Section 7
Modern robots might look like single machines on the outside, but inside they are a synergy of many components working together. Just like living creatures have senses, brains, muscles, and energy sources, robots have sensors, computers/AI, actuators, and power systems.
In this section, we'll peel back the cover to see what makes a modern robot tick – and how these parts interact to give robots mobility and autonomy.
Eyes and Ears
The Brain
The Muscles
Energy Source
Sensors are the devices that allow a robot to perceive its environment and its own state. Different sensors play different roles:
Cameras (RGB, depth, stereo) give robots vision to recognize objects. For example, Spot uses multiple stereo cameras for 3D vision. LiDAR scanners sweep lasers to measure depth, helping humanoids like Atlas navigate.
Act as the robot's inner ear, measuring acceleration and rotation to help maintain balance. Most mobile robots carry an IMU for stabilizing movement - Spot uses one to stay balanced while trotting over rough terrain.
Provide feedback from joints and limbs. Include encoders on motors (measuring angles) and force-torque sensors (measuring loads). In Spot's legs, these constantly report if a foot has hit the ground.
Ultrasonic rangefinders (like bat echolocation), radar (for autonomous cars), GPS (outdoor location), and specialized sensors like thermal cameras or chemical sensors, depending on the mission.
All these sensors act as the robot's "eyes, ears, and skin," converting real-world signals into data. This raw data feeds into the robot's computing system, which must interpret what the sensors are telling it.
Having sensor data is useless without a brain to make sense of it. The computing system is effectively the robot's brain, often running sophisticated AI algorithms on processors ranging from microcontrollers to powerful CPUs and GPUs for heavy AI tasks.
Crucially, robots perform edge computing – processing happens on the robot itself rather than relying on distant servers. This is vital because robots need to react in real-time without waiting for cloud instructions.
Turning sensor data into useful knowledge:
Example: Spot's processors perform SLAM (Simultaneous Localization and Mapping) to navigate buildings autonomously.
Deciding and executing actions:
Control loops run constantly, adjusting motor outputs based on sensor feedback within milliseconds.
Modern robots use advanced processors and neural network accelerators, often running middleware like ROS. The tight sensor-computer-actuator loop lets robots react instantly – from autonomous cars avoiding pedestrians to drones stabilizing in wind.
If sensors and computers are like a robot's senses and brain, actuators are the muscles. Actuators create motion, most commonly through electric motors that convert electrical energy into rotational motion for wheels, joints, or gears.
Use 2-4 motors to turn wheels. Each motor is precisely controlled for navigation.
Each joint has a servo motor with gearing and feedback control. A 6-axis arm has 6 motor-driven joints.
Boston Dynamics Spot: 12 electric motors (3 per leg) act like hips and knees for walking and climbing.
Atlas humanoid: 28 hydraulic actuators provide power for jumping and backflips.
Efficiency & Fine Control
Perfect for precise movements
Raw Power & Strength
Ideal for heavy lifting
Actuators work with power electronics (motor drivers) that take low-power computer commands and output high-power currents. Sensors like encoders provide feedback, ensuring the "muscles" move precisely as commanded.
All this sensing, thinking, and moving requires power. The power system is like the robot's circulatory system, providing energy to all components. Most modern mobile robots rely on lithium-ion batteries due to their high energy density.
90 min
605 Wh battery capacity
20-30 min
Flying consumes lots of power
8-10 hrs
Large batteries, efficient motors
• Voltage Regulators: Distribute correct voltages
• Inverters: Convert DC to AC power
• Battery Management: Monitor health & balance
• Power Controllers: Handle large currents safely
The Balance: Bigger batteries = longer runtime but heavier robot. Engineers optimize efficiency through low-power sensors, efficient motors, and smart software. Some robots feature swappable batteries for continuous operation.
Each component – sensors, computing, actuators, and power – is important, but it's the integration and real-time coordination that truly gives a robot life. Here's how they interact:
Sensors continuously gather data. Example: A self-driving car's cameras see traffic lights, LiDAR scans distances, radar detects vehicle speeds, GPS/IMU pinpoints location.
Data streams to onboard computers. Perception software identifies red lights and pedestrians. Planning algorithms decide to slow down and stop, following rules and AI-learned behaviors.
Plans convert to low-level actuator commands. The computer signals brake actuators with precise timing. Electric current flows from battery through motor controllers to brake motors.
Sensors report speed changes. Control systems verify deceleration and adjust if needed (e.g., slippery roads). This loop continues many times per second while the power system supplies energy.
This loop runs continuously
Trade-offs: Electric motors (Spot) = efficiency & precision | Hydraulics (Atlas) = raw power for acrobatics
Modern robots are a fusion of smart sensors, intelligent computing, powerful actuators, and robust power systems.
A fault in any component brings the whole robot down:
Without sensors
a robot is blind
Without computing
it's dumb
Without actuators
it can't move
Without power
it's dead
But with all components orchestrated in harmony, you get a machine that can drive, walk, fly, perceive, and make decisions autonomously – a robot that truly comes "alive" under the hood.
Ready to explore how robots are programmed?
Click "Next" below to continue to Section 8.