Oliver Mihm, President of Plexus EMEA, takes a look at artificial intelligence and the pace of adoption in today’s electronics manufacturing industry
The electronics industry is going through a period of change, some would say a seismic one, with the introduction of artificial intelligence (AI), alongside Industry 4.0, into the product manufacturing space. Some believe AI will prove disruptive on the manufacturing floor. Others think it is the only way forward. Some consider it a marketing ploy.
AI methodology is, in fact, the necessary next step in the evolution of manufacturing that will move the industry forward in a meaningful way. The question is to what degree? True AI is being implemented, but is it just an overarching term for processes that are not actually users of artificial intelligence, in the true sense of that term?
Defining AI can prove an interesting discussion, as frequently, when people talk about AI, it is in a general sense. The 'all-knowing' machine, capable of learning everything about the world around it, as delivered by popular media. The reality is that adopting AI principles in the near term requires a narrowing of the scope of integration to specific technologies.
A more practical approach, which should be most effective for companies incorporating intelligent technology, is to focus on how large sets of data can be incorporated to improve the effectiveness of the production at hand. The more data there is to capture, the more a machine can be programmed to learn patterns and, therefore, maximise the effectiveness of the process.
It has been a long time since the processes involved in manufacturing have changed significantly. While the production of PCBAs has been automated for years, it is the ability to use significant amounts of data to make that process even more productive that has delivered a breakthrough. Today, optimising the amount of time that equipment can be operating is the key. By introducing AI into the manufacturing process, it will, in the long term, increase the efficiency and effectiveness of equipment.
AI - the first steps
Making an investment in AI technology, in the design through to the manufacturing process, can come at a significant cost. So, it is essential that the right tools are utilised for the right tasks. Currently, vision-guided robotics are being used with sensors and cameras, coupled with robotic arms, to enable a machine to accurately place materials into the process by working on the data it has received. A decade ago, this would have been uneconomical. But, with advancements in the technology, sensor-driven processes are now proving to be an economical and successful addition to the manufacturing process. These are particularly effective in higher mix environments, where the machine has to cope with variations in the production.
Today, the electronics industry could be said to be at a stage of sophisticated manufacturing. There are robotics that work in an ‘intelligent’ way. This means that the non-human technology is able to take in information, make a decision based on that input and then reiterate that process. These are small steps on the road to what is possible with full scale AI within the manufacturing sector. But it does highlight how one can apply the guiding principles of AI.
With full AI on the horizon, it is important that companies prepare their technology to support the future. OEMs need to work with partners that are already preparing for the changes ahead and have the skills in place to optimise their manufacturing processes. The challenge is to identify which applications within a high mix, low volume program will benefit most from a move to specific AI-enabled solutions. An experienced partner will be best placed to advise on this, and its full value stream capabilities will enable the product to be designed to allow for robotic assembly to create a smooth transition from concept to delivery. The partner should also have the ability to extract data to continue the optimisation of the process.
The introduction of true, full scale AI is on the way and it is most likely that this will occur within the customer’s products before the manufacturing floor. It will also bring about a change in business models; not just for the manufacturer, but for its customers too.
Predictability and reliability of AI will continue to improve, enabling increased sophistication in products that enhance the end users’ experience. By incorporating the benefits of AI into the manufacturing process, these will be brought to market at a swifter and more effective pace, with an ability to hit that window of opportunity and innovation much faster.
It may appear to be a slow development process now, but ten years’ down the line, the impact of the change will be felt across the electronics manufacturing sector.