ryan-keenan
commented
7 years ago Thank you for this feedback! I agree. Originally I set out to build a simple first week project but it just kept getting more and more complicated as we made it cooler! The current plan is to have mapping of 40% of the terrain at 60% fidelity and finding at least 1 rock be the minimal passing submission. But I've now created a page called Requirements & Challenges listing tips and tricks for taking on the optional challenge of finding and picking up all the rocks and making a high quality map. Still trying to figure out what kind of prize we can offer to students who solve the challenge or come up with the best result.
867sec
arek 867sec@gmail.com "need to define metric" in writeup template There is "need to define metric" in writeup template in sentence: "Iterate on your perception and decision function until your rover does a reasonable (need to define metric) job of navigating and mapping." under "Autonomous Navigation / Mapping".
My suggestion is to make the project 1: Search and Sample Return into 2 projects: the first the way it is now and the second to collect all the yellow rocks in the autonomous mode. Clear metrics give tremendous motivation to students and are deeply satisfying fostering good learning habits and addiction to Udacity. The way the Project 1 Search and Sample Return is structured now it doesn't provide enough motivation for student to put effort into creating artificial intelligence algorithm (decision tree) to achieve anything in autonomous mode because nothing is required in the rubric. Adding it to the rubric may not work either because the project is already big therefore the idea of splitting it into 2 parts. I'm so excited about the prospect of having a project with clearly defined goal: pick up all rocks using autonomous mode! Sounds so exciting that I'm sure the students will greatly appreciate it because it will teach how hard it is to hard code all the decision, will train in thinking about all the edge cases etc, etc.