What Is Collaborative Robotics? A Practical Guide

Artificial intelligence is rapidly moving beyond the digital world. For AI to learn how to interact with our physical reality, it needs hands, eyes, and the ability to learn from real-world experience. This is where collaborative robotics becomes essential. These robots are no longer just tools for automating simple tasks; they are the physical platforms for embodied AI. By providing a safe and repeatable way to interact with objects, collect multi-modal data, and test machine learning models, cobots are the bridge between algorithms and action. This article explores how cobots are becoming the foundation for the next generation of intelligent systems.

Key Takeaways

• Cobots are partners, not replacements

  : They are designed to work safely with people, taking on repetitive or strenuous tasks so your team can focus on more complex, creative work.

• They are accessible and versatile tools

  : User-friendly programming and lower costs make cobots practical for many applications, including quality inspection and material handling, for teams in labs, startups, and manufacturing.

• Choose the right cobot for your specific project

  : Look beyond the initial price and evaluate key factors like payload, reach, software compatibility, and ease of programming to find a model that fits your workflow.

What Is a Collaborative Robot?

Cobots vs. Traditional Industrial Robots

A collaborative robot, or cobot, is designed to work safely and effectively right alongside people in a shared workspace. This is a major shift from traditional industrial robots, which are typically so powerful and fast that they have to operate behind safety cages to prevent accidents. The goal of a cobot isn’t to replace a human worker, but to act as a partner. They take on the repetitive, strenuous, or precise tasks, which frees up their human counterparts to focus on problem-solving, quality control, or other more complex work. For researchers and developers, this opens up new possibilities for human-robot interaction and embodied AI. Instead of programming a robot to run a task in isolation, you can build systems where a person can guide, correct, or collaborate with the robot in real time.

What Makes Up a Collaborative Robot?

So, what makes a robot "collaborative"? It comes down to built-in safety features that allow it to be aware of its surroundings, especially the people in it. Cobots are equipped with advanced sensors and force-limiting technology that can detect an unexpected collision, even a gentle one. If a cobot bumps into a person or an object, it will automatically stop or reduce its force to prevent injury. This is what allows them to operate without physical safety fences. While cobots are designed for safety, it’s important to remember that a risk assessment is still necessary for any application. The specific task and end-effector, like a gripper or a sharp tool, also play a role in the overall safety of the system. These features are foundational for building practical, AI-ready research kits that can function in labs and classrooms.

How Cobots Keep Human Teams Safe

The idea of working right next to a powerful robot might sound a little intimidating, but collaborative robots were designed from the ground up with human safety as a top priority. Unlike traditional industrial robots that are caged off for a reason, cobots are packed with features that allow them to operate safely in shared spaces. This focus on safety isn't an afterthought; it's what makes human-robot collaboration possible in the first place. These systems use a combination of smart sensors and software to stay constantly aware of their surroundings, ensuring they can work alongside people without putting them at risk.

This built-in safety is what distinguishes a cobot from its industrial predecessors. Instead of relying solely on external barriers, a cobot's safety is integrated into its core hardware and programming. This approach creates a more dynamic and flexible work environment where human ingenuity and robotic precision can combine effectively. By understanding these built-in safety mechanisms, teams can confidently integrate cobots into their workflows, reduce the risk of workplace injuries associated with repetitive tasks, and focus on higher-value work. Let's look at the specific technologies that make this close collaboration a reality.

Force and Torque Sensing

One of the most important safety features in a cobot is its ability to "feel" its environment. Cobots are equipped with internal sensors that constantly measure the force and torque in their joints. If the robot encounters an unexpected obstacle, like a person accidentally bumping into its arm, it detects the spike in force and immediately stops its motion. This force-limiting technology prevents the robot from exerting dangerous pressure, turning a potential collision into a harmless bump. This feature is fundamental to ensuring that a cobot can share a workspace with a human team member without causing injury.

Speed and Separation Monitoring

Cobots don't just react to contact; they proactively try to avoid it. Using a variety of advanced sensors, like cameras or lasers, a cobot can monitor its workspace and detect when a person is nearby. This is known as speed and separation monitoring. If a person gets too close to the operating robot, the system can automatically slow down. If the person moves even closer, the robot might pause its task completely until the person moves away. This intelligent, responsive behavior allows cobots to operate safely in close proximity to people, maintaining a safe distance without constant human intervention.

Collision Detection and Automatic Shutdown

Building on force and speed monitoring, cobots have robust collision detection protocols. These systems are designed to be the ultimate fail-safe. In the event that a cobot does make contact with a person or an unexpected object, the system is programmed to register the collision and halt all movement instantly. Many cobots are designed to automatically stop or slow down when they sense a human within their working range, often before any contact is even made. This immediate shutdown minimizes any potential impact and is a critical layer of protection that allows for safe, side-by-side work.

Working Without Safety Fences

The combination of force sensing, speed monitoring, and collision detection is what allows many cobots to work without the bulky safety cages required for traditional industrial robots. This is a game-changer for facility layouts and workflow flexibility. Instead of redesigning an entire production floor around a caged robot, teams can integrate a cobot directly into an existing workspace. This ability to work alongside humans without the need for physical separation is the defining characteristic of collaborative robotics. It saves space, reduces setup costs, and creates a more fluid and adaptable environment where people and robots can truly work together.

Where Are Cobots Making the Biggest Impact?

Collaborative robots are no longer a niche technology found only in specialized labs. They are becoming essential tools across a surprising number of industries, changing how teams approach everything from assembly lines to order fulfillment. The common thread is that cobots are used to augment human workers, not replace them. By taking over tasks that are repetitive, physically demanding, or require tireless precision, cobots free up people to handle more creative and complex problems. This partnership between human and machine is where the real value appears. Let's look at a few of the sectors where collaborative robotics is making a significant difference.

Manufacturing and Automotive

In manufacturing and automotive plants, cobots are becoming trusted partners on the assembly line. They handle repetitive jobs like screw driving, applying adhesives, or performing quality checks with consistent accuracy. This allows their human coworkers to focus on more intricate tasks, such as final assembly and problem-solving. Because cobots are designed to work alongside people, they can be integrated into existing workflows without the need for large safety cages. Researchers and developers use platforms like Trossen's stationary AI kits to prototype and refine these exact types of collaborative manufacturing tasks.

Electronics and Semiconductors

The electronics industry runs on precision and consistency, which is exactly where cobots excel. They are perfectly suited for handling delicate components, placing microchips, or inspecting circuit boards for tiny defects, tasks that can be tedious and strain a person's eyes over a long shift. A cobot can perform these actions thousands of times with the same level of care and accuracy. This frees up skilled technicians to work on more nuanced parts of the production process. For teams developing these systems, a precise and repeatable robotic arm is the foundation for building a successful application.

Logistics and Warehousing

Warehouses and fulfillment centers are bustling environments where speed and accuracy are critical. Cobots are making a huge impact here by automating pick-and-place, sorting, and packing tasks. Equipped with advanced sensors, they can work safely alongside people in dynamic, shared spaces. Some cobots are integrated into mobile platforms, allowing them to move through aisles to retrieve items for an order. This helps teams process orders faster and reduces the physical strain on workers. Developing these systems often starts with mobile AI platforms that combine mobility, perception, and manipulation in a single package.

Food, Agriculture, and Pharma

In industries like food, agriculture, and pharmaceuticals, hygiene and consistency are non-negotiable. Cobots are ideal for these settings because they can operate in clean environments and perform tasks with a gentle, repeatable touch. You can find them packaging delicate foods, sorting produce by ripeness, or handling samples in a lab. Unlike traditional industrial robots that require physical barriers, cobots are designed with safety as a primary feature. This allows them to work in close proximity to people, making it easier to integrate them into existing production lines and laboratory workflows without a complete overhaul of the space.

What Can You Do With a Collaborative Robot?

Collaborative robots are incredibly versatile tools, capable of handling a wide range of tasks across different industries. Think of a cobot less as a single-purpose machine and more as a flexible platform that you can adapt to your specific needs. Whether you're in a research lab, a startup, or a large manufacturing facility, a cobot can become a valuable partner for your team. Their ability to work alongside people safely opens up applications that were once impractical or too expensive for traditional industrial automation.

The real power of a cobot comes from its ability to take over tasks that are repetitive, physically demanding, or require a level of precision that is difficult for humans to sustain over long periods. This doesn't mean replacing people; it means augmenting their skills. By letting a cobot handle the monotonous work, you free up your team to focus on problem-solving, creative thinking, and other high-value activities that machines can't do. From simple pick-and-place routines to complex assembly and inspection, cobots provide a practical way to improve efficiency and consistency in your operations. Let's look at some of the most common ways teams are putting them to work.

Pick-and-Place and Material Handling

One of the most fundamental and common applications for a cobot is pick-and-place. This involves exactly what it sounds like: picking up an object and placing it somewhere else. While it may seem simple, this task is the foundation of many automation workflows. Cobots excel at moving parts from a bin to a conveyor, sorting components for assembly, or loading items into a testing jig. They are designed to relieve human workers from tedious or strenuous tasks, like lifting heavy objects into position for a person to then work on.

For a research lab, a pick-and-place setup is a perfect environment for testing manipulation algorithms and gripper designs. For a startup, it can be the first step toward automating a key part of your production line. Because cobots can handle delicate components with precision and heavy items with consistent strength, they are ideal for material handling. A well-programmed robotic arm can perform these repetitive actions for hours without fatigue or error, improving overall throughput and consistency.

Quality Inspection and Testing

Maintaining high quality standards requires careful and consistent inspection, but having a person stare at parts all day is a recipe for fatigue and missed defects. This is where cobots shine. Outfitted with a camera or other sensors, a cobot can perform detailed quality inspections with unwavering focus. It can check for cosmetic flaws, verify that components are correctly assembled, or take precise measurements to ensure parts meet specifications. The cobot handles the repetitive, mundane work, allowing human operators to focus on more complex quality control challenges.

This application is also perfect for product testing. A cobot can repeatedly press a button, turn a knob, or open and close a hinge thousands of times to test for durability. This frees up your engineering team from manual testing and provides consistent, reliable data. With a platform like the TOTL Workstation, you can create a repeatable environment to collect structured data from these tests, helping you identify product weaknesses and make improvements faster.

Packaging and Palletizing

Packaging products and stacking them onto pallets are physically demanding and highly repetitive jobs. These tasks are prime candidates for automation with a collaborative robot. A cobot can be programmed to carefully place products into boxes, add protective inserts, and seal the packages. Once packaged, another cobot, or the same one, can handle palletizing, which involves stacking the boxes onto a pallet in a stable, predetermined pattern for shipping. This is just one of a wide range of jobs cobots can perform.

Automating these end-of-line tasks helps reduce the risk of workplace injuries associated with heavy lifting and repetitive motions. It also ensures that every package is handled with the same care and every pallet is stacked perfectly, reducing the chance of damage during transit. For businesses in logistics and ecommerce, using cobots for packaging and palletizing can significantly improve operational efficiency and get products out the door faster.

Machine Tending

In many manufacturing environments, skilled operators spend a significant amount of their time simply loading and unloading parts from machines like CNC mills, lathes, or injection molders. This is called machine tending, and it's a perfect job for a cobot. A collaborative robot can be positioned next to a machine to feed it raw materials and remove finished parts, running continuously through shifts with minimal supervision. Because cobots are easy to set up and can often work without extra safety fences, they can be deployed right next to your existing equipment.

By automating machine tending, you allow a single operator to oversee multiple machines at once, dramatically increasing productivity. The operator can then focus on more valuable work, such as quality checks, machine setup, and process optimization. This application is a practical entry point into automation for many small and medium-sized businesses, as it addresses a common bottleneck and delivers a clear return on investment by maximizing the uptime of expensive machinery.

How Cobots Help Teams Do More

Collaborative robots are powerful tools for making your team more effective. Instead of replacing people, cobots act as partners that handle the dull, dirty, and dangerous parts of a job. This frees up your team to focus on tasks that require human creativity, critical thinking, and dexterity. By taking over repetitive work, cobots allow you to scale your operations, improve consistency, and empower everyone on your team, not just robotics experts, to contribute to automation.

Automating Repetitive Tasks, Not People

One of the biggest myths about cobots is that they are here to take jobs away from people. The reality is much more interesting. Cobots are designed to work alongside humans, complementing their skills rather than replacing them. Think of a cobot as a tireless assistant that handles the most repetitive, mundane, or physically strenuous tasks. This allows human workers to shift their focus to more complex and valuable work, like quality control, process improvement, or managing multiple systems at once. As one analysis points out, many jobs simply can't be fully automated and will always need a human touch. By letting a cobot handle the tedious parts, you give your team the space to do their best work.

Scale Your Operations, Not Just Your Team

Cobots are pivotal assets that help you increase productivity without necessarily increasing headcount. When a cobot takes over a repetitive task like material handling or machine tending, it can run consistently for hours, leading to higher output and more predictable workflows. This allows you to scale your operations and meet growing demand more efficiently. Instead of hiring more people to do the same repetitive job, you can empower your existing team to manage and oversee automated processes. This approach lets you grow your output while investing in your team’s skills, turning them into robot operators and process managers rather than manual laborers.

Empowering Non-Expert Users

You don't need to be a robotics engineer to use a cobot. Many people believe that this technology is a costly investment only accessible to large companies, but modern cobots are designed for accessibility. Their affordability is often matched by their ease of deployment and low maintenance needs. Many systems, like our TOTL Workstation, are created to be intuitive, allowing people without deep technical backgrounds to set up and run automation tasks. This user-friendly approach means that researchers, students, and small teams can start working with advanced robotics quickly, testing ideas and gathering data without a steep learning curve or a massive budget.

How to Choose the Right Collaborative Robot

Once you’ve decided a cobot is right for your project, the next step is picking the perfect model. This isn’t about finding the “best” robot, but the best robot for your specific goals. Just like you wouldn’t use a moving truck to grab groceries, you wouldn’t use a heavy-duty industrial cobot for a delicate lab experiment. Thinking through a few key factors will help you narrow down the options and find a platform that gives you the capabilities you need without paying for features you won’t use. Let’s walk through the most important considerations.

Payload Capacity and Reach

The first and most fundamental question to ask is: what do you need the robot to lift, and how far does it need to move it? Payload is the maximum weight a robot arm can carry, while reach is the maximum distance it can extend from its base. A cobot designed for assembling tiny electronics will have a much lower payload and shorter reach than one built to stack heavy boxes on a pallet.

Start by analyzing your task. Measure the weight of the objects you’ll be manipulating, including the gripper you plan to use. Then, map out the workspace to determine the required reach. Choosing a cobot with the right payload and reach ensures it can perform its job safely and efficiently. Trossen’s lineup of robotic arms offers a variety of options to match different project scales.

Ease of Programming and Deployment

One of the biggest advantages of collaborative robots is their user-friendliness. Many cobots can be programmed without writing a single line of code. Instead, you can use simple hand-guiding, where you physically move the robot arm through the desired motions to teach it a task. Others use intuitive, tablet-based interfaces with drag-and-drop commands.

This ease of use is a game-changer, especially for small teams, researchers, and educators who may not have dedicated robotics programmers on staff. It dramatically shortens the time from unboxing to running your first program. When evaluating a cobot, look for platforms that prioritize accessible programming and quick deployment. This allows your team to focus on innovation and problem-solving, not complex coding. For example, Trossen’s Data Collection SDK is designed to get you capturing data quickly.

Software Compatibility and ROS Support

While simple programming is great for getting started, your long-term success often depends on software flexibility. A cobot is more than just hardware; it’s part of a larger ecosystem. You need to consider how it will integrate with other tools, like cameras, sensors, and custom software. Many manufacturers offer marketplaces with certified accessories, which can be a great starting point.

For researchers and developers, Robot Operating System (ROS) compatibility is especially important. ROS is an open-source framework that provides standard libraries and tools for building robot applications. A cobot that supports ROS is much easier to integrate into complex research setups and custom workflows. This compatibility is a core feature of Trossen’s AI-ready platforms, ensuring you can build on a flexible and powerful software foundation.

Modularity and Extensibility

Your needs today might not be your needs tomorrow. That’s why modularity and extensibility are so important. A modular cobot allows you to easily swap out components like grippers, sensors, or even entire arm segments. This adaptability means you can reconfigure the robot for new tasks or upgrade its capabilities as your projects evolve.

An extensible platform is an investment that grows with you. Instead of buying a new robot for every new challenge, you can adapt the one you already have. This not only saves money but also reduces development time, since your team is already familiar with the base system. Look for platforms designed for modification and expansion, as this will give you the most long-term value and help you build repeatable, scalable robotics workflows.

Finding a Model That Fits Your Project

Ultimately, choosing the right cobot means matching the hardware to your specific application. Are you automating a pick-and-place task, building a quality inspection system, or collecting data for an AI model? Each of these applications has different requirements. For example, a data collection project might prioritize multi-modal sensor integration, while a machine-tending task would focus on reliability and cycle time.

Think about your entire workflow, from setup and programming to daily operation and future expansion. By considering payload, ease of use, software integration, and modularity, you can find a model that fits your project perfectly. Exploring platforms designed for specific use cases, like a mobile AI robot for autonomous tasks or a dedicated workstation for AI training, can help you find a solution that’s already optimized for your goals.

Understanding the Cost and ROI of Cobots

Thinking about adding a cobot to your workflow means thinking about the budget. It’s not just about the price tag; it’s about understanding the full financial picture and what you get in return. A clear view of the costs and potential returns helps you make a smart decision that aligns with your team’s goals, whether you’re outfitting a university lab or prototyping a new product. Let’s break down how to approach the cost and return on investment for a collaborative robot.

Upfront Costs vs. Traditional Robots

One of the biggest reasons teams are drawn to cobots is their accessibility. The upfront cost for a collaborative robot is typically much lower than for a traditional industrial robot. While a large, caged industrial robot can require a significant capital investment, a standalone cobot often falls in a more manageable range. This lower barrier to entry makes automation a realistic option for startups, university research programs, and R&D teams. It's key to getting projects off the ground and into the hands of the people who can start innovating right away.

Calculating the Total Cost of Ownership

The initial purchase price is just one piece of the puzzle. To get a true sense of the investment, you need to consider the total cost of ownership (TCO). This includes the robot itself plus any ongoing expenses like maintenance, software, end-effectors or other peripherals, and training for your team. Thinking through the TCO from the beginning helps you create a realistic budget and avoid surprises. A comprehensive understanding of TCO ensures you’re not just buying hardware but are investing in a complete, sustainable solution for your project’s long-term goals.

Calculating Your Return on Investment

Once you have a handle on the costs, the next step is to calculate your return on investment (ROI). This is where you measure the value the cobot brings to your operation. ROI isn’t just about reducing labor costs; it’s also about gains in productivity, improvements in quality, and the new capabilities your team now has. To accurately assess the ROI, think about what you’ll be able to accomplish. Can you run tests overnight? Can you free up a skilled engineer from a repetitive task to work on a more complex problem? These benefits contribute directly to your bottom line and help justify the investment.

Clearing Up Common Myths About Cobots

Collaborative robots are powerful tools, but they’re also surrounded by a lot of confusion. As cobots become more common in labs, workshops, and even on our desktops, it’s important to separate the facts from the fiction. Let's clear up a few of the most common myths so you can see where these robots really fit into your work.

Cobots will replace human workers.

This is probably the most persistent myth, and it’s simply not the case. Cobots are designed to be partners, not replacements. They excel at handling repetitive, physically demanding, or precise tasks that can be tedious for human workers. This collaboration frees up your team to focus on more valuable work that requires creativity, problem-solving, and critical thinking. Think of a cobot as a tool that enhances productivity, allowing a single person to manage a more complex process or a researcher to run more experiments.

"Cobots are only for large factories."

It’s easy to picture robots belonging only to massive, automated factories, but that’s an outdated view. Modern cobots are designed to be affordable, lightweight, and easy to program, making them a perfect fit for smaller operations. University labs, R&D teams, and startups can use platforms like a WidowX AI arm to automate data collection or test new ideas without the budget or space required for traditional industrial robots. This accessibility is what makes them so valuable for research and education.

"Cobots are completely safe without any protocols."

While cobots are built with safety as a core principle, that doesn't mean you can ignore safety protocols entirely. Features like force-limiting sensors and collision detection make them significantly safer to work alongside than their industrial counterparts. However, every application is different. A robot handling a soft sponge is very different from one handling a sharp tool. A proper risk assessment is always necessary to ensure a truly safe environment for everyone working in the space.

"Cobots don't need any human oversight."

A cobot is not a fully autonomous employee. It’s a sophisticated tool that still needs human intelligence to guide it. Humans are responsible for programming the tasks, supervising the workflow, and making adjustments when things change. In fields like AI research, human involvement is the entire point. Systems for teleoperation and data collection rely on a person to guide the robot and generate the data needed to train new models. The human is, and will remain, in the loop.

What's Next for Cobots and Physical AI?

Collaborative robots are evolving far beyond simple, repetitive automation. They are quickly becoming the physical foundation for the next major shift in technology: embodied artificial intelligence. This is where the hardware of robotics meets the brain of modern AI, creating systems that can perceive, learn, and act in the real world. The future of cobots isn't just about what they can do, but what they can learn. This shift is creating incredible new opportunities for research, development, and practical problem-solving across every industry.

Integrating AI and Machine Learning

The next generation of cobots is learning to think. While early models excelled at repeating a programmed task, integrating artificial intelligence and machine learning allows them to become more adaptive. Instead of just following instructions, cobots can now use advanced sensors and vision systems to react to their surroundings in real time. This means they can learn new tasks through observation, adjust to variations in parts, and work more intelligently alongside human partners. This move from rote automation to adaptive learning is powered by AI-ready hardware designed to support complex, intelligent workflows.

Building a Foundation for Embodied AI

Embodied AI refers to intelligent systems that can physically interact with their environment, and cobots are the ideal platform for bringing this concept to life. To train an AI model to perform a physical task, you need massive amounts of real-world data. Cobots are perfect for this role. They can perform a task thousands of times while a data collection system captures every movement, sensor reading, and interaction. This structured data becomes the fuel for training machine learning models that can then be deployed back onto the robot, creating a powerful cycle of learning and improvement.

Expanding Market Growth and Applications

The market for collaborative robots is growing at a remarkable pace, and for good reason. As the technology becomes more affordable and flexible, it’s finding a home in a huge range of industries, from automotive and electronics to food production and pharmaceuticals. This growth isn't just happening in large factories; it's being driven by small businesses, research labs, and startups who can now access powerful automation. As cobots become easier to deploy and integrate with AI, their potential applications will only continue to expand, opening up new possibilities for innovation in mobile and stationary robotics.

Related Articles

• What Is a Robot? The Ultimate Guide for 2026

• The Trillion-Dollar Market Hidden in Unautomated Human Labor

• Why Pragmatism in Robotics Wins: The Strategic Edge of Simplicity, Scalability, and ROI

Frequently Asked Questions

I'm not a robotics engineer. How difficult is it to actually start using a cobot? This is a great question, and it gets to the heart of why cobots are so exciting. You absolutely do not need to be a programmer to get started. Many modern

What's the difference between using a cobot for simple automation versus for AI development? Think of it as the difference between repeating and learning. For simple automation, you program a cobot to perform a specific, repetitive task, like picking up a part and placing it in a box. The robot executes that exact sequence perfectly every time. For AI development, the cobot becomes a tool for teaching. You might use it to collect data by demonstrating a task hundreds of times, allowing an AI model to learn how to handle variations, like objects in different positions. The first is about efficiency, while the second is about creating intelligence.

The post mentions payload and reach. How do I figure out what my project actually needs? The best way to start is by analyzing your specific task. To determine your payload requirement, find the heaviest object your robot will need to manipulate and be sure to add the weight of the gripper or tool you plan to use. For reach, map out the physical workspace where the robot will operate. Measure the distance from the robot's base to the furthest point it needs to access. It’s always better to choose a model that comfortably exceeds your minimum needs, but you don't need to over-spec it. Starting with a clear picture of your task will guide you to the right hardware.

Are cobots a good investment for a small team or a research lab with a limited budget? Yes, they can be a fantastic investment. Unlike traditional industrial robots that often come with a six-figure price tag and require major facility changes, cobots are designed to be more affordable and easier to integrate. For a small team or lab, a cobot isn't just a machine; it's a force multiplier. It can run experiments overnight, handle tedious data collection, or perform repetitive assembly, freeing up your team to focus on creative and analytical work. The return on investment often comes from this increase in your team's capacity and the speed at which you can test new ideas.

Why is data collection so important for the future of cobots? Data collection is the bridge between robotics and artificial intelligence. For an AI model to learn how to perform a physical task, like sorting objects or assembling a product, it needs to learn from a massive amount of real-world examples. A cobot is the perfect tool for generating this data. By using teleoperation systems to guide a cobot through a task, we can capture every motion, sensor reading, and camera feed. This structured information becomes the training material for building smarter, more adaptive robots that can perceive and act in the physical world.