1
00:00:04,210 --> 00:00:08,950
In this video, we're going to have a quick look at the simulation exercises and examples that we're

2
00:00:08,950 --> 00:00:12,740
going to be using for the unscented coming for a portion of this course.

3
00:00:13,360 --> 00:00:18,190
So just like in the previous section when we used an extended family photo, we're going to be using

4
00:00:18,190 --> 00:00:21,220
the same vehicle model and process model, the same problem.

5
00:00:21,760 --> 00:00:28,200
However, we're going to be solving it using an unscented computer rather than an extended common filter.

6
00:00:29,650 --> 00:00:33,520
So the 2D vehicle filter is the same as follows.

7
00:00:33,520 --> 00:00:40,040
The goal is to estimate the position, velocity and orientation of a moving vehicle based on Leida like

8
00:00:40,060 --> 00:00:45,760
measurements to known landmarks and features, also using GPS measurements and gyroscope measurements

9
00:00:45,760 --> 00:00:46,540
of the vehicle.

10
00:00:49,040 --> 00:00:53,570
Now, the problem is that we're going to assume that the vehicle can travel at a constant speed in a

11
00:00:53,570 --> 00:00:59,720
2D explained in the direction it's facing, and we're going to let the inputs of the system be the turn

12
00:00:59,720 --> 00:01:01,450
right or the vehicle from a gyroscope.

13
00:01:01,850 --> 00:01:06,250
And we're going to assume that the acceleration can be modeled as a random variable.

14
00:01:06,260 --> 00:01:07,730
And this is the process model.

15
00:01:08,060 --> 00:01:12,650
So this is exactly the same as the previous model for the extended common filter.

16
00:01:13,780 --> 00:01:18,730
We will assume that we get positional measurements of the vehicle's location from a GPS like sensor,

17
00:01:18,740 --> 00:01:20,500
and this is going to be the first sensor model.

18
00:01:21,070 --> 00:01:23,970
Now, when we use this in, this is a linear model.

19
00:01:23,980 --> 00:01:29,080
So we're just going to be using the same equations as we did for the linear common filter and the extend

20
00:01:29,080 --> 00:01:29,810
the common filter.

21
00:01:30,130 --> 00:01:36,100
We don't need to do any special handling or anything different for the unscented method when it's only

22
00:01:36,100 --> 00:01:37,450
a linear measurement model.

23
00:01:39,520 --> 00:01:45,190
We also ensuring that we get a number of range and relative bearing measurements of known landmark locations,

24
00:01:45,310 --> 00:01:50,680
i.e. this is from the ladylike sense and measurement model, and this is going to be the second measurement

25
00:01:50,680 --> 00:01:54,130
model that we implement inside the unscented common field.

26
00:01:54,880 --> 00:01:59,320
So basically, we're going to implement the process model, the prediction step and the update step

27
00:01:59,320 --> 00:02:03,790
for the second sensor model for the Leida using the antenna filter.

28
00:02:05,670 --> 00:02:11,400
So just like in all the previous exercises, the first step to get the simulation set up for the intended

29
00:02:11,400 --> 00:02:17,790
coming filter is to take the common to underscore UCP understood, a student found and we want to rename

30
00:02:17,790 --> 00:02:19,900
it to Come and Filter Seip.

31
00:02:20,340 --> 00:02:26,100
So now we're going to be using this base file for the exercises in this following section.

32
00:02:27,290 --> 00:02:33,050
And just like before, if you want a working example to play around with it, you can take this top

33
00:02:33,050 --> 00:02:33,550
fall here.

34
00:02:33,560 --> 00:02:39,110
The common theme to underscore U.K. if underscore answer, rename this to come in here and compile and

35
00:02:39,110 --> 00:02:39,470
run that.

36
00:02:39,470 --> 00:02:42,800
And that will give you a working version to compare against.

37
00:02:44,060 --> 00:02:49,610
So now let's have a look at a working simulation, what the final result of all these exercises should

38
00:02:49,610 --> 00:02:53,960
be once you program them of the uncynical method of running the simulation.

39
00:02:55,000 --> 00:03:00,640
So here's a working example of the onset and common filter, and it's running the motion profile which

40
00:03:00,640 --> 00:03:06,460
tests all the different senses at a constant speed, this simulation should give similar results as

41
00:03:06,460 --> 00:03:11,500
the extent to come and filter as they both operate on the same process model and they're both operating

42
00:03:11,500 --> 00:03:12,970
on the nonlinear system.