Skip to content

UBCAgroBot/Navi24W2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

60 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Navi24W2

Navigation's 2024 term 2 work.

The goal for this term is to explore navigation fundamentals as well as path planning given a simple map (maybe deep learning as well).

Past work:

From term 1 we came out with some reactive navigation models.

Hough algorithm

CenterRowAlgorithm

This term:

There are three big areas to explore this term:

1. Planning drivable paths:

How can we get Agrobot to get to a point given its constrained movement (Ackermann steering)?

Use this chapter of the textbook to help you with this challenge.
https://link.springer.com/chapter/10.1007/978-3-031-06469-2_4
This chapter gives the intro to the topic and is where the last image comes from:
https://link.springer.com/chapter/10.1007/978-3-031-06469-2_5#Sec21

Drive along line

Drive along points

Go around block

2. More advanced reactive path planning

Bug algorithm:

  • This seems very good for Task 1 bug alpgorithm

Potential Field Algorithm:

  • This seems very good for Task 1. How can we place attracting forces to guide the robot around crop rows?

Random Tree Algorithm (RRT/RRT*):
This resouce is amazing for visualizing RRT: https://demonstrations.wolfram.com/RapidlyExploringRandomTreeRRTAndRRT/

  • This one looks useful for exploring an unknown area for weeds, ie Task 3 for FRE.
  • For a robot to use this algorithm it would need to be tweaked so it does not need to backtrack so much, what is the best way to do this? RRT

3. Image processing

All of these can be found in ch 11 of this textbook:
https://link.springer.com/chapter/10.1007/978-3-031-06469-2_11

More advanced image processing can be found in chapter 14, like stiching images together:
https://link.springer.com/chapter/10.1007/978-3-031-06469-2_14#Sec34

Monadic Operations thresholding, histogram normalization, gamma correction, ...

Edge detection
There is no way this is not useful

Boundary Detection
Same story as Edge detection

Distance transform
Maybe useful for preprocessing, definitely useful for some path planning applications.

Template matching?
Not likely that useful but cool to do image detection without ML

Noise Cleanup
Likely not that useful, but cool

There is also always Gazebo:

Look into: https://github.com/FieldRobotEvent/virtual_maize_field https://github.com/UBCAgroBot/navigation-2019-2023

FRE Resources

FRE resouces can be suprisingly elusive so if you find anything interesting put it here.

FRE Website: https://fieldrobot.nl/event/

FRE 24 Competition guidelines: https://fieldrobot.nl/event/wp-content/uploads/2024/04/Rules_FRE24_V4.pdf

Youtube video with footage of robots completing the challenges:
https://www.youtube.com/watch?v=bHzV4xi0JQk&t=78s

Interesting Links:

Sample code for robotics algorithms:
https://github.com/AtsushiSakai/PythonRobotics

Advanced motion planning, planning under uncertainty, sensor-based planning, reinforcement learning, nonlinear systems, trajectory planning, nonholonomic planning:
http://lavalle.pl/planning

About

Navigation's 2024 term 2 work.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 6