Forward Kinematics of Manipulators for GATE RA 2027: DH Parameters, Transformation Matrices & Solved Problems

Forward kinematics answers the most fundamental question in robot manipulation: given the joint variables — the angles of the revolute joints and the extensions of the prismatic joints — where in space does the end-effector sit, and how is it oriented? It is a guaranteed scoring topic in the “Principles of Robotics and Automation” line of the GATE Robotics and Automation (RA) 2027 syllabus, because every problem has a single, deterministic answer that you can grind out with rotation matrices, homogeneous transformations and the Denavit–Hartenberg (DH) convention. This guide builds the theory from the ground up — frames, transformation matrices, the DH parameters, position and velocity kinematics of planar and articulated arms — and then works through five GATE-style numerical problems end to end, with labelled diagrams, formula boxes and the exam pitfalls to avoid.

What is forward kinematics?

A serial manipulator is an open chain of rigid links connected by joints. Each joint contributes one degree of freedom: a revolute joint contributes an angle θ_i, and a prismatic joint contributes a linear displacement d_i. Collect these joint variables into a vector q = (q₁, q₂, …, q_n). Forward kinematics is the map from joint space to task space: given q, it returns the position and orientation (the pose) of the end-effector expressed in the fixed base frame.

The defining feature of forward kinematics — and the reason it is such a reliable source of marks — is that the map is unique. For any set of joint values there is exactly one end-effector pose, so there is never any ambiguity to resolve. (The reverse problem, inverse kinematics, can have multiple solutions or none, which is a separate topic.) Forward kinematics splits naturally into position kinematics — where the tool point ends up — and velocity kinematics — how fast it moves for given joint rates, handled through the Jacobian.

Frames, rotation matrices & homogeneous transforms

Attach a coordinate frame to each link. The pose of one frame relative to another is captured by a homogeneous transformation matrix — a 4×4 matrix that packs a 3×3 rotation matrix R and a 3×1 position vector P together, so that a single matrix multiplication carries a point from one frame into another:

The rotation part obeys a crucial rule you will use in almost every problem: the columns of R are the unit vectors of frame B’s x, y and z axes, written in frame A’s coordinates. So if you ever need the direction of frame B’s z-axis as seen from the base, just read off the third column of R. The elementary rotations about the principal axes are worth memorising:

The full forward-kinematics chain is then just the ordered product of the per-joint transforms, from base (0) to tool (n):

The Denavit–Hartenberg convention

The Denavit–Hartenberg (DH) convention is a systematic recipe that describes each link–joint pair with just four parameters instead of a full six-degree pose, by cleverly choosing where the link frames sit. The four DH parameters are the link length a_i (distance between successive joint axes along the common normal), the link twist α_i (angle between the axes), the link offset d_i (distance along the joint axis), and the joint angle θ_i (rotation about the joint axis). For a revolute joint θ_i is the variable; for a prismatic joint d_i is the variable.

Each single-link transform is the product of a screw along the previous axis and a screw along the common normal:

Building the DH table for a manipulator and multiplying the link transforms in order is the general, machine-crackable route to forward kinematics of any serial arm. For the low-DOF planar and articulated arms that dominate GATE numericals, you can usually write the position equations by direct geometry (projection onto axes), which is faster — but the DH product always gives the same answer and also delivers the orientation.

Forward position kinematics of a planar 2R arm

The planar two-revolute (2R) arm is the canonical worked example. Two links of length L₁ and L₂ rotate in a plane by angles θ₁ (measured from the x-axis) and θ₂ (measured relative to the first link). Projecting each link onto the x and y axes and adding gives the end-effector position directly:

The orientation of the end link is simply φ = θ₁ + θ₂. Notice how the angles accumulate down the chain: the second link’s absolute orientation is the sum of both joint angles. That accumulation is the single most important idea in serial-arm kinematics and it generalises directly to three-link and spatial arms.

Forward velocity kinematics & the Jacobian

Differentiating the position equations with respect to time gives the velocity kinematics — how fast the end-effector moves for given joint rates. For the planar 2R arm, applying the chain rule (and remembering that the derivative of the (θ₁+θ₂) term carries the sum of both joint rates) yields:

In matrix form this is v = J(q)·q̇, where J is the manipulator Jacobian — the matrix of partial derivatives ∂(x,y)/∂(θ₁,θ₂). The Jacobian is central to robotics: it maps joint rates to Cartesian velocities, it maps joint torques to end-effector forces (via τ = J^TF), and the configurations where it loses rank are the singularities where the arm loses a direction of motion. The magnitude of the tool velocity is then |v| = √(ẋ² + ẏ²).

Articulated (RRR) spatial manipulator

A three-revolute articulated (RRR) arm — the classic industrial “anthropomorphic” robot — has a vertical waist joint θ₁ that swings the whole arm about the base axis, a shoulder joint θ₂, and an elbow joint θ₃. The neat trick for its forward kinematics is to work in the vertical plane of the arm first: resolve the reach into a height z above the base and a radial distance r from the waist axis, then let the waist angle rotate that (r, z) point around.

With the shoulder angle θ₂ measured from the horizontal and the elbow angle θ₃ measured relative to the upper arm, the height and radial reach are:

The straight-line distance of the end-effector from the base origin is d = √(r² + z²), and — importantly — it does not depend on the waist angle θ₁, because rotating about the vertical axis keeps the tool at the same radius and height. The Cartesian coordinates come from spinning r by the waist angle: x = r·cosθ₁, y = r·sinθ₁.

Worked examples (GATE-style)

Key formulas

Applications

Forward kinematics is the computational backbone of every manipulator controller. It runs continuously to report where the tool actually is so the robot can be commanded in Cartesian space; it feeds trajectory planning, where a path in task space is checked link by link; it underpins collision checking and workspace mapping, since sweeping the joint variables through their ranges traces out the reachable envelope; and it is the essential first step before inverse kinematics and Jacobian-based velocity and force control can be set up. The same 4×4 transformation machinery describes camera-to-robot calibration in machine vision, tool-centre-point programming on welding and painting robots, and the digital twins used to simulate a cell before commissioning. Mastering it is the gateway to the entire manipulator-kinematics section of the GATE RA syllabus.

Common mistakes to avoid

• Mixing relative and absolute joint angles. In the 2R and RRR arms the outer link’s orientation is the sum of the joint angles, so the second term uses (θ₁+θ₂), not θ₂ alone.
• Dropping the (θ̇₁+θ̇₂) factor in velocity. When differentiating a cos(θ₁+θ₂) term, the chain rule brings down the sum of both joint rates.
• Reading the wrong column of R for an axis direction. The z-axis of a frame is the third column of its rotation matrix, expressed in the reference frame — not a row.
• Multiplying transforms in the wrong order. Homogeneous transforms do not commute; the chain must be built base-to-tool as ⁰T₁·¹T₂·…, and R_zR_y ≠ R_yR_z.
• Measuring an angle from the wrong reference. Always check whether the shoulder angle is given from the horizontal or the vertical, and whether the elbow angle is relative to the upper arm or absolute.
• Forgetting the base offset. When transforming a point between frames, add the origin-position vector after applying the rotation — a rotation alone does not translate the point.

Frequently asked questions

What is the difference between forward and inverse kinematics?

Forward kinematics goes from joint variables to the end-effector pose and always has a unique answer. Inverse kinematics goes the other way — from a desired pose back to the joint variables — and can have multiple solutions (e.g. elbow-up and elbow-down), a unique solution, or no solution if the pose is outside the workspace.

Why do we use homogeneous transformation matrices?

A 4×4 homogeneous transform packs rotation and translation into a single matrix, so the pose of the whole arm is just the ordered product of the per-link transforms. It turns a chain of rotations and offsets into one clean matrix multiplication and automatically carries both the position and the orientation of the end-effector.

What are the four Denavit–Hartenberg parameters?

They are the link length a_i, the link twist α_i, the link offset d_i and the joint angle θ_i. For a revolute joint θ_i is the variable and the other three are fixed; for a prismatic joint d_i is the variable. The DH convention describes each link with just these four numbers instead of a full six-parameter pose.

What is the manipulator Jacobian used for?

The Jacobian J maps joint velocities to end-effector velocity through v = J·q̇, and by duality maps end-effector forces to joint torques through τ = J^TF. Configurations where J loses rank are singularities, where the arm momentarily cannot move in some Cartesian direction. It is the key tool of velocity and force control.

Does the waist (base) rotation change the reach of an articulated arm?

No. The straight-line distance of the end-effector from the base axis, d = √(r² + z²), depends only on the shoulder and elbow angles. The waist joint rotates that fixed (r, z) point around the vertical axis, changing the x and y coordinates but not the distance from the base.

This guide is part of the manipulator-kinematics track of the complete GATE RA 2027 Syllabus overview. It expands the forward-kinematics section of the Principles of Robotics and Automation guide — review that pillar for manipulators, degrees of freedom, rotation matrices and homogeneous transformations.