1. Project number and Title – 3 - Developmentally Learning the Support Affordance of a Platform

2. Should this project be considered for the Best Project award? No

3. Should this project be considered for the top 3 project awards? No

4. On a scale of 1 to 10, how would you rate the overall organization/clarity
of the project report?  9

5. On a scale of 1 to 10, how would you rate the overall project idea?  7.5

6. On a scale of 1 to 10, how would you rate the overall research
contribution of the project idea, methodology and/or results?  8


* Overall, is the project report clear, concise, and well-organized?
This proposal is very clear and well organized. The sections are concise and flow logically to conclusion. Some details are repeated, but they serve more to emphasize and are not jarring.
      
     * How does the project idea and methodology fit within the framework
       of Developmental Robotics?
The team is aiming to solve some fundamental problems that would enable robots to perform general tasks in the future. Their method of developmental learning, that is, letting the robot explore and discover for itself what support really means is fully within framework of Developmental Robotics. This would help the robot develop and define its own schema which it would use for and modify when exposed to novel objects or situations.

* Describe what you like BEST about the project?
  I like the idea of robot learning shape affordances the best, it seems really cool; and the report is also well – organized. 

* Describe what you like LEAST about the project?
The use of the ramp, while necessary to track the object and make sure it remains in the visual space; it introduces some unknowns like friction into the equation. The friction might cause the object to stop or move slowly; the shape of the object is also a factor in making it move faster, for example, a ball will roll down while a block of wood will slide slower.
 
* Do the methods, results and contributions of the final project correspond to what was presented in the initial project proposal?
One implementation detail in the proposal was a construction of an independent platform for the ramp. But due to limitations of the robot’s end-affector space, the setup was changed to include a standard office table and a removable wooden board which could be adjusted to fit many configurations.
While in the proposal it was stated that they would be able to determine the distance that each object could be pushed without it falling off the table, the slow frame rate of data at the platform edge made this infeasible. So, the experiment was changed to include platform variants and a modified classifier.

* Are there any major details left out with regards to the methods, algorithms, or experimental design described in the report?

No, all the necessary details are well explained by the authors. They have provided implementation details along with algorithms used and the resulting experimental conclusions.

* Do the experimental results reported in the paper demonstrate success?
The project shows success in the effectiveness of edge detection but learning affordances of a platforms edge was more difficult to quantify. The team applied data analysis for only two positions of the platform which showed clear depth values and co-ordinate positions. All three classification algorithms returned results that were significant letter than chance. As an alternative, self-detection was also explored to classify control. Here, the object is controlled only if the robot considers it as part of its body. This also returned data consistent with the prior classifications.
* Do you have any suggestions for improvement and future work?
One obvious extension is to apply the algorithms and test the data from all the configurations used for the experiment. The authors have already planned for a new classification that can be applied for different types of measurement for other modalities.

* How close is the final project report to being publishable as a conference or journal paper (consider the research papers that were part of the course reading)? What would it take to get there?
The paper looks pretty good right now. With the above mentioned future work completed, the author’s will have inarguable results that will corroborate the ones found so far.