PRE2023 3 Group3

Group members

Name	Student ID	Current Study Programme
Patryk Stefanski	1830872	CSE
Raul Sanchez Flores	1844512	CSE
Briana Isaila	1785923	CSE
Raul Hernandez Lopez	1833197	CSE
Ilie Rareş Alexandru	1805975	CSE
Ania Barbulescu	1823612	CSE

Problem statement and objectives

Taking care of gardens requires a lot of time and effort. A robot could automate this chore to leave more room for

Deliverables

• A NetLogo environment which implements the algorithm that was found to be most optimal and appropriate for the robot's task, this will aid in confirming that the implementation works and is effective in a simulated environment of a garden.
• A survey conducted on the user group which will be most likely to use the robot, in the survey we hope to identify the most important requirements the robot should implement and features it could have. The survey will also allow for a general idea of the robot's design to be established from the users preferences.
• Interactive UI of an app which will allow the user to control the robot remotely that implements the user requirements that we will obtain from the survey. The UI will be able to be run on a phone and all its features will be able to be accessed through a mobile application.
• Create a design for the robot, confirming which parts are essential and needed in order to visualize a potential design and estimate the cost of the robot.

Work Distribution

Navigation Algorithm and Netlogo Implementation: Patryk

Week	Task	Deliverable to meeting
2
3	Complete initial NetLogo Enviroment
4
5
6
7
8

Navigation Algorithm and Netlogo Implementation: Raul S.

Week	Task	Deliverable to meeting
2
3	euhfei
4
5
6
7
8

Robot Design and Product Cost Estimation: Briana

Week	Task	Deliverable to meeting
2
3
4
5
6
7
8

Robot Design and Product Cost Estimation: Rareş

Week	Task	Deliverable to meeting
2
3
4
5
6
7
8

Interactive UI design and implementation: Raul H.

Week	Task	Deliverable to meeting
2
3
4
5
6
7
8

Interactive UI design and implementation: Ania

Week	Task	Deliverable to meeting
2
3
4
5
6
7
8

General Group Plan

Week 1	Week 2	Week 3	Week 4	Week 5	Week 6	Week 7
Brainstorm initial idea Complete literature review

Individual effort

		Break-down of hours	Total Hours Spend
Week 1	Patryk Stefanski	Attended kick-off (2h), Research into subject idea (2h), Reading Literature (2h), Updating wiki (1h)	7
	Raul Sanchez Flores	Attended kick-off (2h)	2
	Briana Isaila	Attended kick-off (2h)	2
	Raul Hernandez Lopez
	Ilie Rareş Alexandru
	Ania Barbulescu	Attended kick-off (2h), Reading Literature (2h)	4
Week 2	Patryk Stefanski	Meeting with tutors (0.5h), Researched and found contact person who maintains Dommel (1h), Group meeting Thursday (1.5h), Created list of possible deliverables (1h)	4
	Raul Sanchez Flores	Meeting with tutors (0.5h), Group meeting Thursday (1.5h)	2
	Briana Isaila	Group meeting Thursday (1.5h)	1.5
	Raul Hernandez Lopez	Meeting with tutors (0.5h), Group meeting Thursday (1.5h)	2
	Ilie Rareş Alexandru	Meeting with tutors (0.5h), Group meeting Thursday (1.5h)	2
	Ania Barbulescu	Group meeting Friday (4h), Research Literature (2h)	6

Literature Review

1.     TrimBot2020: an outdoor robot for automatic gardening (https://www.researchgate.net/publication/324245899_TrimBot2020_an_outdoor_robot_for_automatic_gardening)

·      The TrimBot2020 program aims to build a prototype of the world’s first outdoor robot for automatic bush trimming and rose pruning.

·      State of the art: ‘green thumb’ robots used for automatic planting and harvesting.

·      Gardens pose a variety of hurdles for autonomous systems by virtue of being dynamic environments: natural growth of plants and flowers, variable lighting conditions, as well as varying weather conditions all influence the appearance of objects in the environment.

·      Additionally, the terrain is often uneven and contains areas that are difficult for a robot to navigate, such as those made of pebbles or woodchips.

·      The design of the TrimBot2020 is based on the Bosch Indigo lawn mower, on which a Kinova robotic arm is then mounted. (It might, therefore, be worthwhile to research both of these technologies.)

·      The robot’s vision system consists of five pairs of stereo cameras arranged such that they offer a 360◦ view of the environment. Additionally, each stereo pair is comprised of one RGB camera and one grayscale camera.

·      The robot uses a Simultaneous Localization and Mapping (SLAM) system in order to move through the garden. The system is responsible for simultaneously estimating a 3D map of the garden in the form of a sparse point cloud and the position of the robot in respect to the resulting 3D map.

·      For understanding the environment and operating the robotic arm, TrimBot2020 has developed algorithms for disparity computation from monocular images and from stereo images, based on convolutional neural networks, 3D plane labeling and trinocular matching with baseline recovery. An algorithm for optical flow estimation was also developed, based on a multi-stage CNN approach with iterative refinement of its own predictions.

2.     Robots in the Garden: Artificial Intelligence and Adaptive Landscapes (https://www.researchgate.net/publication/370949019_Robots_in_the_Garden_Artificial_Intelligence_and_Adaptive_Landscapes)

·      FarmBot is a California-based firm and designs and markets open-source commercial gardening robots, and develops web applications for users to interface with these robots.

·      These robots employ interchangeable tool heads to rake soil, plant seeds, water plants, and weed. They are highly customizable: users can design and replace most parts to suit their individual needs. In addition to that, FarmBot’s code is open source, allowing users to customize it through an online web app.

·      Initially, the user describes the garden’s contents to a FarmBot as a simple placement of plants from the provided plant dictionary on a garden map, a two-dimensional grid visualized by the web app. FarmBot stores the location of each plant as a datapoint (x, y) on that map. Other emerging plants, if detected by the camera, are treated uniformly as weeds that should be managed by the robot.

·      The robotic vision system employed by the Ecological Laboratory for Urban Agriculture consists of AI cameras that process images with OpenCV, an open-source computer vision and machine learning software library. This library provides machine learning algorithms, including pre-trained deep neural network modules that can be modified and used for specific tasks, such as measuring plant canopy coverage and plant height.

3.      Indoor Robot Gardening: Design and Implementation (https://www.researchgate.net/publication/225485587_Indoor_robot_gardening_Design_and_implementation)

·      

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.