A Unified Experimental Architecture for Informative Path Planning: from Simulation to Deployment with GuadalPlanner

TL;DR

Informative Path Planning is hampered by fragmented pipelines between simulation and real deployment. GuadalPlanner provides a unified, end-to-end architecture that decouples high-level decision-making from vehicle control and enforces identical execution pipelines across Algorithm-level Simulation, SITL, and Real deployment using ROS2, MAVLink, and MQTT. The approach is instantiated on a graph-based, 2D environment with fleets of vehicles, demonstrated through GP-based water quality monitoring and other scenarios, culminating in real ASV validation in a lake discretized to a $30\times 49$ grid ($351$ nodes) with $25\,\text{m}^2$ cells, max travel distance $925$ m, speed $0.5$ m/s, and mission durations of $30$–$35$ minutes. Key contributions include an open, extensible toolchain for reproducible IPP experimentation, cross-level compatibility, and demonstrated transfer from simulation to real-world deployment, reducing architectural overhead in IPP research and applications.

Abstract

The evaluation of informative path planning algorithms for autonomous vehicles is often hindered by fragmented execution pipelines and limited transferability between simulation and real-world deployment. This paper introduces a unified architecture that decouples high-level decision-making from vehicle-specific control, enabling algorithms to be evaluated consistently across different abstraction levels without modification. The proposed architecture is realized through GuadalPlanner, which defines standardized interfaces between planning, sensing, and vehicle execution. It is an open and extensible research tool that supports discrete graph-based environments and interchangeable planning strategies, and is built upon widely adopted robotics technologies, including ROS2, MAVLink, and MQTT. Its design allows the same algorithmic logic to be deployed in fully simulated environments, software-in-the-loop configurations, and physical autonomous vehicles using an identical execution pipeline. The approach is validated through a set of experiments, including real-world deployment on an autonomous surface vehicle performing water quality monitoring with real-time sensor feedback.

Figures (10)

• Figure 1: Diagram of the GuadalPlanner implementation. It is structured around a hierarchical architecture of classes that progresses from general to specific.
• Figure 2: Internal structure of the proposed middleware architecture, illustrating the interaction between planning, communication, and vehicle interfaces implemented over ROS2. The green arrows corresponds to external interfaces to other peripherals.
• Figure 3: Diagram illustrating the connection flow between the decision-making module (GuadalPlanner), the ROS2 middleware, and the vehicle autopilot.
• Figure 4: Diagram of the ASV platform illustrating the key onboard components required to support full integration with GuadalPlanner
• Figure 5: Discretization of the scenario used during the experiments. Left: Satellite image of lake environment. Middel: Discretized grid representation. Right: Graph structure derived from the grid.