A robot society is a group of robots (members of the society) which has the ability to communicate and perform tasks jointly. A society is defined by it's information and control structures which make possible common task planning and execution. These structures are of basic interest in this research project.
Towards distributed autonomous robotic systems
In near future the world where most of the autonomous mobile robots are operating is neither stabile nor structured. That is one of the main reasons, why a major trend in robotics currently is going towards multi-robot systems. Furthermore, in many cases, the decomposing of a complex task into parallel subtasks is the only solution or it at least speeds up the performance. With some limitations, i.e. there are several robots working with the same subtask, it can also increase the redundancy of the system. Redundancy itself can be of utmost importance in certain special occasions, like in those long-term and highly expensive planetary missions. Other natural domains for multi-robots are the tasks, where the environmental conditions will set some limitations for example for the size of the robot. In such cases the only possible solution is actually to “divide” the robot into several smaller robots. As an example one can consider environment, where the robot has to go through some very tight spots, like in some nuclear power plant where a cleaning task has to be performed inside a highly radioactive area. Furthermore, there can be tasks where a successful completion of the mission requires close cooperation among the robots. Such case is for example the carrying of a large object together. The performance of a task like this is far from being trivial. It requires some sort of interaction between robots, whether is a direct communication or some sort of indirect communication for example through sensing the forces in the object to be transported. This kind of task as well as many other tasks normally related to multi-robot systems, has clear analogy to biological systems. For example a group of ants solve the problem through sensing the forces and torque in the object. Based on this information they change the direction of forces accordingly or if needed some ants change the position of their hold. Numerous similar examples can be found from Nature around us. Tested by evolution during millions of years these structures are proven to be feasible in dynamic and hostile environments and can thus provide valuable information and inspiration for similar type of engineering tasks.
Efficiency through Cooperation
Cooperative robots are constantly interacting not only with the dynamic environment, but also naturally with each other and with the persons who are using them. This large variety of interactions will produce behaviors that are going to be superior compared to ones performed by current robots. The robots will be equipped to survive as a part of a complex system, where the cooperation is essential for their survival. Collective intelligence emerging from these interactions gives a reason to call these systems at their highest level as “robot societies." When designing multi-agent systems, whether they are software or hardware agent based, some fundamental problems need to be solved. Among these elementary problems are the following ones:
- How do we formulate, describe, decompose, and allocate problems among a group of intelligent agents?
- How do we enable agents to communicate and interact?
- How do we ensure that agents act coherently in their actions?
- How do we allow agents to recognize and reconcile conflicts?
To answer these questions, we are trying to develope a model for defining and designing distributed autonomous robotic systems. All functions of these systems are obviously realized through their members. The members’ behaviors are results from their own needs and from the constraints (dynamic by their nature) set by the system, environment or operator.
It seems obvious that the number of distributed autonomous robotic system applications will increase rapidly as the technology and knowledge improve. Within few years, various robot societies will move out from research laboratories into everyday life. Normal applications will include tasks like cleaning, monitoring, delivering, etc. Besides that a number of more revolutionary applications will probably also emerge. These will take robot societies to distant planets, deep sea nodule collection missions, mining operations, and when the time comes maybe even inside human veins for search of tumors, which they will attack at close range. In these scenarios the multiplicity of robots is more important than the intelligence they possess. Just as in natural systems the intelligence of the system will emerge from the multiple interactions among the members and with the environment.
We have been working from the beginning of the year 1992 with the robot society concept. The first part of the job was to define robot society’s structures. Ant societies were studied in order to find the key issues. Next step was to design a physical society named as A Model Society. It consists of two types of autonomous mobile robots (named as Workers and Energy-carriers), and the task for the society is classical, to gather stones from an unknown environment along with mapping the environment while operating. The society has been implemented both as a physical society(still under work) and as a simulated one. Along with this basic study, we were looking for a suitable application from a real world, where some of our tentative results could be verified. From the beginning of year 1994 we have been building a more realistic robot society application, SUBMAR. The idea is to use this robot society inside industrial processes.
- P Appelqvist,
Mechatronics Design of a Robot Society - A Case Study of Minimalist Underwater Robots for Distributed Perception and Task Execution,
Publication in series
- M Vainio,
Intelligence through interactions - underwater robot society for distributed operations in closed aquatic environment,
Espoo: Teknillinen korkeakoulu, 1999. (Laboratory of Automation Technology, Research reports, 21).
- M Vainio, P Appelqvist and A Halme,
"Mobile robot society for distributed operations in a closed aquatic environment,"
Robotica, Vol. 18, s. 235-250, 2000.
- A Halme, M Vainio, T Schönberg and P Appelqvist,
"Yhteisörobotiikan periaatteelle rakentuva liikkuva anturi/toimilaitejärjestelmä,"
Automaatioväylä, no. 1, s. 30-32, 1997.
- A Halme, P Jakubik, T Schönberg and M Vainio,
"The concept of robot society and its utilisation in future robotics,"
in Advanced Robotics & Intelligent Machines, J. O Gray and D. G Caldwell, London: The Institution of Electrical Engineers, 1996, s. 255-270.
- M Vainio, A Halme, P Appelqvist, P Kahkonen, P Jakubik, T Schonberg and Y Wang,
"An Application Concept of an Underwater Robot Society,"
in Distributed Autonomous Robotic Systems 2, H. Asama, Tokyo: Springer, 1996, s. 103-114.
- M Vainio, T Schönberg, A Halme and P Jakubik,
"Optimizing the performance of a robot society in structured environment through genetic algorithms,"
in Advances in Artificial Life, F Moran, A Moreno, J. J Merelo and P Chacon, Germany: Springer, 1995, 733-746.