ASN1

The Pitch!

Deadline


September 13, 2021.



What you need to do?


The first assignment is to write a document that pitches the project you'll work on for the rest of your semester. Formally, the document should entail the project scope and the contributions you will make in the context to the state-of-the-art. You can change the project from what you pitched in your last semester.

Step 1: Finding an idea


Start with a cool idea or a problem you think is immediately required to be tackled for the greater good. Now search for research papers with related keywords on Google Scholar. Papers in top conferences/journals are generally of higher quality than the lower ones. The obvoius question now is, how do you know what is a top conference/journal? You can refer to commonly used metrics such as h-index, i-10 index or g-index, information about these factors can be found here. To make life easier, Google Scholar also has a top publications list and can be found here. You can filter by area to find the top conferences/journals. Some common publishers known to have top quality papers in robotics are Nature, Science, IEEE, ACM, Springer, Elesevier among others.

Step 2: Breaking down idea into smaller chunks


Now, break down the idea into smaller and more managable chunks by thinking about what the problem entails. In robotics, the standard components include: perception (how your robot sees, feels or precieves the world), planning (how do you plan to go from point A to B, plan next task and so on), control (how do you send control the speeds of your motors or other acutators to achieve the motion you want) and cognition/learning (how do you learn something to be better next time). Feel free to be creative to blend the boundaries among these domains to make them task-specific. Once you have an idea of the input and output for each of the domains, you can start looking for research papers which tackle similar problems in the components you modelled (this might be in a completely different research problem).

Step 3: Look for solutions to the problem you want to solve


In this step, you are deliberatly searching for solutions to the problem you had in mind. The questions to ask are: has anyone done this before? Is this problem even solvable? Should I simplify the problem further? If the problem has been tackled by literature, what is remaining? Make a note of all the related papers here.

Step 4: Modify problem statement


From the previous step, you might notice that the problem you had in mind is either tackled extensively by literature or maybe its too hard to solve. Modify your problem statement by adding/removing in additional constriants to make the prpblem harder/easier. The former is done when the problem is studied extensively in literature and you try to tackle problems (failure cases) encountered by previous solutions. The later is done when the problem you are trying to solve might not be possible in a semester or is just too hard. While adding in additional constraints make sure that they are sensible and plausible in the real-world.

Step 5: Brainstorm ideas


Think of potential ways to solve your modified problem. Try to re-use as many things as you can (save your time and be efficient). Nobody gives you credit for re-inventing the wheel. Look at licences carefully if you want to use somebody else's code for your research. Here is a good description of licences. Find the resources you can potentially use so you have a starting point. Carefully think about the software/hardware tools you need to use. Plan ahead of time.

What you need to submit?


A report in double column IEEE format (maximum of 6 pages) in LaTeX per group. Your report should include the following:

You can use any LaTeX tool you like, I would recommend Overleaf as it is free and you can use it to collaborate online. If you are new to LaTeX, you can learn from this awesome tutorial from Overleaf. Alternatively, if you want to work locally, you can use any of these editors. The LaTeX template can be found here. The submission should be made through ELMS with the name ASN1.