The Inner Workings of Robots

At their core, humans consist of five fundamental components:

Naturally, humans possess intangible qualities like intelligence and ethics, but physically, the aforementioned components encapsulate our makeup.

Robots are constructed using similar elements. A standard robot features a movable frame, a motor, a sensory apparatus, an energy source, and a computational unit that governs these parts. In essence, robots are mechanical imitations of living organisms, designed to mimic human and animal actions.

Joseph Engelberger, a trailblazer in industrial robotics, famously said, "I can't define a robot, but I recognize one when I see it!" Given the variety of machines labeled as robots, crafting an all-encompassing definition is challenging. People's perceptions of what qualifies as a robot vary widely.

You're likely familiar with these renowned robots:

All these machines are regarded as robots by some, though many define a robot simply as something they perceive to be one. Roboticists, however, offer a more technical definition. They emphasize that robots must have a reprogrammable computer brain capable of controlling a physical body.

This definition sets robots apart from other mobile machines like trucks, which rely on direct human control despite advanced onboard electronics. Unlike standard computers, robots are unique in their integration of a physical form with computational intelligence.

In the following section, we’ll explore the key components commonly found in modern robots.

Robot Basics

The majority of robots feature movable bodies. While some are equipped with motorized wheels, others boast numerous articulated sections, often constructed from metal or plastic. These segments are linked by joints, much like the skeletal structure in humans.

Robots employ actuators to rotate wheels and maneuver jointed parts. Actuators can include electric motors, solenoids, hydraulic systems, or pneumatic mechanisms powered by compressed gases. Many robots integrate a mix of these actuator technologies.

To power these actuators, robots require an energy source. Common options include batteries, wall outlets, solar panels, or fuel cells. Hydraulic robots also rely on pumps to pressurize fluid, while pneumatic robots need air compressors or compressed-air reservoirs.

Actuators are connected to electrical circuits, which supply power to motors and solenoids or control hydraulic systems via valves. These valves direct pressurized fluid through the machine. For instance, to move a hydraulic leg, the robot's controller opens a valve, allowing fluid to flow into a piston cylinder, extending the piston and moving the leg. Robots often use dual-action pistons to enable movement in both directions.

The robot's central computer manages all components connected to its circuits. To initiate movement, the computer activates the required motors and valves. Many robots are reprogrammable, allowing their behavior to be altered by updating or modifying the software that dictates their actions.

While not all robots are equipped with sensory capabilities, some possess the ability to see, hear, smell, or taste. The most prevalent robotic sense is motion detection, enabling the robot to track its own movement. This can be achieved using a laser beneath the robot to illuminate the floor, with a camera measuring distance and speed. Similar technology is found in computer mice. Roomba vacuums, for example, utilize infrared light to identify obstacles and photoelectric cells to detect light variations.

These elements form the foundation of robotics. By combining them in countless ways, roboticists can design robots of extraordinary complexity.

The Robotic Arm

The word robot originates from the Czech term robota, which means 'forced labor.' This aptly describes most robots, as they are primarily built for repetitive, heavy-duty tasks in manufacturing. They take on jobs that are hazardous, monotonous, or challenging for humans.

In manufacturing, robotic arms are widely utilized. These arms typically consist of seven metal sections connected by six joints. A computer controls the arm by precisely rotating stepper motors attached to each joint (some larger arms use hydraulics or pneumatics). Stepper motors move in exact increments, enabling the arm to perform highly accurate, repetitive motions. Motion sensors ensure the arm moves precisely as intended.

A six-jointed industrial robot closely mimics a human arm, featuring equivalents of a shoulder, elbow, and wrist. Usually, the shoulder is fixed to a stationary base. This design provides six degrees of freedom, allowing the robot to pivot in six distinct ways, compared to a human arm's seven degrees of freedom.

Just as your arm moves your hand, a robotic arm transports an end effector to various locations. End effectors are specialized tools attached to the arm for specific tasks. Common examples include simplified hands capable of gripping and transporting objects, often equipped with pressure sensors to regulate grip strength and prevent damage. Other end effectors include blowtorches, drills, and spray painters.

Industrial robots are programmed to perform repetitive tasks. For instance, a robot might secure lids on peanut butter jars moving along an assembly line. A programmer uses a handheld controller to guide the robot through the required motions, which the robot then memorizes and repeats for each new unit.

Many industrial robots are employed in automobile assembly lines, where they assemble vehicles with unmatched precision. Unlike humans, robots maintain consistent accuracy, drilling in the exact same spot and applying uniform force to bolts, regardless of how long they’ve been operating. They are also crucial in the computer industry, where their precision is essential for assembling microchips.

Robots are increasingly collaborating with construction workers, plastering walls with greater speed and accuracy than humans. They aid in underwater exploration, assist surgeons in performing delicate procedures, and even flip burgers in kitchens. These tasks are often accomplished using robotic arms.

Robotic arms play a vital role in space exploration. NASA utilizes a seven-degree-of-freedom arm, similar to a human arm, to capture equipment or asteroids. The Perseverance rover’s 7-foot (2-meter) robotic arm is equipped with specialized tools, including a camera for visual guidance, an abrading tool for grinding rock samples, and a coring drill to collect samples stored in metal tubes for future return to Earth. The arm also features an X-ray device called PIXL, which uses a hexapod to adjust angles for optimal analysis.

The SHERLOC instrument identifies minerals by analyzing light scattering, while the WATSON device captures detailed images. Together, these tools help scientists create a comprehensive mineral map of Mars’ surface.

Mobile Robots

Building and programming robotic arms is relatively straightforward since they operate within a limited space. However, challenges arise when robots are deployed into open environments.

First, a robot requires an effective locomotion system. For smooth surfaces, wheels are typically the optimal choice. Wheels and tracks can also handle rougher terrains, but designers often prefer legs due to their adaptability. Developing legged robots also aids researchers in studying natural movement, making it a valuable endeavor in biological research.

Robot legs are usually powered by hydraulic or pneumatic pistons, which connect to various leg segments similar to how muscles attach to bones. Coordinating these pistons is a complex task. Just as a baby learns to walk by mastering muscle coordination, a robot designer must determine the correct piston movements and program them into the robot's computer. Many mobile robots include a balance system, such as gyroscopes, to help the computer adjust movements and maintain stability.

Robotic locomotion often draws inspiration from the animal kingdom. Six-legged insects, known for their balance and adaptability, inspire designs for rugged terrains. Four-legged robots, like Boston Dynamics' Spot, resemble dogs and are often compared to them as they tackle hazardous tasks such as construction inspections. Two-legged robots pose significant balance challenges, but advancements have been made, with robots like Boston Dynamics' Atlas even performing parkour.

Aerial robots also take cues from nature. While many use airplane-like wings, researchers have developed soft actuators mimicking fly wings. Propeller-powered drones, now widely recognized, capture stunning footage for entertainment, sports, and surveillance. Some drones can be networked to form swarms, as demonstrated at the Tokyo Summer Olympic Games in 2021.

Underwater robots may walk along the ocean floor. For instance, Silver 2, a crab-like robot, is designed to locate and remove plastic waste. The Benthic Rover II uses treads instead. Snake robots, inspired by their animal counterparts, function both underwater and on land. They are even effective in medical applications, performing surgical repairs within the human body.

Some mobile robots are operated remotely, with humans dictating their actions via attached wires, radio waves, or infrared signals. These robots are ideal for exploring hazardous or hard-to-reach locations, such as deep oceans or volcanic interiors. Others are semi-autonomous, where the operator sets a destination, and the robot navigates independently.

Autonomous Robots

Autonomous robots operate independently, programmed by humans to react to external stimuli. A simple example is the bump-and-go robot, which demonstrates this principle effectively.

This type of robot is equipped with a bumper sensor to detect obstacles. When activated, it moves straight until it hits an obstacle, triggering the sensor. The robot's programming instructs it to reverse, turn right, and proceed forward after each collision, allowing it to navigate around barriers.

Certain autonomous robots function only in familiar, controlled environments. For instance, lawn-mowing robots rely on buried boundary markers to define their work area. Office-cleaning robots may require a building map to navigate efficiently. Amazon's warehouse robots use colored magnetic tape on floors for guidance, enabling them to transport items while employees focus on order packaging.

Mobile robots frequently employ infrared or ultrasound sensors to detect obstacles, similar to animal echolocation. They emit sound signals or infrared beams and measure the reflection time to determine obstacle distances. Advanced robots may use Light Detection and Ranging (lidar) technology, which relies on light to help the robot determine its position in its surroundings.

Even basic robotic vacuums use multiple methods to navigate. Alongside bump sensors, they feature cliff sensors (to detect drops), wall sensors (to identify obstacles ahead), and optical encoders (to measure distance traveled). This multi-sensor approach to mapping is known as simultaneous localization and mapping (SLAM).

Some robots utilize stereo vision to perceive their surroundings. Dual cameras provide depth perception, while image-recognition software enables them to identify and categorize objects. Robots may also employ microphones and smell sensors to assess their environment. Boston Dynamics' Spot, a dog-like robot, features a 360-degree panoramic camera, along with pan-tilt-zoom and infrared radiometric cameras. This technology allowed the U.S. Marines to test the robot's ability to look around corners and detect threats before moving into open areas.

Advanced robots can analyze and adapt to unfamiliar or rugged terrains. They associate specific terrain patterns with appropriate actions. For instance, a rover robot might create a map of the terrain ahead using visual sensors. If the map indicates rough terrain, the robot chooses an alternative path. NASA's Perseverance rover exemplifies this capability.

Follower robots learn by observing humans. Autonomous farming robot manufacturer Burro combines cameras and GPS for navigation, while its AI system learns tasks by following humans. Piaggio Fast Forward's Gita robots follow their human leaders, carrying items and even keeping pace with bicycles at speeds up to 35 miles per hour (56 kilometers per hour).

Home-made Robots

In previous sections, we explored major robotics fields like industrial and research robotics. While professionals in these areas have driven most advancements, they aren't the sole creators of robots. For decades, a dedicated group of hobbyists has been building robots in garages and basements worldwide.

The homebrew robotics community is a fast-growing subculture with a strong online presence. Amateur roboticists often build their creations using spare parts like old toys, VCRs, and other discarded gadgets. The maker movement has simplified access to components, idea-sharing, and DIY electronics education.

Robots are just one aspect of the maker movement, but many DIY tools have versatile applications. Affordable single-board computers now support complex projects. Platforms like Instructables and Thingiverse enable makers to exchange plans. Makerspaces, hackerspaces, and fablabs in schools, universities, libraries, and communities provide tools and collaborative learning opportunities, often including 3-D printers for custom robot parts.

Home-made robots are as diverse as professional ones. Some hobbyists build intricate walking machines, others design service bots, and some focus on competitive robots. Popular competitive robots include remote-controlled fighters like those on "BattleBots," though some argue they aren't "true robots" due to their lack of reprogrammable brains, functioning more like advanced remote control cars.

More sophisticated competitive robots are computer-controlled. For example, soccer robots play soccer autonomously. A typical team consists of multiple robots communicating with a central computer, which uses a video camera to track the field, identifying teammates, opponents, the ball, and goals by color. The computer then strategizes and directs the team accordingly.

Robots and Artificial Intelligence

Artificial intelligence (AI) is one of the most thrilling and debated areas in robotics. While robots in assembly lines are widely accepted, the possibility of robots achieving true intelligence remains a contentious topic.

Defining artificial intelligence is as challenging as defining the term "robot." True AI would replicate human thought processes, creating a machine with human-like intellectual capabilities. This includes learning, reasoning, language use, and generating original ideas. While roboticists haven't reached this level, they've made significant strides in developing limited AI, enabling machines to mimic specific aspects of human intelligence.

AI systems can already address problems within specific domains. The core concept of AI problem-solving is straightforward, though implementation is complex. The AI collects data via sensors or human input, compares it to stored information, and determines its meaning. It then evaluates possible actions and predicts the most effective one based on the data. However, these systems are limited to solving problems they're programmed for, lacking generalized analytical skills. Chess computers exemplify this type of machine.

Some modern robots possess limited learning capabilities. They can determine if specific actions, like moving their legs in a particular manner, lead to desired outcomes, such as avoiding obstacles. The robot records this data and repeats successful actions in similar scenarios. For example, robotic vacuums memorize room layouts, though their functionality is restricted to vacuuming.

Certain robots are designed for social interaction. Kismet, developed in 1998 at M.I.T.'s Computer Science & Artificial Intelligence Lab (CSAIL), could interpret human body language and vocal tones, responding appropriately. Today, interactive robots are commercially available and serve as companions for the elderly, aiding in cleaning and mobility while also helping to reduce seniors' social isolation.

The core challenge of AI lies in deciphering how natural intelligence functions. Unlike building an artificial heart, scientists lack a straightforward model for AI development. While we know the brain contains billions of neurons and that thinking involves electrical connections between them, the exact mechanisms behind higher reasoning remain elusive. The brain's intricate circuitry is still largely a mystery.

As a result, AI research is predominantly theoretical. Scientists propose theories on human learning and cognition, testing their hypotheses through robotics. M.I.T. CSAIL researchers emphasize humanoid robots, believing that experiencing the world like humans is essential for developing human-like intelligence. This approach also facilitates human-robot interaction, potentially enhancing the robot's learning process.

Just as robotic design aids in understanding human and animal anatomy, AI research provides insights into natural intelligence. For some roboticists, this understanding is the ultimate goal. Others envision a future where humans coexist with intelligent machines, utilizing robots for tasks like manual labor, healthcare, and communication. Some robotics experts predict that robotic advancements could lead to cyborgs—humans integrated with machines. In this future, people might even transfer their consciousness into durable robots, extending their lives for millennia.

Robots are poised to become an integral part of our everyday lives in the future. Over the next few decades, they will transition from industrial and scientific applications to common household use, much like computers became ubiquitous in homes during the 1980s.