2017 is going to be a big year for tech, as we see the rise of both Virtual Reality and Artificial Intelligence. It will also be a good year for the Internet of Things and expanding and improving technology and connectivity across the world. CAD is, by design, an intuitive, responsive program which automates analytical steps of the design process. Whether it is creating a computer model of the parts to check that they fit well, or whether it’s more like a reverse engineering design, where parts can be created from an image, CAD already lends itself well to AI.
With the rise of Siri, Cortana and Watson, we are getting increasingly acquainted with AI as an option, and its uses with engineering will continue to revolutionise, develop and improve on many other sectors from architecture to machines, and all things in between. CAD informs effective reverse engineering practices by making sure that the parts are reproduced rapidly. With the advent of 3D printing, there has never been a better time for CAD services to be implemented. However, an important consideration is intellectual property protection.
Naturally intuitive, CADs determine whether changes should be made, and by using AI this system can become automated. Where design alterations are needed, AI can program the CAD machine to implement the correct changes, reducing human man hours, and ensuring continuity. However, it cannot necessarily be tested for viability using AI, and so AI will not be replacing traditional CAD drawing studios anytime soon. Ultimately, AI can only simulate the decisions and knowledge held by experts with years of training, and every project will still need to be signed off by a human signatory.
Prior to the 1960s, engineers carried out all drawing through physical manual exercises. After this point, computers were also implemented, helping CAD emerge as a way to produce accurate and ergonomic engineering. AI simulation software has been used for creating mobile phones, GPS and body motion simulators used by police force tracking devices and other such software. With the advancement of tech, AI will continue to inform and improve CAD, producing merging and collaboration opportunities that will enrich its scope for intuition, resolution and knowledge based alterations.
To successfully integrate CAD with AI, lengthy rules, based on decision making and problem solving would need to be implemented. This first involves heavily strategic information harvesting from experts on their practises. After much data has been collected, different knowledge based rules could be added, which could then be modified appropriately to continually refine and improve on existing methods.
A successful integration of the two would need to involve the following limitations:
• Individual components used within products need to be kept in an ordered sequential form, where there are clear relationships between components, in order for AI to select the correct component.
• The structure of the entire product would also need to be in the same type of hierarchical form, and labelled in a way that makes sense to the computer. Links between products and their composite components would need to adhere to rules based methodology, with no exceptions or deviations.
• End result product behaviour and component behaviour needs to be deduced by simulation.
This way, adding new components or products into the database and setting rules accordingly should be fairly straightforward. So long as each one is added with a knowledge based programme, the methodology should be able to inform the AI where to select parts and how to integrate them. This, of course, will still need overseeing by a human eye, and so AI is unlikely to completely overtake humans just yet!