1
00:00:00,650 --> 00:00:05,990
In the last with you, we developed the go to Google algorithm, which should, in theory have the car

2
00:00:05,990 --> 00:00:07,070
reach the exit.

3
00:00:07,400 --> 00:00:12,770
It relied on breaking down the large stars into smaller molecules or configurations, which the board

4
00:00:12,770 --> 00:00:14,690
alternatively need to achieve.

5
00:00:15,320 --> 00:00:17,120
So was this enough?

6
00:00:18,850 --> 00:00:21,010
Unfortunately, not quite.

7
00:00:21,520 --> 00:00:27,640
The problem was that the configuration states that we computed and wanted the robot to achieve was considering

8
00:00:27,640 --> 00:00:31,390
that all stages of the robot navigation worked ideally.

9
00:00:31,870 --> 00:00:34,330
But we don't live in a perfect world, do we?

10
00:00:34,840 --> 00:00:41,080
The robot suffered two problems number one, uneven surfaces, and the errors in the localization and

11
00:00:41,080 --> 00:00:44,860
configuration caused the robot to get stuck in obstacles.

12
00:00:45,100 --> 00:00:51,280
And number two, going at high speeds caused the robot to overshoot its goal and enter the wrong lane.

13
00:00:51,850 --> 00:00:53,770
So how do we address these issues?

14
00:00:54,220 --> 00:00:59,050
Primarily, we require some sort of obstacle navigation that is backtracking.

15
00:00:59,800 --> 00:01:01,690
Now, let's look at what we came up with.

16
00:01:02,830 --> 00:01:08,620
So if the robot is actively solving the maze and no change has been recorded in the angle of set and

17
00:01:08,620 --> 00:01:13,120
the distance for a significant amount of time, then the robot is definitely stuck.

18
00:01:13,870 --> 00:01:20,560
It should backtrack and try again if the next minute goal is closed and it is on top of an obstacle.

19
00:01:20,800 --> 00:01:25,240
Then we need to try next minute goal one that is not too close to the last one.

20
00:01:26,140 --> 00:01:32,230
So let's go try it out and see whether our robot is finally able to solve the more complex maze to.

21
00:01:32,800 --> 00:01:33,580
Fingers crossed.

22
00:01:33,970 --> 00:01:34,720
Let's go.
