PRE2018 3 Group17

Group members

Group Members	Student nr.
Diederik Geertsen	1256521
Cornelis Peter Hiemstra	0958497
Joël Peeters	0939193
Benn Proper	0959190
Laila Zouhair	1260529

Ideas

• Omnidirectional 3D printer
• Breakfast bot
• Robot to remove microplastics from water
• Clothes folding robot
• Building guidance robot
• Disaster observation drones

Disaster observation using a network of drones

Problem statement and objectives

The search for survivors after disasters such as floods and earthquakes is an extremely time sensitive matter. Fast localisation and retrieval of victims is essential for increasing surviving chances, which creates an obvious application for automated systems.

The main goal of the design is to be able to identify the high priority zones following a natural disaster. These high priority zones could for example be the area with the highest density of survivors, or locations which are the easiest to reach for emergency services. The system should also plot the most efficient and safest route that can then be used by emergency services to quickly reach the survivors. This solution should ideally be deployed as fast as possible to provide a detailed overview of the situation the first moments after disaster struck. It should also be applicable in a large range of natural disasters, these should take into account the varying conditions of different disasters, for example the system should be able to deal with bad weather, rough terrain, high temperatures etc.

Users

What do the users require

Our users require a way to quickly get information about a large disaster. This would mean that we must automate this information gathering on different scales. The users have the problem that they cannot get to a disaster quick enough, and when they are at the disaster, they cannot get information quick enough because of the scale. For example, when there is a very large earthquake. The emergency services have no good way to get to the disaster, they do no immediately know the scale of the disaster and they do not know which parts really require there attention. This all costs much time, which can be greatly reduced. To get all this information really quickly, sometimes they use drones. These are manual controlled. This means that they can only gather information at as many places as they have people available. If we can make the robots independent and automated, while communicating with each other and giving important information to the users, this process would become much faster.

Who are the stakeholders?

There are different stakeholders with different roles in this project involved. They would all take advantage of a solution we provide to their problem. The three stakeholders are Users, Society and Enterprise. We will describe per category why this particular stakeholder is involved into our problem and how our project will contribute to a solution for their problems.

Users

The biggest group of stakeholders are the users, which consists of civilians, government organizations, and private organizations or non-government organizations. These would all take advantage of the solution we provide, in particular those which are our intended end-user, i.e. the groups which will be involved during a natural disaster. We shall describe how these groups use our solution

• Government Organizations

Organizations formed by the government to combat natural disasters will take the most advantage from our solution. When a natural disaster will take place on large scale, emergency services or other organizations want to gather information as quick as possible. With our solution, this will become automated and much quicker.

• Civilians

Civilians struck by natural disasters benefit from our solution. The quicker help comes, the smaller problems arising for civilians will be. This counts for medical care, but also search and rescue and preventing loss of private property.

• Private organizations/non-government organizations

Organizations could also use our solution to work for different purposes. For example as security of property. Next to that, our solution to the described problem could be used as a good solution for similar problems as government organizations are describing.

Society

The society in a whole would benefit greatly from our solution. Our solution is relative cheap, and would be a great addition or replacement to existing solutions. Our solution would contribute to prevent loss of life, loss of property and would help organizations greatly. Next to that, since it is not a expensive solution, it would be much more cost effective than existing solutions such as the manual controlled drone.

Enterprise

The enterprise would also benefit from our solution. Firstly, the usage of drones would be far greater than before. This would mean that enterprises could cash in into our solutions.

Approach and deliverables

In order to complete the project, first, a thorough literature study will be conducted that will be used to determine the state-of-the-art regarding drone technology as well as their current use in natural disasters and regarding the current other methods of observation during natural disasters and the problems and restrictions associated with these methods. Based on this information, an optimal observation strategy for a certain set of defined scenario's will be designed, based on the abilities of drones and maximizing the amount of new information a single drone can deliver. To provide evidence of the working of this strategy, a simulation will be made that shows how a network of drones would operate in such a disaster area. If enough time is left after these deliverables are finished, a small-scale test setup using a few drones can be used to show the functionality of the network in reality. The final report will exist of a literature study, an explanation of the optimized observation strategy, and the simulation. It will also include an expansive explanation of the choices made during each part of the project.

Milestones

The three main milestones that need to be reached are the finishing of the literature study, the observation strategy and the simulation.

literature study/state of the art

Flood-survivors detection using IR imagery on an autonomous drone

https://pdfs.semanticscholar.org/048c/3193942a9fa6aa416b669b9a3dc72167ab2b.pdf
Diederik list: (5/5)
https://ieeexplore.ieee.org/abstract/document/6929384
https://ieeexplore.ieee.org/abstract/document/7303086
https://www.computer.org/csdl/mags/pc/2017/01/mpc2017010024-abs.html
https://www.academia.edu/30915171/Unmanned_Aerial_Vehicles_in_Response_to_Natural_Disasters
https://www.researchgate.net/profile/Agoston_Restas/publication/283537233_Drone_Applications_for_Supporting_Disaster_Management/links/5688148e08ae051f9af5b166/Drone-Applications-for-Supporting-Disaster-Management.pdf
End diederik list
Laila :
https://ieeexplore.ieee.org/abstract/document/8214963, Massive MIMO for communications with drone swarms
https://ieeexplore-ieee-org.dianus.libr.tue.nl/document/8316776, Multi-Tier drone architecture for 5G/B5G cellular networks: Challenges, trends, and prospects
https://ieeexplore-ieee-org.dianus.libr.tue.nl/document/7807176, Help from the sky : Leveraging UAVs for disaster management
https://ieeexplore-ieee-org.dianus.libr.tue.nl/document/7823357, A decade of research in opportunistic networks : Challenges, relevance, and future directions
https://ieeexplore-ieee-org.dianus.libr.tue.nl/document/6940426, Spectrum policy challenges of UAV/drones

Problems of a trajectory planning in autonomous navigation systems based on technical vision and AI

https://ieeexplore.ieee.org/document/8317265
This article mainly talks about the problems that exist when trying to navigate an environment using just vision and AI, it also attempts to give a solution to this problem. It was chosen on the basis that for the planned autonomous system drones are used to navigate the environment. It is therefore important that these drones can navigate the environment themselves. This article is also useful for the idea that the system can plan a route to the survivors based on the images provides by the drones.

A Disaster Response System based on Human-Agent Collectives

https://jair.org/index.php/jair/article/view/11037
This article gives a proposal for a disaster response system that, when combined with rescue personell, can improve the effectiveness of the rescue operation. This paper can be used to compare our ideas against an already established. This information can then be used to add functions that are not present yet, or to improve on the functions that are present using the new information.

Motion planning with temporal-logic specifications: Progress and challenges

https://content.iospress.com/download/ai-communications/aic682?id=ai-communications%2Faic682
Here the concept of time sensitive decision-making is central. This is essential in a search-and-rescue mission as quick decision making and optimal strategies are important to save as many lives as possible. This article can then be figure out a process to incorporate this time sensitive decision making system into our system.

Unblinking eyes: the ethics of automating surveillance

https://link.springer.com/article/10.1007/s10676-012-9291-0
This paper talks about the problems that occur when automated surveillance becomes the norm. This might not seem that relevant for our situation, but if one would consider that a search-and-rescue robot that can search for people is the fact that it could also be implemented into general surveillance systems. This could then mean that everyone could be surveilled at any time, combined with the detection of people this could cause an enormous breach in privacy if it were implemented. Careful considerations should be made to prevent this from happening as this could potentially cause a "Big-Brother" type scenario if it were to come to fruition.

Ethical Aspects of Facial Recognition Systems in Public Places

https://www.emeraldinsight.com/doi/pdfplus/10.1108/14779960480000246
Here the idea of facial recognition incorporated in surveillance systems is discussed. This relates back to a previous point about automated surveillance and the prevention of a "Big Brother" scenario. The system used to find people after a natural disaster should never have access to facial recognition, it should only need general recognition of a human to be succesful. If facial recognition were to be added, this would then make it easier to implement this idea into surveillance systems, which is precisely what should be avoided.