Robotics: Science and Systems (RSS)
Response to 2014/15 survey feedback.
The RSS course is running in 2015/16.
We have taken note of the feedback provided through the Student Online Survey. We have the following responses/clarifications:
- The extensive comments include as a major theme the time commitment required of the students and the difficulty of some of the milestones. As we have said in response to earlier year's comments, the need for substantial time investment in robotic experiments, perhaps more so than assignments in other courses, is not something that can be easily avoided. One of the objectives of this course is indeed to expose students to this aspect of the field. That said, we are reviewing aspects of the organisation of the practicals, and the lectures, which would hopefully alleviate some of this burden. This includes the following points below:
- We will reduce the number of milestones, making each of them significant but with sufficient time to do the required work despite the pressures of other modules, assignments and so on.
- We will be replacing C++ with Python for both vision, using OpenCV wrapper, and control, using phidgets wrapper (brand of the I/O hardware used), to ease the burden of dealing with real-time programming issues in C++. Additionally, we have installed the python libraries on the DICE machines, and will provide a step-by-step guide to install these on the students' own laptop computers, which will enable the students to develop and debug their code faster.
- We will provide a skeleton practical application (a sandbox), with documentation and examples. This sandbox provides low level interfaces that students have found time consuming to implement in the past. We will also run short 10 minute tutorials/discussions during every week when the practicals are being worked on, to demonstrate different aspects of the sandbox or to discuss implementation of useful algorithms.
- For the vision component of the course, we will add demonstrations using the openCV software which is also available on the robots.
- The timing of lectures will be revised so that topics needed in practicals will be covered earlier in the semester.
- Overall, we recognise that this course covers a lot of ground. In keeping with its role as an introductory and compulsory course, we will eliminate some of the more specialised topics, attempting to replace them with a more in depth discussion of key algorithmic ideas that will also be reinforced in the practicals.
Subramanian Ramamoorthy, Sethu Vijayakumar, Chris Williams, September 2015.