This work describes an underwater robotic system to be used for exploration, fruition and surveillance of underwater cultural heritage.
This system is based on a number of cheap autonomous underwater vehicles (AUV) organised on the swarm rules. Its development is currently in progress in our laboratory in the framework of the HARNESS project (Human telecontrolled Adaptive Robotic NEtwork of SensorS) . The project aims at the realization of an underwater multi body robotic system able to perform tasks in a fast and reliable way. The key points of the HARNESS project are: the development of a novel underwater acoustic channel with very high performances in routing and data throughput capacities and the design of a reliable swarm rule-based control system with an interface towards a supervisor operator acting as a priority definition arbiter. The project includes a demonstration with a limited number of physical vehicles and with a larger number of simulated ones.
In a swarm the members operate with a common objective, sharing the job workload; the lack of one member can be easily taken care of by redistributing the job among the others. This feature is especially useful if we consider, as application the discovery and the surveillance of submarine archeological sites. Moreover a swarm can be considered as a single body, offering the advantage of a simple way of interfacing with the human end-users and overcoming the problem of the control of a large number of individuals. The geometrical distribution of the members of this system is flexible and adaptable to the task and environment characteristics. In the underwater world the physical medium makes the acoustical channels as the most convenient one, since electromagnetic waves are rapidly damped. The acoustical technology has limited performances too, being affected by a fast decrease in the signal bandpass (limited carried frequency) as the distance increase. The swarm technology allow to avoid this drawback by a suitable and intelligent spatial distribution of transmission nodes (the swarm members themselves), allowing the exploitation of ultra-high frequencies and an enhanced data transmission through logical and physical routing. One of the aims of the project is the study and implementation of different behaviors in the swarm, to generate a collective shaping as a response to environment stimula and to modifiy of the communication parameters in order to maximize the performance of the system.
The swarm control must balance the different requests of the operator (e.g. modify the mission task), the swarm needs and the single members management (e.g. obstacle avoidance, loss of communication link).
The result is the selection of collective behaviours that must be compatible with all the before mentioned conditions.. The methods currently under evaluation to this aim include neural network techniques, fuzzy logic and genetic algorithms.
In the swarm there is no central brain, mainly because of the excess needs in band pass requested by such a brain. Instead each individual must possess an intelligent local control system capable to manage its choices according to the choices of the neighbours on the basis of the available data. Data coherence along the swarm, being affected by the the position of the member and by the data propagation speed is also a research topic.
The foreseen applications of this underwater swarm are the safeguard of the underwater cultural heritage in terms of: exploration of submerged sites, searching for possible new relics, surveillance known sites against damage, realization of a remote museum facility where tourists and archaeologists can explore underwater sites without “getting wet”.