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

All Robot Ideas

Below are all ideas that were thought up for the robots everywhere project, they are ordered by the order in which they were thought up. The final idea that was elaborated on was the disaster observation drones.

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

Problem Description

The employment of drones in disaster areas is an obvious application which can aid in the fast gathering of information concerning the situation and locating survivors. In recent years, extensive research has been conducted into using drones for this cause. The focus of these studies were, among other topics:

• Computer vision for recognizing survivors [Rivera, Villalobos, Monje, Mari ̃nas OppusRivera .2016].
• Communication among drones [Saha .2018].
• Optimal routing [Mersheeva .2015].

Past research has created a good foundation towards the development of an actual product, however a practical issue that still remains is the limited operation time of these drones. This problem is especially challenging because when applied to disaster areas, a solution must be independent of existing infrastructure which is likely to be damaged. The aim of this project is to develop a solution to extend the operation time of aerial drones, which are very suitable for disaster area monitoring due to their independence of ground infrastructure, but have an especially limited operation time compared to e.g. ground vehicles. Drones can be applied to many kinds of disaster, for the scope of this project we will focus specifically on earthquakes.

Users

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. 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 their attention. This all costs a lot of time, which can be greatly reduced. To get all this information really quick, drones are often used. 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. A major problem in the automation of these search processes is the limited operation time of the drones used. This is a problem for the operators, as it severely limits the range of the drone system, and requires the users to manually make sure drones are recharged. A system that allows these drones to be autonomously brought back into operation can save a lot of time which the people can save elsewhere.

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 with 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 of 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 be possible over a wider area and with less involvement of personnel necessary.

• 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, our solution to the described problem could be used as a good solution for similar problems as government organizations are describing.

Society

The society as a whole would benefit greatly from our solution. Our solution is relatively cheap and would be a great addition or replacement for 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 an expensive solution, it would be much more cost effective than existing solutions such as the manually 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

There are four functions that the system will need to be able to perform, each of which is listed below:

• 1. Drones need to be able to reach the system. This may include something that allows the system to move around in a potentially chaotic and hard to traverse disaster area.

• 2. The system needs to be able to either recharge or swap the batteries of a drone so that the drone can continue its search afterwards.

• 3. The system needs to keep working long enough to be able to increase the effectiveness of the swarm before having to be recharged/refuelled itself.

• 4. The system needs to be able to communicate their location to drones and receive information on battery levels and location of drones.

In order to keep a narrow and well-defined problem, a general solution for the design will be chosen based on the benefits and drawbacks of some chosen concepts. This general design will then be used to choose an optimal design for each function in order. That means, for instance, that the design for function 4 will be chosen to be chosen to work best combined with the designs for the other three already chosen functions. Once each function has been chosen, they will be worked out in greater detail, this time making sure all designs are fully compatible. In short, only one design option will be researched at a certain time, and the design of any option will be constrained by choices already made for the other options.

RPCs

The RPCs for the system will be defined as follows:

Requirements

• Can swap batteries of the drones.
• Needs to be mobile.
• 25 km system range extension.
• Fully autonomous positioning and task execution.
• Return to base after task completion or when in need to service.
• Max. one minute battery swap.
• Can service 10 drones in its operation cycle (about 2 kg payload).
• Can service at least 1 drone at a time.

Preferences

• The system should still function in "bad" weather conditions (rain, wind up to a certain speed).
• The vehicle should be safe for human interaction.
• Low manufacturing and operation costs.
• The system should support a manual override.
• The system should be easy employable.

Constraints

• The system needs to function independently of available infrastructure.
• Can be operational on rough terrain.

Phase 1: Traversal of environment

Here the concepts will be discussed on how the drone moves around the environment, a choice will also be made afterwards.

Concepts

In this section, several different concepts for the transport of the system will be discussed. The benefits and drawbacks of each concept in the given environment will be discussed so a well-informed choice can be made.

Ground vehicle

A ground drone that can move around the disaster area that can serve as a charging platform for drones. This comes with some up- and downsides. First of all the drone could move a bigger payload with the same amount of fuel when compared to a flying drone. It will also be less sensitive to weather conditions than an airborne drone. It does, however, have limited speed and movement depending on the state of the terrain.

This last point could prove troublesome after an earthquake. After such a disaster the rubble will prevent proper movement for smaller drones, larger drones could, however, have less trouble with this but can also need more space to move through. It is clear that a ground vehicle cannot follow the searching drones as well as it could if it were flying, this further complicates its function as a mobile charging platform.

It also operates in a more dangerous environment compared to aerial drones. It must be mindful of unstable areas and if they hit an obstacle. It could, for example, cause some walls or a building to collapse. This then further endangers the lives of both the drone and survivors of the disaster.

Ground drones do need to be given more attention to how they move around a disaster area. In general, some path planning has to be done from beginning to end. Considering that the area will be difficult to traverse and certain locations are destroyed, the drone will need some type of advanced vision analysis [2][4][5]. This would either require a constant connection to a computation server, or a considerable amount of processing power to achieve this.

Balloon UAV

A balloon is added to a quadcopter, this balloon's lift partially cancels the weight of the quadcopter. This allows it to stay in the air for a considerably longer amount of time. The drone also does not have any problem with broken infrastructure due to the ability to fly.

This idea also has some downsides. First of all it would be difficult to keep stable in windy areas [6]. It would, therefore, have difficulty with charging the drones using any method due to the movement of the two drones. The size of the balloon would also have to be large to cancel the weight enough for it to fly considerably longer [7]. This is therefore not an ideal solution for the main drone design.

Continuous flying large drone

A drone that can fly continuously without having to land. This drone uses up a lot of energy especially considering that drones usually only have a battery life of about 30 minutes during operation. The obvious solution to this would be by using alternative fuel sources or additional batteries. Alternative fuel sources could include diesel or petrol, which have a considerably higher energy density and therefore extend its range, even considering the lower efficiency of combustion engines. Such a large drone is however very costly to manufacture and uses a lot of resources during operation, which may be limited in a disaster scenario. More power cells could also be an improvement, but this would also increase the total weight of the drone and therefore decrease the flight time. This drone also suffers from other problems that aerial drones usually face. Its relatively large size and mass do make it more stable in daunting weather conditions [6].

Hopping drone

Drones fly ahead and land at certain spots creating an infrastructure for the observation drones to have their battery swapped and extend operation time. A certain amount of drones can form a grid of charging stations. The observation drones can fly to the nearest station when in need of service. Because the drone lands on its designated location, it can stay relatively simple and cheap compared to a drone that has to stay airborne for a longer time. The system itself takes the form of a larger drone, which should be large enough to carry what is needed to recharge or swap the batteries of the other drones.

Large drone with chain of recharging drones

Fuel powered large drone. Can fly for around 1 hour without being refuelled or recharged itself. With a chain of refuelling drones, the large drone will be refuelled/recharged. The hopping drone is a concept that combines the best of both worlds in regards to aerial and terrestrial drones. It solves the traversal of the environmental problem in the ground drone, but also solves the stability problem of aerial drones, as it can land and charge the drones on the ground.

Choosing

Out of the ideas and concepts discussed, the hopping drone seems to be the best solution. it's hard to keep any of the other designs stable while keeping them in the air and resting on the ground instead of flying in the air allows the system to last much longer. Drones will have an easier time reaching a still target. The hopping drone is a concept that combines the best of both worlds in regards to aerial and terrestrial drones. It solves the traversal of the environment problem in the ground drone, but also solves the stability problem of aerial drones, as it can land and charge the drones on the ground.

Phase 2: Increasing flight time of drones

For the drone to succeed in its function, it would need a method to extend the flight time of the drone. This can be done in a multitude of ways, below several concepts are discussed and a decision is made.

Concepts

Below several concepts for increasing the lifetime of the search drones are discussed. The pros and cons will both be discussed so an informed decision can be made.

Changing Batteries

Changing the batteries of a drone is one of the quickest ways to make sure that a drone can fly for longer periods of time. This means that the downtime of the drone only consists of the time it takes to change the battery. This method also has 0 power loss as the power is not transferred to the drone by cable, but by physical movement.

It does, however, have a few drawbacks. Considering the variable placement of the two drones relative to each other, the drone might have difficulty with changing the battery. Multiple solutions have already been made for this problem, the solutions range between a 50% success rate to an almost 100% success rate [9][10][12]. Needless to say, this second option would be ideal to implement.

The other problem is that the replacement system would need to carry all the batteries, be it full or empty, to the next location. This would require a bigger drone to be able to store them somewhere, stronger motors to be able to carry everything, and would need to have enough power to move everything around. These are problems that cannot be solved easily, but the effects could be diminished by selecting lighter batteries for the drones without compromising the stored power too much.

Planning

Week 4

• Reformulate Idea and specify the exact approach
• Rewrite wiki
• Finish Phase 1
  • Concepts and Choice
• Write State of the art Part 1
  • Supports Phase 1 of the design

Carnaval holiday

• Finish Phase 2
  • Concepts and Choice
• Write State of the Art Part 2
  • Supports Phase 2 of the design
• Finish Phase 3
  • Concepts and Choice
• Write State of the Art Part 3
  • Supports Phase 3 of the design

Week 5

• Finish Phase 4
  • Concepts and Choice
• Write State of the Art Part 4
  • Supports Phase 4 of the design
• Start work on the design

Week 6

• Continue working on the design
• Add finished aspects to the wiki

Week 7

• Finish Design
• Finish sections design wiki
• Write Conclusion, Reflection and Discussion
• Prepare presentation
• Finalize the wiki

Week 8

• Presentation
  
• Hand in report
  

State of the art

Pathfinding through rough terrain

Path planning for robots can be done in multiple ways, but finding the right choice for rescue operations can prove cumbersome. Research has been done to improve the path planning in regards to time planning and determining if the path taken can be completed. One such research is using genetic algorithms to determining a path as shown in source [3]. This does, however, have the drawback as it assumes to know what terrain is difficult to traverse and what isn't. However, this method is able to deal with unexpected situations and plans a new path that is close to optimal to reach its goal.

Another method is to evaluate the chance the robot will tilt when moving through the disaster area [4]. This is done by determining the height of the area using sensors and constructing a height gradient. The robot can then decide on a path through this gradient after nodes have been set, it takes into account the length of the path and the chance of tilting over. This method is ideal for small case areas but would need some considerable computation power to reliably do this continuously.

A variation on this idea is to change the configuration of actuators depending on the terrain [2]. This combines the path planning of the previous idea with additional functionality to further decrease the chance of tipping over. This can, therefore, be added as an extra to existing robots, given that it knows what the path will be like when moving towards it.

Using deep reinforcement learning is also an option for terrain navigation [5]. This method uses an elevation map as well and can learn what route it should take to reach the goal. This can then be applied to a robot and it should select a successful route, it could even learn if it makes a mistake. This aspect of self-improvement is unique to deep learning.

The final option that was researched is the option of using a guidance system that will guide a robot through dangerous areas [1]. This guidance system can use a multitude of lightweight sensors that can be placed all around the area. These will then connect with the main network and determine what areas are hazardous. This system is not ideal to move around obstacles. It is however useful in finding survivors as this is another functionality of this design.

Sources

• [1] Aliali, S., & Benchaiba, M. (2018). Safe route guidance of rescue robots and agents based on hazard areas dissemination. Proceedings of the 2017 International Conference on Mathematics and Information Technology, ICMIT 2017, 2018–Janua, 29–37. https://doi.org/10.1109/MATHIT.2017.8259692
• [2] Brunner, M., Bruggemann, B., & Schulz, D. (2012). Motion planning for actively reconfigurable mobile robots in search and rescue scenarios. 2012 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2012, 00(c), 2–7. https://doi.org/10.1109/SSRR.2012.6523896
• [3] Fries, T. P. (2019). Autonomous Robot Navigation in Diverse Terrain Using a Fuzzy Evolutionary Technique. IECON 2018 - 44th Annual Conference of the IEEE Industrial Electronics Society, 1, 5618–5623. https://doi.org/10.1109/iecon.2018.8591210
• [4] Ohki, T., Nagatani, K., & Yoshida, K. (2010). Safety path planning for mobile robot on rough terrain considering instability of attitude maneuver. 2010 IEEE/SICE International Symposium on System Integration: SI International 2010 - The 3rd Symposium on System Integration, SII 2010, Proceedings, 55–60. https://doi.org/10.1109/SII.2010.5708301
• [5] Zhang, K., Niroui, F., Ficocelli, M., & Nejat, G. (2018). Robot Navigation of Environments with Unknown Rough Terrain Using deep Reinforcement Learning. 2018 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2018. https://doi.org/10.1109/SSRR.2018.8468643
• [6] Santoso, F., Garratt, M. A., Anavatti, S. G., & Petersen, I. (2018). Robust Hybrid Nonlinear Control Systems for the Dynamics of a Quadcopter Drone. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 1–13. https://doi.org/10.1109/TSMC.2018.2836922
• [7] Turns, S. (2006). Thermodynamics - Concepts and Applications. New York, NY: Cambridge University Press.
• [8] Choi, C. H., Jang, H. J., Lim, S. G., Lim, H. C., Cho, S. H., & Gaponov, I. (2017). Automatic wireless drone charging station creating essential environment for continuous drone operation. 2016 International Conference on Control, Automation and Information Sciences, ICCAIS 2016, 132–136. https://doi.org/10.1109/ICCAIS.2016.7822448
• [9] Fujii, K., Higuchi, K., & Rekimoto, J. (2013). Endless flyer: A continuous flying drone with automatic battery replacement. Proceedings - IEEE 10th International Conference on Ubiquitous Intelligence and Computing, UIC 2013 and IEEE 10th International Conference on Autonomic and Trusted Computing, ATC 2013, 216–223. https://doi.org/10.1109/UIC-ATC.2013.103
• [10] Lee, D., Zhou, J., & Tze, W. (2015). Autonomous Battery Swapping System for Quadcopter. In 2015 International Conference on Unmanned Aircraft Systems (ICUAS) (pp. 118–124). Denver, Colorado: IEEE. https://doi.org/10.1109/ICUAS.2015.7152282
• [11] Liu, Z., Leo, D., Zhao, H., Wang, Z., & Liu, X.-Q. (2017). QUADO: and Autonomoys Recharge System for Quadcopter. In 2017 IEEE 8th International Conference on CIS & RAM (pp. 7–12). Ningbo, China. Retrieved from https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8274740
• [12] Swieringa, K. A., Hanson, C. B., Richardson, J. R., White, J. D., Hasan, Z., Qian, E., & Girard, A. (2010). Autonomous Battery Swapping System for Small-Scale Helicopters. In 2010 IEEE International Conference on Robotics and Automation (pp. 3335–3340). Anchorage, Alaska. Retrieved from https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5509165

Old Sections

• Old State of the art