-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathreadme.txt
43 lines (25 loc) · 2.04 KB
/
readme.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
# Autonomous
# Abstract
An oscillating lidar mount to extract raw 2-D maping of environment to be further used in SLAM and obstracle avoidance.
It is a team project aimed at making a self-driving bot capable of following a defined path while avoiding obstacles.
Created a Rotatory 2-D mapping device using a point LiDAR mounted on a arduino controlled Motor-encoder pair.
Implemented ICP SLAM algorithm on the collective data from encoder and LiDAR to create a real-time global Map.
# Hardware
1. dc motor
2. encoder
3. lidar
4. motor driver l298
# Software
python
cv2
ICP source: https://engineering.purdue.edu/kak/distICP/ICP-2.1.1.html
# Rotating Lidar
Basically the lidar gives distance data of the point straint infront of it, the lidar is the mounted over a dc motor and oscillated under feedback from encoder, when rotated we get the data of encoder and coresponding distance, this data is sent serially from arduino to python on windows system serially, On python this data as polar cordinate is then converted and plotted on a cartesian system, this instantaneous map is made through some filters and then to be further used for SLAM and obstracle avoidance algorithms.
# SLAM
Once we have the instantaneous map the next step to apply SLAM is to create a GLOBAL map by "stitch" with succedding instantaneous maps. Our first attempt was to extract points as features, using hough line transform of cv and twerking its parameters points were obtained but we switched to ICP(Iterative Closed Point) algorithm which is standard for SLAM application. we used an already packaged version of ICP, due to poor quality data of lidar, ICP was tested on dummy images prepared by us. results are attached.
# Limitations
Poor quality data from Lidar
ICP works only for small shifts in maps, thus reqireing high speed lidar data extraction and processing or limiting the speed of bot.
# Applications
object avoidance in autonomous vehicles, closed room mapping for indoor bots etc