Track 03
Kinematics & Dynamics
Where the robot is, how fast it's moving, and what forces it takes to make it move. The mechanical spine of the field.
10 published · 0 planned · 10 lessons total
- 01→
Forward kinematics of a 2-DOF planar arm
PublishedTwo joints, two link lengths, and high-school trig. Compute where the end-effector is, plot the reachable workspace, and you've unlocked the foundation of every arm problem.
~15 min
- 02→
Rigid-body transforms: SO(3) and SE(3) demystified
PublishedRotations and poses in 3D, without trig soup. Learn SO(3), SE(3), homogeneous matrices, and quaternions — and when to use each.
~18 min
- 03→
Denavit-Hartenberg parameters (and why we're moving past them)
PublishedDH was the standard arm-kinematics formulation for 50 years. It still shows up in textbooks and FK code. Here's how it works, why it survives, and what modern textbooks (and URDFs) prefer.
~13 min
- 04→
Product of exponentials: the Modern Robotics approach
PublishedThe clean replacement for Denavit-Hartenberg. Each joint contributes a screw axis; kinematics is the product of their matrix exponentials. Same answers as DH, with dramatically less bookkeeping.
~18 min
- 05→
Inverse kinematics: analytical solutions
PublishedGiven a target (x, y), find the joint angles. Closed-form IK for a 2-DOF planar arm, the two-solution trap (elbow up vs elbow down), and when closed-form falls apart.
~14 min
- 06→
Inverse kinematics: iterative (Jacobian-based) methods
PublishedWhen closed-form IK doesn't exist (7+ DOF, redundant arms), you iterate. Damped least squares, null-space projection, and the practical recipes that make Jacobian-based IK actually converge.
~14 min
- 07→
The Jacobian and velocity kinematics
PublishedThe matrix that connects joint speeds to end-effector velocities — and joint torques to end-effector forces. Master it once and most arm-control intuition follows.
~13 min
- 08→
Lagrangian dynamics of a serial arm
PublishedFrom kinetic + potential energy to the equation of motion. M(q)q̈ + C(q,q̇)q̇ + g(q) = τ — the canonical form behind every torque-control loop and feedback-linearization scheme.
~15 min
- 09→
Mobile robot kinematics: differential, Ackermann, omnidirectional
PublishedThree wheel configurations cover almost every wheeled robot. Each has different kinematics, different motion constraints, and different planners. The full comparison in one lesson.
~12 min
- 10→
Quadrotor dynamics in 15 minutes
PublishedFour propellers, six degrees of freedom, four control inputs. Underactuated, fast, and gorgeous to control. The minimal model, the controller hierarchy, and the differential-flatness trick that makes it work.
~14 min