Real robots don’t drive straight
This is the title of a publication by Fred G. Martin of the Computer Science department of the University of Massachusetts in Lowell.
Basically, what he is trying to tell us is that sometimes we tend to cut corners and hide the interesting stuff from our students, so that they do not fully learn to solve the problems that are inherent to programming robots in a real world environment. When we learn students elementary things like ‘turn left’, turn right’, ‘turn around’, ‘forward n rotations’, then they will not learn anything but programming things in the right order. Which is fine if that’s our goal, but it won’t teach them to have their robots react in the real world where things tend to be fluid instead of fixed. In short, the whole subject of seemingly intelligent interaction gets lost in this kind of programming.
So, when we want a robot to drive straight, we have to come up with feedback of some sort, i.e. from a gyro sensor or a GPS module and have the students puzzle on that! From my experience they can do that. It takes time and guidance, but boy, the satisfaction that is achieved when they manage to do that!
It is an article that can help us guide them:
Abstract of the article
Over last fifteen years, robot technology has become popular in classrooms across our whole educational sys- tem. Both engineering and AI educators have developed ways to integrate robots into their teaching. Engineering educators are primarily concerned with engineering science (e.g., feedback control) and process (e.g., design skills). AI educators have different goals—namely, AI educators want students to learn AI concepts. Both agree that students are enthusiastic about working with robots, and in both cases, the pedagogical challenge is to develop robotics technology and provide classroom assignments that highlight key ideas in the respective field. Mobile robots are particularly intriguing because of their dual nature as both deterministic machines and unpredictable entities. This paper explores challenges for both engineering and AI educators as robot toolkits evolve.