Inference Grids for Environmental Mapping and Mission Planning of Autonomous Mobile Environmental Robots

Mobile sensing platforms provide a new modality for exploring the natural world. Robotic vehicles can be quickly deployed to new areas of interest and can have their sensor payload configured to the specific natural and environmental processes to be investigated. Building integrated sensing architectures that coordinate the operation of stationary networks and mobile platforms will allow researchers to take advantage of the strengths of both modalities, opening up new opportunities for scientific research and environmental monitoring. Among the various challenges to be faced, there are two key interrelated issues that are common to autonomous mobile platforms: the representation and modeling of natural processes using the sensor data being collected, and the use of this information to provide guidance, navigation and control for the mobile platforms. Both are addressed using a stochastic lattice-based framework for robot mapping, planning and control called the Inference Grid. In this paper, we will review our work on environmental robotic platforms, discuss how Inference Grids are used for natural process representation as well as for planning and control of autonomous robot vehicles, and show selected experimental results from field tests.