Each of us has limitations, strengths, and weaknesses. Our associations with social groups—including our friends, family, teams, schools, companies, towns, counties, countries, etc.—enable us to combine our strengths into a collective, such that we all contribute to an overall measure of excellence. There is strength in numbers. This most human of principles needs to apply to IIoT, smart manufacturing, and AI if we are to reach the next step of smart manufacturing achievement.
There are a lot of very clever software engineers working to develop AI software; though, in all cases seen so far, the intelligence continues to be static algorithm-based, rather than true intelligence that would fundamentally adapt by itself. Perhaps the most common example of such software is emerging from the automotive industry, where many carmakers have established teams of engineers, combining sensors and cameras together with their smart software and striving to make the leap toward completely autonomous driving. Based on this level of engagement, forget it. No matter how clever any individual car becomes, working alone, they will always have to tiptoe around, needing the support of humans, just in case; significant risk of blindness will endure, and no potential risk of an accident will ever be acceptable.
To make a step-change that will attain a higher level of intelligence requires autonomous cars to simply talk to each other. This can let cars know what is around the next corner and then be able to coordinate actions such that cars as a group decide what to do together, safely. There may even be input from fixed stations, such as at intersections so that there is less dependency on live traffic at critical points. The dependency on having faultless internal sensors is then avoided, as confirmation of observations and events can be confirmed, taking data from multiple sources and immediately exposing any inconsistent or erroneous input.
To read this entire column, which appeared in the April 2020 issue of SMT007 Magazine, click here.