Abstract
This work presents the development and field testing of a novel adaptive visual information gathering (AVIG) framework for autonomous exploration of benthic environments using AUVs. The objective is to adapt dynamically the robot exploration using the visual information gathered online . This framework is based on a novel decision-time adaptive replanning (DAR) behavior that works together with a sparse Gaussian process (SGP) for environmental modeling and a Convolutional Neural Network (CNN) for semantic image segmentation. The framework is executed in mission time. The SGP uses semantic data obtained from stereo images to probabilistically model the spatial distribution of certain species of seagrass that colonize the sea bottom forming widespread meadows. The uncertainty of the probabilistic model provides a measure of sampling informativeness to the DAR behavior. The DAR behavior has been designed to execute successive informative paths, without stopping, considering the newest information obtained from the SGP. We solve the information path planning (IPP) problem by means of a novel depth-first (DF) version of the Monte Carlo tree search (MCTS). The DF-MCTS method has been designed to explore the state-space in a depth-first fashion, provide solution paths of a given length in an anytime manner, and reward smooth paths for field realization with non-holonomic robots. The complete framework has been integrated in a ROS environment as a high level layer of the AUV software architecture. A set of simulations and field testing show the effectiveness of the framework to gather data in P. oceanica environments.
Highlights
IntroductionThe main purpose of robotic exploration is to gather data in areas that humans can not reach, such as, in deep waters, in the space, or in missions that involve unsafe or hazardous environments or actions
This work presents the novel design of a adaptive visual Information Gathering (IG) framework (AVIG) that integrates a novel decision-time adaptive replanning behavior (DAR)
Such DAR behavior is coupled with a DF-Monte Carlo tree search (MCTS) strategy for information path planning (IPP), and joins two advantages of graph-based and sampling-based methods: (a) initializes a node network to set neighbor relations between sampling locations, and (b) samples paths in the node network through tree search following a decision-time strategy
Summary
The main purpose of robotic exploration is to gather data in areas that humans can not reach, such as, in deep waters, in the space, or in missions that involve unsafe or hazardous environments or actions. Such exploration can be performed by means of an operator controlling remotely a robot system. This relies on having a proper bidirectional communication between the operator and the robot (for instance to transmit robot images and operator commands), and in many scenarios such communication is not possible, or is very limited, and the robot has to be equipped with some degrees of autonomy. Data gathering missions with autonomous robots are normally limited to the use of preprogrammed paths, and their performance is supported only on a reliable localization and control
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
