philosopher76 said:
Hi:
I am 36 years old. I like very much artificial intelligence and in particular speech recognition. However, I would like to study a career in telematics and after that a master in artificial intelligence. The reason for my first election is that there are some subjects like signal transmission, microprocessors, digital and analog electronics, which I would like to understand and get acquainted on how mobile devices receive and process the information.
I hope some of you can give me some advice.
Thanks!
Hey philosopher76 and welcome to the forums.
Although there are going to connections between the two for real-world applications, I think, and this is just my opinion, that the two are separate enough from each other when it comes to actually designing these kinds of platforms and applications in the real world.
If you focus on telemetry, then you will be focusing on issues that are completely of a different nature, although not completely disjoint.
AI refers to a bunch of concepts that involve things like learning, planning, classification and so on and telemetry focuses on providing a sound basis for long-distance communication.
What I imagine would go on is that the overlap between the two would involve that the telecommunication experts would tell the system designers for the software and hardware components what the communication specifications are for the platform itself, and from this the software and hardware people (in relation to the functioning of the platform that does not deal with the network or telecommunications components) will use that to design this part of the platform.
There would be no need for those guys to do any more than that since it would take away from their focus and because the only way things get done nowadays is because everyone focuses on what they know and what they are good at so that the sum of the parts is greater.
If you want to learn AI for learnings sake, then that's great but just take note that in modern design, what usually happens is that things are decomposed usually by functionality in some way, and from this interface specifications are drawn up between different functional parts that form a kind of 'contract' between the different designers responsible for the different functional parts.
From this template, it allows the people to work more or less autonomously and independently from one another and when the relevant specifications and constraints are drawn up for every other design component to use, then they will use that and go on doing what they need to do.
The advantage I think you will have though in this situation is understanding the data limits which will help you create algorithms that will work under these constraints as well as ones that give a good trade-off between functionality and real-time response.
