Electricity is the lifeblood of every robot. Motors need current to produce torque, sensors need stable voltage to generate accurate readings, and microcontrollers need regulated power to execute programs. Understanding how electrical energy flows through a circuit is essential before you can design or troubleshoot any robotic system.

The three fundamental quantities in electronics are voltage (V), current (I), and resistance (R). Voltage is the electrical “pressure” that pushes charge through a circuit, measured in Volts. Current is the rate of charge flow, measured in Amperes. Resistance is the opposition to current flow, measured in Ohms. These three quantities are governed by Ohm’s Law.

• Series Circuit: Components connected end-to-end share the same current. Voltage divides proportionally across each component: V_n = I × R_n.
• Parallel Circuit: Components connected side-by-side share the same voltage. Current divides among branches: I_n = V / R_n.
• Microcontroller: A self-contained computer on a single chip (e.g., Arduino Uno / ATmega328P) that reads sensor inputs and drives outputs through programmable digital and analog pins.
• Breadboard: A reusable prototyping board with internal spring clips that allow components to be connected without soldering, essential for rapid circuit iteration.

From Circuit to Code

Modern robotics relies on microcontrollers — small, inexpensive programmable chips that bridge the gap between sensors and actuators. An Arduino Uno provides 14 digital I/O pins and 6 analog inputs. A single line of code like analogRead(A0) reads the voltage on pin A0 (0-5V mapped to 0-1023), while analogWrite(9, 128) outputs a PWM signal at roughly 50% duty cycle on pin 9. This direct hardware access is what makes microcontrollers so powerful for robotics.

The simulator below demonstrates Ohm’s Law in action. Adjust the voltage and resistance and watch how current and power change. Notice that power scales with the square of voltage — doubling the voltage quadruples the power dissipation through the same resistor. This is why proper component selection and heat management are critical in robot electronics.

Hardware without software is just metal and plastic. Programming is what brings a robot to life — translating sensor data into motor commands through a logical sequence of decisions. Robot programming combines real-time data processing with control theory, often running on resource-constrained hardware where every microsecond and byte of memory counts.

Most beginner robotics platforms use C/C++ (Arduino) or Python (Raspberry Pi). The core programming loop reads inputs, computes outputs, and writes those outputs to actuators — repeating this cycle hundreds or thousands of times per second. Within this loop, the programmer must handle sensor noise, implement state machines for complex behaviors, and manage timing constraints.

• PWM (Pulse Width Modulation): A technique for simulating variable voltage by rapidly switching a digital signal between HIGH and LOW. The duty cycle (percentage of time HIGH) determines the effective average voltage delivered to a load like a DC motor.
• State Machine: A programming pattern where the robot exists in one of several discrete states (e.g., “searching,” “approaching,” “grabbing”) and transitions between them based on sensor inputs.
• H-Bridge: A circuit of four switching elements (transistors or MOSFETs) that allows bidirectional current flow through a DC motor, enabling both forward and reverse rotation from a single supply.
• Sensor Noise: Random fluctuations in sensor readings caused by electrical interference or environmental factors. Filtering (moving average, exponential smoothing, Kalman filter) is needed to extract reliable data.

PWM and Motor Speed Control

A DC motor’s speed is roughly proportional to the voltage applied to it. However, microcontrollers output digital signals (5V HIGH / 0V LOW), not true analog voltages. PWM bridges this gap: by switching the signal HIGH for a fraction of each cycle at a frequency high enough (typically 490 Hz or 980 Hz on Arduino), the motor’s inertia smooths the pulsed signal into a proportional average speed. The simulator below lets you change the duty cycle and supply voltage to see how the PWM waveform and effective voltage respond.

While wheeled robots move by steering along surfaces, articulated robots — robotic arms — must compute precise joint angles to position their end effectors in space. This is the domain of kinematics: the mathematics of motion geometry without regard to the forces that cause it. Kinematics bridges a robot’s mechanical structure and its software control.

A robotic arm is a kinematic chain — a series of rigid links connected by joints. Each revolute (rotating) or prismatic (sliding) joint adds a degree of freedom. A 2-link planar arm confined to a flat plane is the simplest non-trivial case and serves as the foundation for understanding complex industrial manipulators with 6 or more DOF.

• Forward Kinematics (FK): Computing the end-effector position from known joint angles. Uses straightforward trigonometry and always yields exactly one solution.
• Inverse Kinematics (IK): Computing the joint angles needed to reach a desired end-effector position. More complex — may have zero, one, or multiple solutions.
• Workspace: The set of all points the end effector can reach. For a 2-link arm, this is an annular (ring-shaped) region with inner radius |L1 – L2| and outer radius L1 + L2.
• Singularity: A configuration where the arm loses a degree of freedom (e.g., fully stretched out). Near singularities, small changes in end-effector position require disproportionately large joint movements.

Elbow-Up vs Elbow-Down

The inverse kinematics of a 2-link arm generally yields two valid solutions — the “elbow-up” and “elbow-down” configurations. Both place the end effector at the same target position, but the arm’s shape differs. Choosing between them depends on workspace obstacles, joint limits, and the desired motion trajectory. Industrial robots often use elbow-down for tasks below the arm and elbow-up for overhead work. The simulator below computes the elbow-down solution. Try moving the target outside the dashed reach boundary to see what happens when a point is unreachable.

The ultimate goal of robotics is autonomy — the ability for a robot to operate independently, making decisions and adjusting its behavior in real time without human intervention. At the heart of most autonomous behaviors is a feedback control loop: the robot continuously measures its state, compares it to a desired state, and applies corrections to minimize the difference. This is the principle of closed-loop control.

The PID (Proportional-Integral-Derivative) controller is the most widely used feedback control algorithm in robotics and industrial automation. Its elegance lies in its simplicity: three tuning parameters (Kp, Ki, Kd) that can be adjusted to handle a remarkable range of control challenges, from maintaining a drone’s hover altitude to guiding a mobile robot along a curved line.

• Proportional (Kp): Produces correction proportional to the current error. Higher Kp means faster response but can cause overshoot and oscillation around the setpoint.
• Integral (Ki): Accumulates past errors over time. Eliminates steady-state offset that P-only control cannot correct, but excessive Ki causes integral windup and instability.
• Derivative (Kd): Reacts to the rate of error change. Acts as a predictive dampener that reduces overshoot, but amplifies high-frequency sensor noise.
• Steady-State Error: The persistent offset remaining after a system settles, which a P-only controller cannot eliminate for certain plant types (e.g., velocity control with constant disturbance).

Tuning a PID Controller

PID tuning — finding the right balance of Kp, Ki, and Kd — is both a science and an art. The Ziegler-Nichols method provides a systematic starting point: increase Kp until the system oscillates at constant amplitude (the “ultimate gain” Ku with period Tu), then compute initial gains from published formulas. From there, manual fine-tuning adjusts the response to meet specific settling-time, overshoot, and stability requirements.

For a line-following robot, the process is intuitive: start with Kp only and increase until the robot tracks the line but oscillates. Add Kd to dampen oscillation. Finally, add a small Ki to correct any persistent bias from sensor miscalibration or uneven wheel diameters. The simulator below lets you experience this tuning process directly — adjust each gain and observe its effect on the robot’s trajectory.

This capstone module brings together every concept from the course — sensing, electronics, programming, kinematics, and control — into a single integrated challenge: autonomous navigation. A mobile robot must reach a goal position in an arena populated with obstacles, using sensor-based perception and intelligent decision-making to plan and execute a safe path.

Real-world autonomous navigation involves multiple layers of intelligence. At the lowest level, reactive behaviors (like obstacle avoidance) provide immediate safety. At a higher level, path planning algorithms compute efficient routes. This capstone focuses on the foundational layer: reactive navigation using potential fields, where the goal exerts an attractive force and obstacles exert repulsive forces, and the robot follows the combined vector.

1. Environment Sensing: The robot’s sensors (ultrasonic, LiDAR, or IR) detect nearby obstacles. In this simulation, obstacles are rectangular regions that generate repulsive force fields.
2. Goal Attraction: The target position exerts an attractive force proportional to distance, pulling the robot toward it. The force magnitude is capped to prevent runaway speed.
3. Obstacle Repulsion: Each obstacle generates a repulsive force when the robot is within a threshold distance. The force increases sharply as the robot approaches.
4. Navigation Execution: The robot’s heading is determined by the vector sum of all forces. By continuously recomputing these forces as it moves, the robot navigates around obstacles while progressing toward the goal.

• Potential Field Method: A navigation approach where the goal attracts and obstacles repel, creating a virtual “force landscape” the robot flows through.
• Local Minima: A limitation where balanced attractive and repulsive forces trap the robot in a non-goal position. Advanced methods (navigation functions, RRT*) address this.
• Sensor Fusion: Combining data from multiple sensor types (wheel encoders + IMU + LiDAR) to produce a more accurate state estimate than any single sensor could provide.