Laura Sorbi

Target detection and recognition: A mission planner for Autonomous Underwater Vehicles

This paper proposes a mission planner controller, developed to manage a group of Autonomous Underwater Vehicles (AUVs) performing a target detection and recognition task. The whole mission is divided in two fundamental steps: i) an area covering task, aimed at finding objects of interest and implementing the Behavior-Based (BB) approach introduced in [1]; ii) a real-time path planner, leading the group towards the target, based on a modified Artificial potential Field (APF) controller. Moreover, a new formation control technique based on a modified Van der Pol oscillator is studied and integrated. To test the proposed mission planner, a global simulation environment has been implemented, including a real vehicle model. Results demonstrate the good overall performances of the proposed approach in several scenarios.

A Behavior-Based Mission Planner for Cooperative Autonomous Underwater Vehicles

Due to its applications in marine research, oceanographic, and undersea exploration, Autonomous Underwater Vehicles (AUVs) and the related control algorithms has been recently under intense investigation. In this work, we address target detection and tracking issues, proposing a control strategy which is able to benefit from the cooperation among robots within the fleet. In particular, we introduce a behavior- based planner for cooperative AUVs, proposing an algorithm able to search and recognize targets, in both static and dynamic scenarios. With no a priori information about the surrounding environment, robots cover an unknown area with the goal of finding objects of interest. When a target is found, the AUVs’ goal become to classify it (fixed target) or track it (mobile target), with no information about target trajectory and with the constraint on maintaining the formation. Results demonstrate the good overall performance of the proposed algorithm in both scenarios.

Behavior-based Control: An Approach to Cooperative AUVs Missions

Cooperation of multiple underwater vehicles can offer great improvements to a variety of applications, opening the possibility to perform more complex tasks (e.g., target recognition, environmental monitoring). This paper addresses control algorithm issues for multiple Autonomous Underwater Vehicles (AUVs) employed in a target detection mission. A novel Behavior-Based (BB) algorithm, which evaluates each single AUV motion as a reaction to the perceived surrounding, while taking into account the mission history, is proposed. Defining each robot behavioral action, the algorithm leads to an emergent behavior of the whole fleet, which can even be improved with data sharing. To test the proposed algorithm a real vehicle model together with its motion controller have been implemented. Results demonstrate that a realistic AUV can follow the reference path produced by the high-level controller with a negligible motion error.

Communication in a Behavior-based Approach to Target Detection and Tracking with Autonomous Underwater Vehicles

In this paper, we address the challenging topic of target detection and tracking in underwater environments, proposing a behavior-based algorithm suitable for both fixed and moving targets. Taking into account the stringent constraints of the acoustic underwater channel, we apply this method to fleets of collaborative AUVs. The resulting control strategy is able to benefit from the behavior-based approach and the cooperation among robots. Results show that the performance of the algorithm can be improved by tuning the range of vehicles communication, meeting the tradeoff between energy consumption and behavior efficiency.

On the Performances of a Target Detection Algorithm in Underwater Sensor Networks with Aqua-Sim

This paper is aimed at investigating the performance of a behavior-based controller for area coverage and target detection missions, in which multiple Autonomous Underwater Vehicles (AUVs) are employed. The main contribution of this work is the analysis of the performance improvement when a realistic underwater communication network is considered among robots, exploited in a realistic mission. In our model, we integrate the algorithm proposed in [1] and implemented in MATLAB with the Aqua-Sim simulator, a network simulator for underwater sensor network based on NS-2. Results show that the performance improvement is deeply related to the communication parameters, which have to be rightly chosen in order to find a good trade-off between energy consumption and algorithm efficiency.