it feels to me quite contradicting)? * ![](https://i.imgur.com/pKsPhtd.png) * As we are having only two degrees of freedom we can analyze only ${J}_v$ instead of whole Jv and Jw. * Now for 3 DOF arm in ${2}^{nd}$ slide we get Jacobian as below: ![](https://i.imgur.com/diYMt7I.png) * How is he reducing J matrix to 3 rows he has elemenated 3 rows of zeros(i didn't understood the maths part)? * ![](https://i.imgur.com/6HfkmaN.png) * ![](https://i.imgur.com/s2xad65.png) * When ${\theta}_5$ goes to zero we get below matrix where column 4 and 6 are identical and rank of matrix becomes 5. * ${V}_e$ is end effector velocity . ![](https://i.imgur.com/SkzirTt.png) * We have written it in reverse order because now we can write it in matrix form using Identity. * Here we are using cross product operator i.e, ${\hat P}_{ne}$. * The ${J}_n$ (is defined in frame n) is related to ${J}_e$ with matrix. * Now if Jn is in frame zero then the P_hat opereator should be also in frame zero. * How we will convert P operator from nth frame to 0th frame?ye yahi hai jo mujhe samajh nahi raha hai assignment ke 1st question mai bhi usne yahi kiya tha.: We use here similarity transform i.e, we premultiply and postmultiply. Would like to get further clarification ....[48:08](https://youtu.be/XrNdB4k5kUk?t=2928) * ![](https://i.imgur.com/YSd5frj.png) * how this second matrix came? ![](https://i.imgur.com/uWCaZzx.png) * ![](https://i.imgur.com/QKUMCTI.png) * He has not explained anthing here would require more clarification.: Here he has used product operator formula to get $\hat P$ from P ## Resolved Motion Rate control: * so here we know $\theta$ then we find $\delta x$ then with it $\delta \theta$ then we find increment $\theta$ with which then we move the robot. Is it enough to know the desired angle to move robot like hoe it works ? * ![](https://i.imgur.com/YeNPPbE.png) * This motion rate control might work for small displacement otherwise it gives lots of errors. ## Static Forces: * We are going to find relation btw forces excerted by the end effectors and torques produced at the joints. * ![](https://i.imgur.com/o7jBRE0.png) * first equation represent sort of a Jacobian relating velocities V and Omega or theta dot and the second relation is representing sort of transfers of the Jacobian ### Fundamental equations in kinematics of robotics: * These two eauations are fundamentas in kinematics in robot: $\nu = J\dot{\theta}\,and \,\tau = {J}^T F$. * ![](https://i.imgur.com/8btuAin.png) ![](https://i.imgur.com/qF1gfQ8.png) * for The arm to be in static equilibrium summation of forces on each link = 0 as well as summation of moments at a point = 0; ![](https://i.imgur.com/cRFWCEy.png) * ![](https://i.imgur.com/q2VKV3L.png) * Here we need to project forces along axes. * There are 3 ways to calculate jacobian velocity propogation, Jacobian transpose can be calculated with back propogation and expicitly calculate by analyising the structure. ### virtual work principal: * internal forces can be ignored and only active forces are taken into account. * Here ${\tau}^T\delta q$ represents dot product btw $\tau \, and \, \delta q$ ![](https://i.imgur.com/4i68r5u.png) ### Example(Static forces): ![](https://i.imgur.com/aAOj0VJ.png) ![](https://i.imgur.com/Y1beFiu.png)