Scalable multirobot planning for informed spatial sampling

Abstract

Persistent monitoring of a spatiotemporal fluid process requires data sampling and predictive modeling of the process being monitored. In this paper we present PASST algorithm: Predictive-model based Adaptive Sampling of a Spatio-Temporal process. PASST is an adaptive robotic sampling algorithm that leverages predictive models to efficiently and persistently monitor a fluid process in a given region of interest. Our algorithm makes use of the predictions from a learned prediction model to plan a path for an autonomous vehicle to adaptively and efficiently survey the region of interest. In turn, the sampled data is used to obtain better predictions by giving an updated initial state to the predictive model. For predictive model, we use Knowledged-based Neural Ordinary Differential Equations to train models of fluid processes. These models are orders of magnitude smaller in size and run much faster than fluid data obtained from direct numerical simulations of the partial differential equations that describe the fluid processes or other comparable computational fluids models. For path planning, we use reinforcement learning based planning algorithms that use the field predictions as reward functions. We evaluate our adaptive sampling path planning algorithm on both numerically simulated fluid data and realworld nowcast ocean flow data to show that we can sample the spatiotemporal field in the given region of interest for long time horizons. We also evaluate PASST algorithm's generalization ability to sample from fluid processes that are not in the training repertoire of the learned models.

Mori-Zwanzig Approach for Belief Abstraction with Application to Belief Space Planning

“…Burks et al (2019) develop a Gaussian mixture model, and develops a clustering-based technique for mixture condensation that scales well to large belief space. Manjanna et al (2022) leverage a multi-resolution aggregation in the feature space to solve the multirobot coverage problem as a MDP. Hoerger et al (2022) propose a multi-level POMDP solver that uses multiple levels of approximation to the system dynamics to reduce the complexity of the forward simulations used in Monte-Carlo Tree Search.…”

3D Gas Sensing with Multiple Nano Aerial Vehicles: Interference Analysis, Algorithms and Experimental Validation

“…Several approaches allocate optimally a set of predetermined locations to the robotic team, usually based on information content and location [20]. Other approaches exploit prior knowledge of the phenomenon, to make informative decisions [21], or they rely on a large amount of data, often coming from cameras [22,23]. Moreover, global knowledge can be injected into the planning, in the form of attractor landmarks [22] or by exploiting an underlying model of the phenomenon [24].…”

Space-Fluid Adaptive Sampling: A Field-Based, Self-organising Approach

“…Other research focuses on mobile sensors, e.g., robots [25,7,15,8], hence the goal is to move them to the most informative sensing locations. Some work aims to adapt the sampling process to cope with spatial (and, often, temporal) phenomena, preserving some spatial properties of the WSN or the phenomenon under observation [9,12,26].…”