Next step is to build a hull...
Thursday, March 15, 2012
Wednesday, March 14, 2012
Stijn Van Iersel, Martijn Lammers, Roel Smeets, Paul Staal
Protei has as a goal to create a vessel that can “live” autonomously and that can perform tasks as a swarm. In this description, we found similarities with the properties a species has. We are, at least for our project, proposing to approach our activities in the Protei project as if we are designing a newt species.Viewing Protei as a species provides us with design opportunities (or research questions).
Here, we will set the properties apart that the new species should possess and how these properties will, according to us, apply to Protei. We do not want to base these properties on any specific sort of species, but more on general aspects of nature to have a framework on which we can build up our Protei version and through which we can find opportunities to develop all of Protei’s functionalities. This framework is dynamic and the properties inside it can therefore be changed and new properties can be added.
Firstly, individuals get born. In the case of a Protei, this will happen through an interested contributor, building their own Protei. The design opportunity for this aspect lies in how the information of how to build a protei will be obtained. Designing a building kit is an interesting opportunity. Others are to create a better, clearer community where information about Protei is distributed. This way of working will create a ecosystem of similar but not identical Protei individuals and biodiversity may be a positive side effect.
The species called Protei can:
A new Protei will be born through the open source community. Interested “parents” can download plans or buy a kit with which they can build their own Protei. To ensure compatibility between different “breeds” of Protei, a certain standard will need to be set.
Design opportunities: development of a Protei building kit, development of open source community (a clear communication channel) including standards.
Respond to stimuli.
A protei responds to its environment. For example to cope with different wind directions when sailing, to objects blocking it’s path and to other Protei with which it is hunting for oil.
Design opportunities: What sensors should Protei have? Designing behavioral rules to which Protei should act, designing the learning algorithm inside protei (from what should it learn and from what should it not learn?)
As many species Protei will evolve over time. This will happen through development of the Protei concept by the open source community.
Design opportunity: create a evolution road map, development of the open source community.
Show collective behaviour
Being able to show collective behavior can have multiple advantages;
By interacting with each other Protei can share information on winds or oil spills (aka food).
With cooperation for protection, multiple Protei can, for example, stick together to protect themselves better from a storm. Together, Protei may be able to establish more efficient locomotion and quicker oil absorption.
Opportunity: Design a communication standard for communication between individual Protei.
Live in symbiosis
Being able to live in symbiosis with other organisms. In this case Protei will for example help birds who are covered in oil. While this is a rather extreme example, it might be more realistic to have Protei co-evolve. This can happen between different species of Protei (they can, for example, deliver software updates to each other. But maybe also between Protei and “natural” species.
Next step: Design briefs.
Based on these design opportunities each group member has written their own design brief.
Roel Smeets’ design brief
The design opportunity for me is to design and test the behavior of the drones. How will they communicate and how will they collect the oil? By creating visual models I want to test what methods are effective. These models should be based on existing technology to provide me with helpful data. If I base my model on futuristic and non-existing technologies it cannot be implemented within the Protei project.
Furthermore, I think that the environment as well as the behavior of the drones has a great impact on the design of the drones. It will be an opportunity for me to use my designed behavior and environmental the aspects to create a design for the drones of the Protei species.
List of proposed activities during the project:
- Research into species and swarms
- Analyses of environment
- Ideation of swarm behavior
- Concept of collective swarm behavior
- Testing of behavior
- Improving behavior (if necessary)
- Physical design of species
- Physical prototype of swarm behavior
At the midterm presentations I want to present my research and my first ideas of group behavior. In the presentation I will show a visual of the environment in which the drones will be placed. This visual will include all the important aspects of the ecology. It displays the flora of the region, the fauna and also the influence of humans in the environment. The research I did on swarm behavior will be summarized and presented. The main point I will hope to receive feedback on is my ideas for swarm behavior. So this will be the key part of the presentation.
For the final exhibition I want to make a visual model that accurately displays a Protei swarm cleaning up oil spills. In the most ideal situation I will also have a small scale physical swarm that can execute swarm behavior on a small scale. Using the data a gathered about the swarms and the environment I will make a design for the drones. This design will be based on the tasks that the drone has to accomplish, the environment and the swarm behavior. It will not be restricted by current technology or feasibility.
Paul Staal’s design brief
The general aim of the team is to look at the Protei living together as a species. A species that can reproduce, respond to stimuli, evolve, show collective behaviour and is able to live in symbiosis with other species.
My personal direction for this project is to get an understanding of how the boats communicate with each other. Growing up on a farm, I’ve always been around animals and I want to research their communication and integrate it in the communication of Protei. By looking at how animals communicate with each other to hunt or to prevent from being hunted upon, I want to create a ‘language’ for the Protei that is stripped down to its essentials. They should only communicate if they have to, so there should be as few rules as possible to ‘hunt down’ the oil-spills.
For the midterm presentation my goal is to finish research on animal behavior and make a computer model in Processing in which the Protei already have a basic intelligence to communicate where to find the oil and absorb it.
Stijn van Iersel & Martijn Lammers’s design brief
To suit the Next Nature theme the project is in, the main focus of the project will be looking at wild robots as a species of its own. This focus will be applied on the client Protei, there sailing boats will be seen as a species that reproduces, responds to stimuli, evolves, shows collective behaviour and is able to live in symbiosis with other species.
At this point the general understanding is that the protei species will reproduce itself by people that buy parts or a set/kit of parts (could be partially pre-assembled). The assembled product should be able to find its way to a swarm autonomously, allowing users to put it into the water anywhere. To reach a realistic design in these aspects, attention needs to be paid to compatibility and affordability.
The design opportunity in this is to design a basic boat assembly with kits available worldwide. The boat and kit itself need to be designed in a way that the final assembly will be very simple and low cost for all users. When designing the kit the functionality needs be designed to be pre programmed. This is necessary to assure compatibility with the whole protei swarm. After designing the physical and operational part there is also a marketing/ business area that is to be designed. To give a clear image of what actions are needed to accomplish the reproduction of the Protei species business models are also needed.
Midterm; General Boat design + Business model
Final; Specified Boat design + Kit design + Business model/analysis
Other design opportunities that add to the ones above are:
- Way to follow your own Protei robot / personal bonding.
- New born boats sharing their new software to update the existing swarm.
Midterm: General boat design of an RC Protei boat, with at least multiple parts of this boat prototyped.
Research boat design (involve expert)
Build and test parts of the boat that can later be put together to form a complete boat.
Final: RC boat design that is semi autonomous and of which a kit for home building is available.
Assemble all previously design boat parts into one complete boat. Which is semi-autonomous, this means the user will be in direct control of steering the boat (determining the course) and the other aspects of sailing, such as adjusting the sails, will be done by a sensor-actuator system on board of the boat.
The kit will be designed focused on it being reproducible through websites such as instructables, this means drawings + explanation of all the parts. In this phase I want to involve the instructables community very much as they will be the target group of this kit.
Thursday, March 1, 2012
Over the past 2 project-weeks, various brainstorm sessions, discussions and research have been performed in order to get a clear view on what we want to get out of this project. On the right are the possible directions we've come up with on Monday the 13th of February. We've also performed a brainstorm session that day with random items, to see whether new ideas sprang to mind using these items.
After these brainstorm sessions, it became clear that we want to focus on the Artificial Intelligence, especially autonomous sailing and swarm control. How can we ensure that a fleet of - for instance - 10 vessels can harmoniously work together without crashing each other.
We then continued searching for the optimal building solutions in order to create an easy, yet efficient sailing drone.
- One sail instead of two sails. If you build a gaff rig up front of the drone, you can get proper pulling power without extra control difficulties.
- A canting keel aids the boat to stay at a more horizontal level, ensuring a bigger surface to catch wind in the sails. We might look into this in the future.
- We might want to use a special material for the sails: There exist a very tin fabric of solar panels which can be used as a sail. This way, you have a big surface where energy can be caught to charge the batteries. This is also for in the future.
Plan for now
Our plan is to build a quick and easy sailing drone. We then want to actuate it by using servo's etc., while the final step will be autonomising it. The specs for this boat will be uploaded soon.