Walking, Rolling, Flying, Swimming
Nature solved the locomotion problem in dozens of ways: fish undulate, birds flap, cheetahs sprint on four legs, humans balance on two. Robotics researchers have spent decades borrowing each of these solutions and engineering them into machines. The result is a rich family of locomotion types, each with its own strengths and weaknesses, each suited to a different environment and task.
Rolling: Fast, Efficient, Flat-World Friendly
Wheels are the simplest and most energy-efficient way to move a robot across a smooth, flat surface. A wheeled robot converts motor rotation directly into forward motion with minimal mechanical complexity. There are no impacts, no balance problems on smooth terrain, and the engineering is well understood after more than a century of automobile development. Differential drive is the most common wheeled configuration for simple robots: two driven wheels on opposite sides, steered by spinning them at different speeds. Turn the left wheel faster than the right and the robot turns right. Spin them in opposite directions and the robot pivots in place. The Mars Exploration Rovers Spirit and Opportunity used a rocker-bogie six-wheel system — no steering mechanism at all, just differential speeds across six wheels — to traverse rocky Martian terrain for years. The limitation of wheels is terrain. Wheels work brilliantly on pavement and floors but struggle with steps, loose sand, and rough rubble. Even the best wheeled robot cannot climb a standard staircase.
Differential drive steers by varying the relative speed of left and right wheels. No separate steering mechanism is needed. It is mechanically simple and very common in classroom robots, warehouse robots, and Mars rovers.
Walking: Versatile but Complex
Legged robots can step over obstacles, climb stairs, walk on loose terrain, and move through environments designed for the human body. This versatility comes at a steep engineering cost: each leg needs multiple actuated joints, every step involves managing balance and impact forces, and the software to coordinate gait across four or six or two legs is enormously complex. A gait is the pattern of leg movements that produces forward motion. A quadruped — a four-legged robot like Boston Dynamics Spot — can use a walk gait (always at least two legs on the ground), a trot (diagonal pairs moving together), or a bound (both front legs then both back legs). Each gait has different stability, speed, and power characteristics. Bipedal robots, which walk on two legs like humans, face the additional challenge of dynamic balance: a person standing still on two feet is already balancing, not statically resting. We will explore this challenge in detail in Lesson 7.
Flying: Freedom in Three Dimensions
Aerial robots — drones — escape the ground entirely. A multirotor drone (the familiar quadcopter shape) uses four or more motor-driven propellers arranged symmetrically. By varying the speed of each rotor, the flight controller can produce thrust, tilt, and yaw independently. The resulting aircraft can hover motionless, fly in any direction, and navigate complex three-dimensional spaces. The flight controller on a quadrotor runs a closed-loop PID controller many hundreds of times per second, adjusting each rotor's speed to maintain the commanded attitude. Without this fast feedback, a multirotor is aerodynamically unstable and would flip immediately. The control loop is what makes hovering look easy. Flight comes with its own constraints: battery energy is consumed rapidly fighting gravity, payload capacity is limited, and weather — especially wind — creates constant disturbances the controller must fight.
A quadrotor tilts forward by speeding up the rear rotors and slowing the front rotors, shifting the net thrust vector. It yaws (rotates on its vertical axis) by spinning diagonal pairs faster or slower. All four motors are adjusted simultaneously, hundreds of times per second.
Swimming: Navigating a Dense Medium
Water is 800 times denser than air. A robot moving through water faces far greater drag than one moving through air, but it also gets buoyancy for free — a properly designed underwater robot can hover at depth without fighting gravity. Underwater robots are used for ocean pipeline inspection, coral reef surveys, and recovering objects from sunken ships. AUVs — Autonomous Underwater Vehicles — typically use thrusters (underwater propellers) to maneuver and pressure-rated housings to protect electronics from the crushing depths. ROVs — Remotely Operated Vehicles — are tethered to a surface ship and human-piloted. Fish-inspired robots use undulating body movements or oscillating fins, mimicking the energy efficiency of real fish at low speeds. Communication is the biggest challenge underwater: radio waves do not penetrate water. Underwater robots communicate via acoustic signals (sound), which are slow and low-bandwidth compared to the radio links a ground or aerial robot would use.
Match each locomotion mode to its defining characteristic or best-suited environment.
Terms
Definitions
Drag terms onto their definitions, or click a term then click a definition to match.
Why does a multirotor drone require a fast, continuously running closed-loop controller to fly?
Which locomotion mode is most suitable for a robot inspecting the hull of an oil tanker 20 meters below the ocean surface?
Locomotion Mode Design Challenge
- Step 1: Read each deployment scenario and choose the best locomotion mode: wheeled, legged, aerial (multirotor), or underwater.
- A) Mapping the interior corridors of a large warehouse 24 hours a day.
- B) Searching a collapsed building for survivors in debris-filled rooms with rubble, broken stairs, and uneven floors.
- C) Filming a sports event from above a crowd without a fixed camera mount.
- D) Surveying coral reef health at 15 meters depth in a tropical ocean.
- Step 2: For each choice, write one sentence explaining why your selected mode fits and one sentence explaining what would make it fail.
- Step 3: For scenario B, describe one specific engineering challenge a legged robot would face that a wheeled robot could not handle.