Artificial Intelligence in Embedded Systems
No longer is artificial intelligence just a sci-fi story or only possible in a lab setting. It is now practical, and can be done in embedded systems. Deep learning systems can automate tasks that previously required a person or thousands of lines of hard-coded conditional catches.
What is it? What does the platform look like?
Artificial intelligence platforms can take various forms ranging from large 12"+ boxes connected to networks and dozens of sensors down to 2" credit card sized modules and just a few cameras. Gwentech specializes in embedded applications, whereby you can specify the parameters important to your application.
For a traditional embedded system to distinguish error or abnormal conditions, typically the conditions must be pre-determined or defined by some other manual means. For example, a closed-loop leveling system on an assembly line platform might need limits defined that encompass worst case normal conditions. If every 'worst case condition' wasn't pre-programmed at design time, a momentary imbalance could false trip the whole line and shut it down unnecessarily. An artificial intelligence based system can learn more advanced normal and abnormal conditions rather than simply comparing simple threshold levels.
A few real-life applications:
- People counter in a line
- Security, face recognition for shoplifting offenders
- Product quality control in a manufacturing environment
- Loading characteristics in a smart power supply
We design for Manufacturing and Service
A high performing system design is only part of the equation. Manufacturing costs and reliability are always working against keeping you competitive. With this in mind, Gwentech designs from the end to the beginning, promising to deliver a whole solution, not just a product design. We want to help you succeed.
