|
Great, you know what you want to make. Go ahead, and do it. Look at projects previous folks have done, see how then can be adapted for your need. If you are interested in getting this One Amazing Thing working, great! You have the motivation and cranking through what you'll need to lean to succeed will be possible. IF, on the other hand, you want to make 1e6 of these things ("tech giants seems to capture most of the conversation")-- that a completely different story. You're going to need a skilled team, ideally one that has been to a rodeo before. I can't help you there. But if you're building one, for fun --- Just DO It. Ask friendly robotics club members when you get stuck. IF your problem is particularly and singularly unique, then whatever domain that entails, you will need to master; that's your Secret Sauce. Ivan Sutherland, who was Thesis Advisor to CMU's first Robotics Institute PhD said something to the effect that after 4 years of a Robotics PhD program, you end up with the ability to solder and attach connectors with a high degree of confidence. A True Roboticists requires deep knowledge in mathematics, software engineering, mechanical engineering, manufacturing small lots, reliability engineering, materials engineering, data collection & graphical evaluation methodologies..... What one might call a Systems Engineer's handbag. It also helps to be inspired by the wonders of nature that biology affords us. The details? eh, they change. Whether you used some hot cpu, some hot language, some hot OS --- Marc Raibert's hopping & jumping robots used C and BSD4.3 Unix, with the kernel locks cleaned up (reduced) so a 1 kHz kernel interrupt stream could be supported on a 1 MIPS machine (Vax 780). Timesharing was never stopped, but the kernel did get quite a large percentage of CPU cycles. I'll repeat: Just do it. Start. Move forward. The project will teach you the questions you must ask. |