How to Reduce Waste With Autonomous Robots

The birth of robotics – all the way back to GM’s Unimate – was all about doing jobs that were simply too dangerous, repetitive and physiologically difficult for humans to execute safely and securely. While robots have evolved up to (but not beyond) many manufacturers’ wildest dreams, a new generation of challenges poses will force robots to adapt in new ways.

Two of these forces are most concerning: environmental change and commodity scarcity. The most important aspect of this is maximizing the efficiency of robots in order to minimize their total impact on the environment and the use of energy in executing their necessary processes. While adjusting robot design to make this more practical is possible, the real challenge comes in where you don’t see robots.

Outside of automotive and consumer electronics manufacturers, very few manufacturers make effective use of robots. This is because robots need extensive programming and fixturing in order to reach outputs compliant with most manufacturers’ quality goals. Autonomous robots effectively automate the programming element within the factory environment, working according to the real position of parts. With these in hand, manufacturers can eliminate much of the unnecessary waste they see in their non-robotic processes – without wasting money on trying to get it to work!

How Robots Save Consumables

When a human being plans an action, they are usually able to follow it – at least the first time they do it. As time goes on, however, they can tend to take shortcuts, lose focus, get tired – all things that HAPPEN and should be accommodated to a reasonable degree, but do not meet the standard that people today expect from industrial goods. That’s where robots can make the most impact. 

Robots that have a coherent plan always follow it to a tee. This is because robot programming effectively defines positions and orientations in space with limited error, at tolerances of plus or minus .05 millimeters in many cases. This kind of reliability means that – if a program can be generated – the use of materials like varnishes, finishes, paints, coatings, topcoats, gels or other unique and sometimes scarce chemical combinations can be avoided because there is no need to substantially over-apply them in order to guarantee sufficient coverage.

This is of course unless you want to be as generous as possible. Aside from the consumables themselves, the accuracy and consistency of well-programmed robots means that rework, rejections and scrapping of poorly worked parts becomes almost non-existent since the robots only fail if the parameters around them fail. This means that, yes, there could be incidences of failure or a bad batch, but in generating an optimized process, these would be the exception instead of the rule, and ultimately over the full time horizon of a robot’s operation, waste becomes significantly decreased compared to existing non-robotic workflows.

And how long is that time horizon? Some robots can last more than 4 decades, and very often when a robot becomes out of date, there are simpler, smaller or less demanding tasks for which a robot can still find great use – all meaning that all sorts of consumables waste can be minimized for generations to come with an investment in one single, lone, powerful robot.

How Robots Save Energy

As you can imagine, reducing waste can also limit energy waste. There are a few important things to realize, however: robots can function in environments that are inhospitable and unsafe to humans, and robots can also run on relatively low energy consumption compared to the actual task a human may execute. 

For instance, one study has shown that as much as 70% of industrial energy expenditure is simply composed of the energy requirements used in air conditioning and environmental protection in order to get humans working in shops. When considering the idle time of many machines, as much as 87% of energy expenditure doesn’t go directly into the value-added processes that manufacturers actually profit from.

Robots increase both labor capacity and labor durability, while improving the quality, ease and safety of the manufacturing jobs executed by humans. At the same time, this energy savings results in an environmental benefit that may open manufacturers up to new funding and tax benefit opportunities that otherwise would not have existed before. 

How Autonomous Robots Improve Efficiency

What’s important to understand here is that a well-programmed robot can do it all, but generating a robot program is no easy task – it can take days, weeks, months depending on the workflow and further requires high degrees of precision, accuracy and even trial and error to realize a reliable finished product.

Autonomous robots are different – by seeing, planning and executing operations without the need for programming but instead through the use of vision systems, autonomous motion planning and highly advanced AI, programs can be generated in real production time and use existing available hardware to automate human-level processes with machine-like consistency.

When you have a system that you can throw new challenges at, you can save the best of your workforce for real problems.

With AutonomyOS™ and AutonomyStudio™, it’s never been easier to deploy an autonomous robotic system. Using 3D Perception with AI-based Task Planning and Motion Planning, manufacturing engineers and integrators can configure autonomous robotic systems for value-added processes that allow manufacturers to achieve more consistency and flexibility in production than ever before.

Why You Should Stop Programming Your Robots

Robot programming is a complicated tax. Even novice programmers are often hot in demand, and very often they will work for integrators who operate on the basis of one-off or serial integrations with different manufacturers. These are surprisingly small businesses that are often unable to address new customers after a very short period of time, as their queue of clients (relative to the work they can provide) can fill up quickly. 

At the same time, there are a variety of advanced manufacturing systems, industrial automation techniques and more possible process improvements that engineers can be working on to create productivity and growth for manufacturers. Ultimately, the opportunity cost of trying and failing repeatedly to integrate robots is very high compared to the other benefits that a manufacturer can get from new systems. While robot programming is complicated and tedious, the actual capabilities found within existing programming languages are highly limiting. Systems that try to make robot programming easier often don’t do enough, and AI can instead allow engineers to focus on bigger problems. 

Robot Programming is Complex

At its base, programming a robot is fundamentally more complex than programming a computer or generating the usual piece of software. For instance, a six-axis robot can have six degrees of freedom, an indefinite number of positions in space, as well as have positions and joint constraints that cannot be exceeded without risking the integrity and performance of the robot. 

In this circumstance, you can imagine that to achieve a single output, you must do six or seven times the work of a traditional programmer just to ensure that your work achieves a reasonably consistent output. At the same time, the costs of organizing and testing these programs is difficult to manage because industrial settings are scarce and expensive to put together on short notice, meaning that they are typically reserved for actual production. When a new program is tested and deployed, that means production must be shut down.

At the same time, testing in a production environment doesn’t come with any guarantees. You can be limited in your ability to actually ensure that your production goals are met, and if there are any quality or fixturing issues, you may see your process go back to the drawing board entirely. 

Robot Programming Languages are Designed to Limit Flexibility

Robot programming originated as a discipline to describe simple tasks for industrial robot arms. Robot programming was only ever meant to describe points, motions, speed, geometries and other control mechanisms – it is not designed to intuitively understand human reasoning and adopt the same strategies humans do.

This is in part because robots are such a leap forward that they could only ever originate in a lab. The first manufacturing robots simply execute a pre-recorded series of positions, and were not even based on computers, transistors or servos because at the time they were too expensive. 

No decision or response to their environment was required here, and what this implies is that ultimately robots are simply a mechanical system, not even an intelligent machine in themselves. This makes robot programming more mundane (in some ways) than traditional computer programming, and while hand guiding, teach pendants and other methods have made things easier, there are ultimately far fewer resources for engineers to call on in generating their motion programs. Building those resources is almost further out of the question, since they’re not really necessary if robot programming can be automated on the whole. 

Offline Programming and Other Systems Don’t Help All That Much

Another environment to understand is the virtual world. Robot programming has primarily moved to offline systems in an effort to save material, manpower and factory space. While there are a variety of tips that can make robot programming easier, the real question is if you want to endeavour to program a robot at all.

Why is this the case? When you have an industrial robot accessible to you in a production-type environment, the cost may be high, but it is far easier to anticipate and structure the needs of the robot within the environment, making it easier to design and model the process you want to execute and provide clarity to your programming task.

An offline programming solution may take some time out of the actual preparation required in this environment, but if it duplicates work by not serving as a perfect proxy for all the unexpected imperfections that a real production environment offers, then you can expect offline programming to perhaps limit some costs but actually involve more work than traditional programming. 

Usually work is the largest cost center of programming tasks, while any wasted material or damaged equipment renders the concept of “payback” on OLP to be totally irrelevant. The Romans might have said “caveat emptor” – we can put it more simply as: buyer beware. 

AI Can Allows Engineers to Focus on Other Problems

Age of autonomous manufacturing is upon us, and the real benefit is labor-saving and entirely new skills paradigms that no longer require tedious robot programming as a core competency. 

This is fundamentally an enabler for the productivity of thousands of engineers who are managing a variety of industrial operations without the benefit of real automation technologies. In fact, high-mix manufacturing very often uses no manufacturing automation whatsoever despite the critical role these firms play in producing so many of the essential and critical goods we use in our daily lives.

With autonomous robots, these engineers will be able to go from “knowing they can do better” to actually doing it, because despite the fact that they can know they can do better, very few systems that allow them to do so exist today. 

With AutonomyOS™ and AutonomyStudio™, it’s never been easier to deploy an autonomous robotic system. Using 3D Perception with AI-based Task Planning and Motion Planning, manufacturing engineers and integrators can configure autonomous robotic systems for value-added processes that allow manufacturers to achieve more consistency and flexibility in production than ever before.