Subject RV

study ITshoe deformations using finite elements. Validade the wearable system in terms of mechanical resistance and properties. it is also an asset not only related to the system defense but also to have a better interpertation on how the forces are being translated inside the wearable system contributing this way to possible future enhancement of the system as well as an easy interpretation or analysis of the force that that the system provides to the user. 


shoe model for NAO,
appropriated method to mesh the cad
boundaries, degrees of freedom representation. 

force representation robot- shoe - floor
evaluation against material properties example. how much going to sides. different frictions would represent different forces propgation, can this be modeled. 
like to use C Cpp or python to do it other than use a software like abaqus, is it possible, and what are the steps to achieve it . 
 
abstract goals set from fct application, left with element finit methodoligy and application.

How can this be usefull? and how can this be prepared to be continued on RS subject 2nd semester. talk with both to see if its possible
maybe simulation by means of software to predict the behavior and then adjust calibration by means of software of the work itself.


sub,j,t RM  

create a database and follow readins real-time to this database. with a constructed well thought structure; ;; full of data with different exemplary floors:
different gait parameter-


the manipulation of this data is still an issue that needs to be well studied so that the further machine learning can be simplified without the need to overcome all the noise and faulty hardware that would easily overload the learning system, which contributes to a not optimized network. nowadays inside the learning methodologies we have the deep learning which requires thousand of datasets with the addition to accept and re-adjust further everytime the system is being used for the purpose.

gather thousands of dats with different gaits and floors with a proper analysis can be usefull for this project once, the real results and contribution come as better as the data is interpreted, on the side of the machine but also for us as programmers or users. 

sub VS
main 3 goals:

enhance the wearable system
create a controller to take actions in order to improve robot mobility over different floors, by different we define as different friction coeficients, not considering deformable materials nor obstacles.

the learning system is the most interesting as well as the most challanging task to be done, because it is the point where the system becames usefull to the society. having the tools without having the way to make the tool usefull isnt what we trying to achieve. and so, deep study over learning systems is where we want to move foward on the last goal. 

elaborate whyY, sota to sustain and fundament this work . 

1- record some steps with different ITshoe placement on the robot using the same gait parameters;
2- For F1uptoF8 check the initial pressure as well as the progression while walking. 
3- if F1toF8 look alike for several experiences, %e < 5%
4- record data for N floors
5- train NN
6- test real.time g%+80
7- 4to7 changing N gait parameters
8- if NN not enough moving foward to extreme learning or deep learning.
 


SUB,j,t: VS

Controller C/C++ or python - external perturbance agst ability to recover. 
later use. NN detect floor and controler adjust to fulfill the need to walk over it. 

: how to control, Need to have acess to joints, or somehow simplify into the extremely necessary joints to mantein robot's stability while walking over sliperry floors.