Enhancing Autonomous Navigation in GNSS-Denied Environments: Integrating Collision Avoidance with Observability-Based Motion Planning

This paper introduces a collision avoidance (CA) method integrated into the Observability Based Motion Planning (OBMP) framework. OBMP facilitates autonomous robot navigation in GNSS-denied outdoor environments, leveraging the concept of observability degree. The CA-OBMP method proposed here is an extension of the OBMP algorithm, which constructs paths based on terrestrial landmarks considering the obstacles. The CA approach is developed by redefining a dataset of in range landmarks while all of the landmark in vicinity of the obstacles are removed from the dataset. To evaluate the performance of the proposed method, simulations of the CA-OBMP algorithm are conducted for a 6-Degree of Freedom (DOF) quadrotor using MATLAB. The simulations encompass various arrangements of obstacles and diverse initial positions to assess the efficacy of CA method in different scenarios.