Post Reply 
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Asymmetrical robot
07-08-2016, 10:52 AM
Post: #1
Asymmetrical robot
Hi everybody!

I have ordered the electronic kit for building the B-ROBOT for my kids. Ia want to make the body by myself, with a different design.

I have one doubt. I wanted to make an asymmetrical design of the body, so the center of gravity could not be in the vertical of the stepper motor axis (may be a bit forward, may be a bit backwards, I still do not know because I have not finished the designed...). The question is that being unbalanced in vertical position could be a problem to keep balance.

I have seen that more or less all design are quite symmetric, but I suppose depending on components distribution inside de body the balance would also not be 100% perfect! So in the end I imagine there won't be too much difference. And also the motors will be driven by the IMU sensors measuments to compensate that...

I think i am answering myself... Big GrinBig GrinBig Grin

Anyway, comments from experienced memebers will be apreciated.

Regards,

Sergio
Find all posts by this user
Quote this message in a reply
07-12-2016, 09:12 AM
Post: #2
RE: Asymmetrical robot
It should not be a problem. But you will have to adjust a little the weights and yes... you are answering yourself Tongue
Find all posts by this user
Quote this message in a reply
07-12-2016, 10:53 AM
Post: #3
RE: Asymmetrical robot
(07-12-2016 09:12 AM)JohnQ Wrote:  It should not be a problem. But you will have to adjust a little the weights and yes... you are answering yourself Tongue

Big GrinBig GrinBig GrinBig Grin

Ok. Thanks JohnQ. Anyway, I am also following another thread from the forum which is actually having problems in this way.

I am looking forward to get all parts and start the project!
Find all posts by this user
Quote this message in a reply
07-12-2016, 08:32 PM
Post: #4
RE: Asymmetrical robot
RU
Попробую объяснить суть проблемы образно, с привязкой к постоянному и переменному току.
Все процессы ПИД-регулирования симметричны относительно "земли". Это означает, что в случае идеального баланса программный интегратор ошибки регулирования на выходе выдает ноль. В этом случае положительная полуволна переменного тока компенсируется отрицательной полуволной и моторы робота с одинаковой затратой энергии компенсируют наклоны робота как вперёд, так и назад.
В случае же смещённого центра тяжести, когда полуволны не компенсируются, накапливается интегральная ошибка, которая лчень быстро достигает программного ограничения.
Какие могу дать рецепты?
Первое и самое правильное, обрабатывать положительные и отрицательные ошибки с различными ПИД-коэффициентами, которые учитывали бы дисбаланс.
Второе решение:
- раздельная фильтрация полуволн (нахождение среднего за несколько периодов обработки);
- вычисление разности между абсолютными значениями разнополярных усреднённых ошибок;
- на базе полученной разности и её соотношения с абсолютными биполярными значениями рассчитываются поправочные коэффициенты для каждой из полуволн соответственно;
- текущие значения ошибки регулирования умножаются а зависимости от знака ошибки на соответствующий коэффициент.
Иными словами, в программном коде банально не хватает высокочастотного фильтра или "разделительного конденсатора", если привязываться к электрическим цепям.

EN by Google
Let me explain the problem figuratively, with reference to both AC and DC.
All PID control processes are symmetric with respect to "ground". This means that in case of perfect balance adjusting program error integrator provides a zero output. In this case, the positive half-wave of the alternating current is offset by the negative half-wave, and the robot motors with the same energy expenditure as a compensated robot tilts forward and backward.
In the case of a displaced center of gravity, when the half-wave are not compensated, accumulated integrated error that lchen quickly reaches the software limit.
What can give the recipes?
The first and most correct, handle positive and negative errors with different PID coefficients that take into account the imbalance.
The second solution:
- Separate filtering half-waves (being the average of the number of treatment periods);
- Calculation of the difference between the absolute values ​​of bipolar averaged error;
- On the basis of this difference and its relations with bipolar absolute values ​​calculated correction factors for each of the half-waves, respectively;
- The current values ​​of the loop error and multiplied depending on the sign of the error by the appropriate factor.
In other words, the code is simply not enough high-pass filter or "capacitor" when attached to electrical circuits.
Find all posts by this user
Quote this message in a reply
07-15-2016, 07:56 PM
Post: #5
RE: Asymmetrical robot
I thinķ there is no problem with asymmetric desigh. Only You need is balance them with component placing or additional weights.
Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump:


User(s) browsing this thread: 1 Guest(s)