Introduction to Robotics & Automation
What robots are and how automation works. Learn the history and definitions of robots, the anatomy of a robot (links, joints, actuators, sensors, controller, end effector), joint types and degrees of freedom (Grübler–Kutzbach mobility), robot classes (articulated, SCARA, Cartesian/gantry, cylindrical, delta/parallel, collaborative, mobile, humanoid, drones), key specifications (payload, reach, repeatability vs accuracy, speed), workspace, fixed vs programmable vs flexible automation, and the robotics software stack — with Python for mobility, a 2R arm's workspace and payload torque.
A robot is a programmable, multi-functional machine that senses, decides and acts in the physical world. Industrial arms weld, paint and assemble cars; SCARA and delta robots pick electronics and food at high speed; mobile robots move goods in warehouses; cobots work beside people.
Robotics combines mechanical design, electronics, control, computer science and AI. Automation more broadly uses machines, PLCs and software to run processes with minimal human intervention.
An analogy
A robot arm is like a human arm: joints are the shoulder, elbow and wrist, motors are the muscles, encoders and cameras are the senses, the controller is the brain, and the gripper is the hand.
Anatomy of a robot
Every robot has:
- Mechanical structure — links and joints (revolute R, prismatic P, spherical)
- Actuators — electric servomotors with gearboxes (harmonic, cycloidal), pneumatics, hydraulics
- Sensors — encoders (proprioceptive), cameras, lidar, force/torque, proximity (exteroceptive)
- Controller — computes trajectories and closes servo loops (1–10 kHz)
- End effector — gripper, vacuum cup, welding torch, spindle, tool changer
Robot types
Articulated 6-axis arms (general purpose: welding, handling); SCARA (4-axis, fast planar assembly); Cartesian/gantry (large workspaces, CNC-like); cylindrical and polar (legacy); delta/parallel robots (very high-speed picking, stiff, small workspace); collaborative robots (force-limited, safe near people); autonomous mobile robots (AMRs) and AGVs; legged and humanoid robots; aerial and underwater robots; surgical robots.
Specifications that matter
Payload (including gripper), reach, number of axes, repeatability (how closely it returns to a taught point, typically ±0.02–0.1 mm) vs absolute accuracy (how closely it reaches a computed point, often ±0.5–1 mm unless calibrated), maximum joint speeds, cycle time, IP rating, mounting, controller features and safety functions. Choose robots by payload at the required reach and speed, not just nominal payload.
Automation and the robotics stack
Fixed (hard) automation — dedicated machines for high volume; programmable — robots and CNC for batches; flexible — quick changeover and mixed models. The robotics software stack: drivers and real-time control, kinematics and dynamics, motion planning, perception, task planning and HMI — often built on ROS 2. Trends: cobots, AI-based picking, digital twins and simulation (Isaac Sim, Gazebo), humanoids and robot foundation models.
Formulas
- Spatial mobility (Grübler–Kutzbach):
M = 6(n − 1 − j) + Σfᵢ - Planar mobility:
M = 3(n − 1 − j) + Σfᵢ - 2R workspace:
|L₁ − L₂| ≤ r ≤ L₁ + L₂ - Static payload torque:
τ = m·g·r - Reach of 2R at elbow angle q₂:
r = √(L₁² + L₂² + 2L₁L₂cos q₂)
Worked example: Mobility of a SCARA
Given: 4 joints (R, R, P, R), 5 links incl. base
Find: DOF.
- 6(5 − 1 − 4) + 4
Answer: 4 DOF (x, y, z, rotation about z)
Worked example: Workspace
Given: L₁ = 0.5 m, L₂ = 0.5 m
Find: Reach range.
- |0.5 − 0.5| to 1.0
Answer: 0 to 1.0 m (full disc)
Worked example: Payload torque
Given: 3 kg at 0.6 m
Find: Shoulder torque.
- 3 × 9.81 × 0.6
Answer: 17.7 N·m (static)
Mobility, workspace and payload torque
import math
# Robot basics: mobility (Grubler-Kutzbach) and the reach of a 2-link planar arm
def grubler(links, joints): # spatial: M = 6(n - 1 - j) + sum(freedoms), revolute joints give 1 each
return 6 * (links - 1 - len(joints)) + sum(joints)
print(f"6-axis serial arm: {grubler(7, [1]*6)} DOF")
print(f"planar 4-bar linkage: {3*(4-1-4) + 4} DOF (planar formula 3(n-1-j) + sum f)")
L1, L2 = 0.40, 0.30 # m
print(f"2R arm workspace: annulus from {abs(L1-L2):.2f} m to {L1+L2:.2f} m, area {math.pi*((L1+L2)**2-(L1-L2)**2):.3f} m2")
# Payload moment check at full reach: 5 kg payload
print(f"static torque at shoulder with 5 kg at {L1+L2:.2f} m: {5*9.81*(L1+L2):.1f} N*m")
# 6-axis serial arm: 6 DOF
# planar 4-bar linkage: 1 DOF (planar formula 3(n-1-j) + sum f)
# 2R arm workspace: annulus from 0.10 m to 0.70 m, area 1.508 m2
# static torque at shoulder with 5 kg at 0.70 m: 34.3 N*mSix independent joints give six DOF — enough to place a tool at any position and orientation in space (3 + 3). A four-bar linkage has one DOF: a single input angle determines everything. The 2R arm can reach anywhere in an annulus from 0.1 to 0.7 m, and holding just 5 kg at full reach already needs 34 N·m at the shoulder — before the arm's own weight and accelerations — which is why arms are rated with reduced payload at full extension.
Common mistakes
- Wrong: Choosing a robot by nominal payload only. — Right: Check the payload–reach–inertia diagram including the gripper.
- Wrong: Confusing repeatability with accuracy. — Right: Repeatability is return-to-taught-point; accuracy needs calibration.
- Wrong: Automating a bad process. — Right: Simplify and standardise the process first.
Applications
- Automotive: Spot welding, painting, body assembly.
- Electronics: SCARA and delta pick-and-place.
- Logistics: AMRs, palletising, parcel sorting.
- Healthcare: Surgical and rehabilitation robots.
Key takeaways
- Links, joints, actuators, sensors, controller, end effector.
- Mobility from Grübler–Kutzbach; 6 DOF for full pose.
- Choose by payload at reach, repeatability, speed.
- Automation ranges from fixed to flexible.
Going further
- Redundant robots: 7-DOF arms for obstacle avoidance.
- Parallel kinematics: Delta and Stewart platforms.
- Robot foundation models: Generalist policies from large datasets.