Indian Journal of Geo-Marine Sciences Vol. 41(6), December 2012, pp. 581-588 A calibration framework for swarming ASVs’ system design Z Z Abidin 1 , M S M Hamzah 1 , M R Arshad 1 & U K Ngah 2 1 Underwater Robotics Research Group (URRG) 2 Imaging & Computational Intelligence Group (ICI) School of Electrical & Electronic Engineering, Universiti Sains Malaysia, 14300, Pulau Pinang, Malaysia. [E.mail: zzulkifli@iiu.edu.my, mohd_salzahrin83@yahoo.com, rizal@usm.edu.my and eeumi@eng.usm.my] Received 26 July 2012; revised 18 August 2012 This paper is concerned with the virtual simulation of Autonomous Surface Vessels (ASVs) named, Drosobots, using virtual simulation software i.e. Webots™, and the pre-deployment in a swimming pool environment based on an improved simplest navigation technique. Swimming pool provides as a controlled calibration framework for the proposed swarming algorithm. The performance of the system is determined by firstly, its capability to allow the various robots to communicate amongst themselves in order to reach the desired location and secondly, the use of optimization in its searching strategy. By using basic theories of GPS steering, low-cost microcontroller and straightforward wireless communication method, a framework which takes into consideration both mechanical constraints in its physical setup and the suitability of control methods is presented. Swarming robots work as a team, propelled by slim-line water pump with cylindrical shape of body hull. In order to increase the robot’s buoyancy, high density foam has been added to the previous design and results of the new rudder simulation effect is also been presented. Due to the delay of the NMEAs data and the limitation of an 8-bit micro-controller, complex control has been deferred until sometime in the future. [Keywords: Webots, Drosobots, Algorithm, Bio-Inspired, Methodological approach, Algorithms] Introduction Most of the current swarming robotics and bio- inspired projects involve work based on a miniature platform where the mobile robots' agents are assigned to search the brightest light intensity within the fieldwork, with a camera based localization technique 1,2,3,4,5 . The information is used to emulate the minimum/maximum searching area or virtually named Artificial Potential Fields (APF). For the real situation, that particular location might be the most hazardous zone, the highest peak on the ground or the deepest part of the marine area (i.e. the potential site that the system is looking for). In marine expedition, finding the deepest part of the aquatic area such as oceans, lakes, dams, ponds and rivers, are presents as one of the most interesting challenges. This kind of mission requires a lot of efforts and proper planning is necessary. One of the most outstanding expeditions carried out at this particular moment in time is the one which involves locating the actual GPS position of Mariana Trench. Most of the voyages used multi-beam sonar and their movement is a simple "lawn mower" formation to find the specific coordinate 6 . For biologists, the area serves as keen interest as it may contain various new organisms and microorganisms. Of late, work on Bio-Inspired Robots for swarming applications has gathered interest. One would wonder how the ants, bees, birds and flies exhibit effective team work when compared to a single handed animal. This research area empathizes more on the biological study, optimization and multi-agents robots development. Unlike the Multi-Autonomous Ariel Vehicles trajectory planning using the Ant Colony Optimization (ACO) approaches 7 , our project objective is mainly to perform contour mappings of calm water environment and ultimately determine the deepest location within the vicinity of the water area in a short period of time 8,9 . Though small in size, the intelligence that the multi robots exhibit while searching for food and mates, may be adopted in finding an optimal path. The most important factor is that they normally move in a small number. Therefore, in this study, we propose the “Fruit Fly” as agents and their searching patterns as an alternative algorithm 10 . According to Afek Y. et al., the fruit fly searching strategy is better than ants and bees, leading to an