Tuesday 2 January 2018

Part I - Computers, Robotics, Automation and Disasters are on a collision course


86 Seconds.

The level of assembly and automation used in the Ford Motor Company Cologne factory, shown in the video at the bottom of this article does not surprise me. Nor does the path how we got to this point. Many have ignored the why and how manufacturing  of consumer products and goods have got to this level. To be ignorant of creators and innovation and why our world has developed this path should not be taken lightly or for granted. Are we to be saved by the robotic arm and computers in the future? Can a robot replace humans when a disaster strikes? Will a satellite detect and prepare an army of humanoid appearing drones to repair damage to your home after a hurricane strikes?

The video is an impressive demonstration why science matters and in the use of mathematics, applied to complex geometric constrained found during assembly of complex parts, that are programmed and managed by computers. The level of repeatable accuracy using robots cannot be debated and should not come as a surprise. They are routinely accurate to within 0.000001". Some Computer Numeric Controlled machining centers are capable of creating components with nanometre precision.

The computers and robotic machines in use today are designed by some of the best mechanical, electrical and chemical engineers in the world. The software used to develop, design and create the factory floor from raw materials to car in your driveway has transformed how thousands of products are created and built. And yet, many are shocked and fear the technology and the implementation of robots and the use of computer based technologies. Many believe it is only a recent development and should be stopped. To halt progress would require our world to go back over 50 plus years in our evolutionary time scale. To put this genie back in the bottle is impossible.

The first generation of this technology was first implemented on a large scale beginning in the mid 1960's for very basic X and Y dimension limited manufacturing requirements after its development by General Motors in 1959 at its die casting plant in Trenton New Jersey. George Devol invented the Unimate 1900 series robotic arm with over 450 built and is considered the pioneer in the use of artificial limbs to replace humans that were placed in dangerous work environments like stamping plants that formed steel body panels under immense loads. It's adoption was quickly implemented in Japan, migrating to Germany and the U.S. by 1971 through 1979 and soon began crossed manufacturing domains, spreading and being used around the world in every industry and specialty vertical.

Education focused on STEM - Sciences, Technology, Engineering and Math disciplines is critical to the successful implementation of these production solutions. Nations are racing to be the best in each of these domains; computer numeric control (CNC), computer assisted / aided design (CAD), computer assisted modeling (CAM). Early adopters were to change our world forever. The development of these products could not be stopped anymore than water evaporating when the sun is brightly shining.

The advent of a rich computer based graphical user interface developed by Xerox in 1974 was soon leveraged by IBM, Microsoft and Apple. It would take a mere 5 years before Steve Jobs and Bill Gates would turn the world on its head. In the early 1970's, a car manufacturer would require 5,000 engineers to design, develop, test and prepare an assembly line to be ready for a new car model. No longer was the computer to be limited to calculating spreadsheets and financial numbers using a IBM mainframe developed in 1952 some 22 years earlier.

Some of the most successful computer assisted design engineers were early graduates and visionaries because they were exposed to the personal computer that became widely popular with the release of the Intel 8086 microprocessor chip in 1978 and Motorola 68000 in 1979. Those that bought these early systems saw the future potential for computer controlled processing of more than just mathematical equations, but fully managed development processes for every conceivable application including manufacturing. These whiz kids began to graduate from universities around the world in the mid 1980's and were the first generation to be computer literate and push the boundaries of what a microprocessor was capable of.

It is where the majority of sustainable jobs over the next 20 years will originate worldwide. The demand for STEM graduates is skyrocketing at multinationals and in your local hometown USA small job shop business. This has largely been proven true with the growth of Silicon Valley just outside of San Francisco between 1982 and 2005.

Crowdsource engineering is also becoming more prevalent that enables specialty engineers and STEM graduates to collaborate on building new products. Individuals are not long constrained to be employed by a large enterprise corporation.

The next logical and likely level of innovation will be community and regional manufacturing plants that are capable of being used for by any collaborative remote based developer team. These plants will house raw materials and universal component assemblies that can serve multiple types of products, which can then be placed onto a shared production assembly line and be tracked by QR barcode from component and parts to finished product ready for local shipping and delivery.

Access to industrial quality 3D Laser Printers, 3/4/5/6 Axis CNC machining centers along with automated pallet sub-assembly construction systems and interconnected assembly lines and shared resource computing power is not in the far distant future. We've already seen the size and scale of large robot controlled plant floor logistics management used in manufacturing plants and consumer goods warehouses which solves the final hurdle in the supply chain cycle.

It's not a question of if it is possible, but when it will become available directly to small groups or directly to the consumer. Over the next 20 years, you will likely see the first 'brands' make the leap. They will change how products are developed and sold from the current model of design, develop, test, contract, market, produce, dealer and finally the consumer to something like this;

- Modules (size of product and capacity)
- Format (requirements for above)
- Options (features and built-in apps or services)
- Personal preferences (colors and fabrics)
- Production (plant)
- Direct delivery (consumer)

The warehousing model could eventually be eliminated and direct to home logistics will be implemented using automated platforms including driverless trucks, trains and ships. Some products will be available in their modular sub-assembly configurations to complete the assembly yourself for the DIY market that is already beginning to flourish.

There is the distinct possibility that when combined with the online shopping experience many are already transitioning too, the retail floor space domain will likely disappear for many consumer products and goods by 2030.

Our world is experiencing an industrial revolution that will have consequences and influence human social behavior that many are not be prepared for. Which is a lesson we have failed to learn during the second industrial revolution. Humans have been able to develop and assemble cars at rates of less than a minute since the late 1960's. But 55 years ago, it required a plant to be over 2 million square feet, warehouse millions of dollars in parts and require over 25,000 people to work on the shop floor and manage the facility. Today, that number is less than 1,500 - 2,000 and is half the plant size of its predecessor. Some multi-national owned manufacturing plants are currently expanding in physical size to reduce the number of sub-assembly plants used in some regions of the world. As more and more robotic automation is implemented onto the shop floor, it makes less and less sense to have distant based sub-assembly facilities because the need or requirement for specialized manufacturing techniques is reduced.

It is likely we will continue to see a blend of complete end to end manufacturing at some large manufacturers while others focus on being a 'flex' plant, capable of producing multiple types of sub or complete assemblies ready for delivery. In either scenario, the number of people required in these facilities is beginning to shrink and migrate away from the shop floor to the head office design and development laboratories or in some cases, the employees residence based  home office. 

This is the backdrop to what lies in store for many countries. Shop floor jobs are disappearing and with it, the level of education needed to supervise their operation. Even these jobs are in jeopardy as as robots begin to take care of basic warehouse and robotic arm service and maintenance functions. 

This raises some fundamental organization and agency policy questions how society prepares response options to natural, environmental and pandemic disasters. How do we plan for the future events. Many aspects of disaster response have already undergone major technological change and procedural process change as a consequence. How far into the future do we anticipate robotic disaster response? Don't look now, but robots have already been deployed to some of the biggest disasters in recent history. Their use is likely to expand. In the future, they may be capable of responding in as little as 86 seconds.



Ford Motor Company Fiesta Car plant in Cologne Germany. Video Source: Ford Motor Company on YouTube.

For now, sit back and watch the video. In part II, we will dive into how computers, robots and software are changing how we deal with disasters today and what is potentially possible in the near future. Today their accurate interpretation and analysis of a the destructive power of a hurricane is still (relatively) crude. In the future, their precision may rival that of CNC machining centers. I'm not sure emergency management agencies are ready for what is in store.

Atlas Robot built by Boston Dynamics. Image Source; Wikipedia