1
00:00:00,060 --> 00:00:04,770
In previous section, we have learned how to generate trajectory for the robotic manipulators.

2
00:00:05,160 --> 00:00:12,030
And this was pretty interesting actually how we didn't learn trajectory planning in order to plan and

3
00:00:12,030 --> 00:00:14,550
plot them and look how beautiful they are.

4
00:00:14,970 --> 00:00:16,350
This is not a drawing lesson.

5
00:00:16,410 --> 00:00:22,230
We have learned them in order to do trajectory planning and then make a robot manipulator to track these

6
00:00:22,230 --> 00:00:22,980
trajectories.

7
00:00:23,640 --> 00:00:26,670
We have seen the first part, namely trajectory planning.

8
00:00:26,680 --> 00:00:29,460
Now it's time to learn how to do the second part.

9
00:00:29,850 --> 00:00:36,140
Namely, in this second part, we will riot control algorithms in order to make robot manipulator track

10
00:00:36,150 --> 00:00:40,950
to give trajectory with as small errors as possible, ideally zero.

11
00:00:42,390 --> 00:00:45,990
So before going on, let's see some terminologies.

12
00:00:46,260 --> 00:00:50,850
First of all, we can categorize controller into controllers in two ways.

13
00:00:51,300 --> 00:00:54,450
The centralized control algorithms and centralized ones.

14
00:00:54,840 --> 00:01:01,890
The difference between them is the centralized control algorithms take each one as individual and try

15
00:01:01,900 --> 00:01:08,070
to control it separately and accept more minorities and couplings in robot dynamics as disturbances.

16
00:01:08,490 --> 00:01:10,800
And they are tried to be rejected.

17
00:01:11,160 --> 00:01:14,420
This assumption can be accepted under low velocities.

18
00:01:14,430 --> 00:01:22,320
However, as the velocity increases, the coupling effects in robot dynamics increases in that level

19
00:01:22,470 --> 00:01:25,710
that it cannot be taken as a disturbance to be rejected.

20
00:01:26,040 --> 00:01:31,080
Indeed, there has to be some method to handle them at this moment.

21
00:01:31,090 --> 00:01:33,570
Centralized control methods come to the help.

22
00:01:33,930 --> 00:01:42,240
They don't take each joint individually but try to provide control to the whole robot together, so

23
00:01:42,240 --> 00:01:46,920
it takes into account more minorities and the couplings in robot dynamic model.

24
00:01:47,370 --> 00:01:51,960
Because of that, centralized metal, centralized control methods are superior.

25
00:01:52,350 --> 00:01:57,690
However, as we will see in the future, these methods will also not solve the problem completely.

26
00:01:58,170 --> 00:02:04,350
This is because while they are considering robot dynamics model non-linear this uncoupling, we assume

27
00:02:04,350 --> 00:02:10,920
that we have perfect model, which models disturbances perfectly with the perfect knowledge of dynamic

28
00:02:10,920 --> 00:02:11,640
parameters.

29
00:02:11,970 --> 00:02:17,370
Additionally, we don't take into account external disturbances that affect the robot manipulator.

30
00:02:17,760 --> 00:02:25,140
All these factors will affect the stability of the robot control algorithm, which we will see in future

31
00:02:25,140 --> 00:02:25,650
lessons.

32
00:02:26,130 --> 00:02:33,060
In order to solve these issues, we will introduce advanced control algorithms such as robust control

33
00:02:33,060 --> 00:02:35,120
and adaptive control method.

34
00:02:35,160 --> 00:02:41,040
They will take into account external disturbances and uncertainties in the knowledge of dynamic parameters.

35
00:02:41,640 --> 00:02:48,450
Robust control will try to reject disturbances and parameter uncertainties with high energy input,

36
00:02:48,690 --> 00:02:56,280
while adaptive control will accept disturbances as changes in model dynamic parameters and so control

37
00:02:56,280 --> 00:02:56,970
parameters.

38
00:02:56,970 --> 00:03:01,410
So control parameters will be adapted based on these changes.

39
00:03:01,920 --> 00:03:06,000
So without further ado, let's jump into the robot control methods.