A Framework for Safe Local 3D Path Planning Based on Online Neural Euclidean Signed Distance Fields

This paper presents an open-source framework that integrates a distance-aware 3D local path planning al-gorithm based on Euclidean Signed Distance Fields (ESDFs) with an online generated Sinusoidal Representation Neural Network (SIREN) to estimate the required ESDF. The main con-tribution of the paper is a software framework that incorporates an online generated ESDF into local planners for efficient and safe 3D path planning by leveraging the ESDF properties. The framework includes a neural network that can be used by the local planner as an up-to-date representation of the environment. Experimental validation shows favorable results in exploiting the intrinsic characteristics of online generated ESDFs and acknowledges this framework as a feasible method to perform local path computation. The source code of the frame-work and more details about the software implementation is available at: https://github.com/robotics-upo/neural_esdf_local