SmartExplorer1.0: A Novel Intelligent Framework for Path Planning in Robotics

Abstract

Autonomous systems require advanced path planning to ensure safe and efficient navigation through complex terrains. The SmartExplorer1.0 algorithm (SE1.0) utilizes Digital Terrain Models (DTMs) as input to identify navigable areas and plan optimal routes by incorporating terrain gradients, obstacle types, and environmental conditions. Compared to traditional methods such as A* or the Artificial Potential Field (APF), and modern metaheuristic methods such as Rapidly-Exploring Random Tree * (RRT*), SE1.0 offers superior adaptability. It complies with varying vehicle specifications while considering factors like soil granularity, slope consistency, heat sources and electromagnetic fields, elements often overlooked by conventional approaches. This integration allows for more precise and context-aware navigation, especially in dynamic or complex environments. The algorithm defines accessible regions and penalizes hazardous zones. Starting from specified initial and target points, it iteratively evaluates potential movements by optimizing an objective function to determine the optimal path. This approach not only visualizes accessible and non-accessible areas but also produces actionable data files for practical use, such as a list of trajectory points, which can be used to activate control sensors with pinpoint accuracy and enables the calculation of both completed trajectory and remaining path to be traveled. Its flexibility and comprehensive integration of environmental and vehicle-specific parameters make SE1.0 a versatile solution for autonomous navigation in challenging environments. The efficiency and versatility of SE1.0 are demonstrated by solving various path planning problems on DTMs also considering real world applications that involve the use of the DUNE rover.