PRE2024 3 Group10

The OneDrive can be accessed via https://tuenl-my.sharepoint.com/:f:/g/personal/l_kirkels_student_tue_nl/Ev3UBVOtfktApfNvxowBf-YBFpdR1uAtqIWQzg_Px7CosQ?e=URlCi2

Every week we have divided the task between chairman, minute taker and a wiki updater. The wiki updater updates the OneDrive at the end of the week to the Wiki page.

Group members

Member	Student number	Program
Loes Kirkels	1898477	Applied Physics
Joost Schuurmans	1893297	Applied Physics
Jamin van Amelsvoort	1829998	Applied Physics
Tom Weegels	1883410	Applied Physics
Norah Bouma	1902121	Applied Physics

Problem statement and Objectives

Beaches are beautiful places where a lot of fun is to be had, but they are being terrorized by plastic. On average there is about 20-80 kg plastic per kilometer beach. Without proper cleanups the beaches would be unusable. Luckily there are big beach cleanups like the Boskalis Beach Cleanup which is a tour where all beaches in the Netherlands connected to the North Sea are cleaned by volunteers. Here also various data is collected on how much, what type and at what places the plastics are. These cleanup tours are however mostly done during the summer, when the weather is good. This means the data on plastics on beaches is mostly skewed towards days when the weather is good, but to build the best possible plan of battling the plastic problem, a better representation of the dats is needed. To get a proper representation of the data on plastics on beaches, also the data in worse weather conditions are needed, like in the winter. We will be investigating how data on plastics on the Castricum beach specifically can be gathered effectively and efficiently.

Users and their needs

The primary user is Stichting de Noordzee which is an organization that specializes in cleaning the North sea in the Netherlands, including the beaches. They need a lot of data, over all periods, to get the full picture of the types of plastic, where exactly they are, in which time periods, etc. We have been in contact with them via mail, and are conducting an interview on the 10th of March. In this interview we want to get to know what they are currently doing, and what exactly they need, so maybe focus on a specific type of plastic, a specific weather condition, a specific period of time, etc. After this interview even deeper research will be done in specific elements of a solution of plastic data gathering in bad weather conditions.

Interview tactics:

Primary users:

• With our research, environmental Cleanup Organizations will gain access to new robotic solutions, improving their ability to achieve the goal of removing plastic and debris from the oceans. They require reliable and efficient robot designs that can handle different situations and different kinds of ocean debris. In addition, the solution should be cost effective to fit into their budgets.

• Government agencies gain new tools for ocean cleaning policies and reducing nature pollution, creating public recognition for addressing environmental issues and increasing their support base. They would require robots that comply with national regulations, for example in terms of safety. In addition, they would require enough research data and actual results to justify government funding.

• Non-profit groups would gain examples of current research and solutions for ocean cleanup, to attract possible donors for their organisation. For that, they would require easy-to-understand and transparent results to showcase.

Secondary users:

• Researchers could gain new insights on possibilities to clean up the ocean, helping them to improve their results as well. Therefore, they would require access to our research and data.

• Engineers could get inspired to build the design, or even improve it. Also, they can gain insight into struggles we found, incorporating it into their own work. For this, they would require technical specifications of the design, and the feedback and evaluation of the research.

• Policymakers would gain additional evidence to support their policies and funding decisions. For this, they need accessible reports and results and a low-cost solution.

State of the art Literature research

Drones in bad weather conditions:

Paper 1: A.B. Bello et al. Fixed-Wing UAV Flight Operation under Harsh Weather Conditions: A Case Study in Livingston Island Glaciers, Antarctica. Drones, 6(12), 384; https://doi.org/10.3390/drones6120384, 2022

Drone used: fixed-wing RPAS, in particular, a Trimble UX5 UAV with electric pusher propeller by brushless 700 W motor, chosen for its ability to fly long distances and reach inaccessible areas.

Various hypotheses are proposed and tested, based on the main variables that can affect the flight operations (aircraft stability in flight under gusty winds, behavior of the drone’s materials at low temperatures, battery life, camera configuration to reduce the effects of albedo, etc.). Weather conditions of measurement:

Important variables for drone performance: air temperature, relative humidity, atmospheric pressure, precipitation, wind speed and direction, insolation (zonnestraling), and albedo (lichtweerkaatsingsvermogen).

The general rule of thumb for flying drones is that the wind speed should be no more than two-thirds of the maximum speed of the drone. It has also to be borne in mind that the higher the wind speed, the greater the consumption of battery power.

Hypotheses:

1. withstand lateral winds of up to 50 km/h and gusty winds of up to 15 km/h with a cruise speed of 80 km/h.

1. battery life decreases less than 30% due to cold weather

1. The main materials were high-density polyethylene foam and carbon fiber frame structure and composite elements. Accessory elements, such as the shuttle, also had to be taken into account. None of the materials could severly be affected

1. Camera requirements (not in the scope of this research)

1. the aircraft could travel 60 km in 50 min, reserving 5 min for takeoff and another 5 min for landing.

You cannot fly drones in all weather conditions. à Not too windy or foggy. The reason is that, due to the low temperatures, the fog is formed by ice crystals, which cause Pitot tube obstruction and airframe gelation.

Results: The duration of the batteries was reduced by 30% due to cold conditions, and 5 min must be reserved for return and landing. Biggest problems are wind and rain. The plane was able to fly with constant winds of up to 35 km/h, withstanding gusts of 65 km/h. Because of cold temperatures, the battery live became considerably worse. This implied that flights had to be planned with a reserve of 30%.  In general, the materials were not affected by low temperatures. Due to the low temperatures at flight height there was sometimes icing of the aircraft wings, making them crash.

Paper 2: Gao, M., Hugenholtz, C.H., Fox, T.A. et al. Weather constraints on global drone flyability. Sci Rep 11, 12092 (2021). https://doi.org/10.1038/s41598-021-91325-w  

The study analyzes the impact of weather on drone operability using a 10-year dataset from the ERA5 reanalysis. Two drone classes were considered: Common Drones (CDs), which are inexpensive but have limited weather tolerance (0–40°C, max wind speed 36 km/h, no precipitation tolerance), and Weather-Resistant Drones (WRDs), which are more robust (-20–46°C, max wind 50 km/h, can operate in 50 mm/h rain).

Conclusion: DC’s are very sensitive to rain and seasonal temperature changes, as WRD’s are not really affected by it. Flyability of drones in Netherland’s coastal region should not be a problem with WRD’s.  

Situation in Castricum:

In the winter months it can get up to about -6 degrees celsius in the night, but on average the lowest will be about 2 degrees during the whole day.

During the winter months, the wind speed can go up to about 35 knots or more, which is about 65 km/h or more, so that is very fast.. On average though the wind speed will be about 30-40 km/h.

All the plastic found during a cleanup on the Canticum beach is laid out here. The small pebbles are plastic pellets which are very small in size, about ten times maybe even more smaller than a cigarette bud, so very tiny. The rest are mainly broken pieces of plastic, cigarette buds, or other various types of plastic.

Cameras and Sensors:

(https://www.deltares.nl/en/news/detecting-plastic-pollution-remote-sensing)

Satellites can carry both microwave and optical sensors to detect plastic. Microwave sensors are particularly useful because they can see through clouds and rain, which often hinders optical sensors.  

With microwave sensors, we collect data on the roughness of the water surface

(https://link.springer.com/article/10.1007/s11082-024-06564-8)

A combination of laser induced fluorescence and hyperspectral imaging can be used. LIF can be used to identify between sand, a plastic bag piece, and a wood piece based on the molecular structure differences between plastics, wood, and beach sand.

A cloudy condition could actually improve the working of LIF due to the reduced solar UV interference. However, if the cloud is very thick, it may cause light scattering in the atmosphere, which would slightly reduce the efficiency.  

AI plastic detection:

(https://uu.diva-portal.org/smash/get/diva2:1899954/FULLTEXT01.pdf)

This paper investigates multiple AI learning models on the TACO dataset, which is a database consisting of all kinds of different plastics. They look at three different CNN (Convolutional Neural Network) algorithms, Mask Region-based Convolutional Neural Network (Mask R-CNN), You Only Look Once (YOLOv7), and U-Net, they then use two different machine learning models on these, k-Nearest Neighbors and Logistic Regression, were trained and evaluated. How this roughly works is that you have a dataset, in this case the TACO database consisting of images of all kinds of plastics. The CNN algorithms break down these pictures into very small pieces and look for similarities. These are mostly used for breaking down these pictures which are then classified with a different machine learning algorithm. Now you have classified a lot of pictures and if you use this trained model on a totally new picture it should be able to classify this new picture into one of the classes created by the machine learning algorithms. Their findings are that firstly the TACO dataset is quite imbalanced, favouring some types of plastic a lot more than others, a lot of pictures of cigarettes and plastic bags/wrappers, while straws, styrofoam and aluminium foil are very limiting. The confusion matrices of the trained models with Mask R-CNN and YOLOv7 can be seen below. The ones in the diagonal row from top left to bottom right, are correctly classified objects of a specific class.

An example of how the model detects plastic can be seen below, this is the YOLOv7 trained model. Boxes are roughly put around the object, a class is assigned to them with some specific probability of it being correct based on the data it has been trained on. So, if we figure out what exactly we want to detect it can be solely trained on this specific aspect to be very good at this one thing. The k-NN trained model had a total accuracy of 64%, while being very good in detecting bottles (82%), and plastic film (74%). The logistic regression model had a overall accuracy of 73%. So the logistic regression model is much more balanced than the k-NN trained model

They state that the current detection with these trained models is okay, but not far from perfect. More types of different litter categories are needed with a lot more data for some categories, because the data is skewed right now.

Planning

	Approach	Milestones and deliverables
Week 1	-Make planning with milestones and deliverables (everyone) -Literature study (5 articles per person), summarized on the wiki (state-of-the-art research) -Make problem statement and objectives (Jamin) -Write down users and what they require (Joost)	Problem statements and objectives Plan (subject, objectives, users)
Week 2	-Make a state-of-the-art from the summarized articles (Norah) -Think of the requirement and goals the robot should have (Loes&Tom) -Do further research (Joost) -Do a user study (Jamin)	State-of-the-art Concrete ideas for a concept
Week 3	-Write down the research in the wiki (Joost) -Investigate the specifics of the robot (Jamin&Loes) -Write the user study in the wiki (Tom) -Write down the process of approach in the wiki (Norah)	Finished research Concept of the robot Finished user study
Week 4	-Design concept of the robot (Loes) -Write the research and results in the wiki (Joost) -Start designing the model of the robot (Loes&Norah)	Research and results in the wiki
Week 5	-Finish the model of the robot (Loes&Norah) -Finish the wiki (Joost&Jamin&Tom)	Finalized model Finalization of research, everything written down in the wiki
Week 6	-Make the presentation -Room for things that still need to be done	Presentation
Week 7	Preparation for final presentation

Weekly updates

	Loes	Joost	Jamin	Tom	Norah	General
Week 1		Research papers (4h)				Monday morning lecture + discussion (2h), Wednesday morning working session (4h)
Week 2						Monday morning meeting (1h), monday afternoon meeting (4h), contact users (4h), generalize and research the subject (2h)
Week 3		Research drones in bad weather conditions (4h)