UAV Trajectory Planning in Cave Environments Based on Improved Artificial Potential Field and Raccoon Optimization Algorithm

UAV Trajectory Planning in Cave Environments Based on Improved Artificial Potential Field and Raccoon Optimization Algorithm

Abstract: This study addresses the complexity of unmanned aerial vehicle (UAV) path planning in challenging cave environments. Caves present unique challenges, including flying obstacles, complex U-shaped passages, limited GPS availability, and intricate spatial topologies. To overcome these obstacles, this study proposes a hybrid path planning algorithm that combines a modified artificial potential field (APF) method with the raccoon optimization algorithm (ROA). The improved APF method tackles issues like unreachable targets and local minima, ensuring efficient UAV navigation to the desired location. The ROA further enhances the algorithm by automatically optimizing the APF parameters, adapting them to the specific requirements of diverse cave environments. This results in improved path planning efficiency and overall performance. Preliminary experiments and validations conducted in actual cave settings demonstrate the potential of this hybrid approach for applications such as cave exploration and environment modeling. Further research and testing across a broader range of realistic scenarios are planned to validate the algorithm's robustness and feasibility. This study provides a significant step towards the effective deployment of UAVs in challenging cave environments for various potential applications.

Keywords: cave environment, artificial potential field, raccoon algorithm, trajectory planning, UAV, path planning, obstacle avoidance

1. Introduction

Underground environments, such as mines and unexplored caves, pose significant challenges for exploration. Mines often feature vast open spaces, active mining areas, and potentially hazardous passages [1]. Caves, on the other hand, present intricate arrangements of overhead obstacles and complex U-shaped passages. Both environments share common challenges, including unreliable GPS signals and complex spatial layouts.

Traditional exploration methods involve human personnel physically navigating these environments while carrying cumbersome detection equipment. However, these approaches are inherently risky and limited by human factors like physical constraints, safety concerns, and restricted operating durations [2]. Rotary-wing UAVs, with their compact size, exceptional hovering capabilities, and high maneuverability [3-4], have emerged as promising tools for underground exploration.

Efficient UAV path planning becomes paramount in such challenging settings. Researchers have dedicated extensive efforts to address this, exploring various algorithms such as the A* algorithm [5], genetic algorithms (GA) [6], particle swarm optimization (PSO) [7], ant colony optimization (ACO) [8], Rapidly-exploring Random Trees (RRT) [9], reinforcement learning [10], and the artificial potential field (APF) method [11]. When it comes to cave environments, the focus shifts towards the UAV's ability to plan locally and effectively avoid obstacles.

Among these algorithms, the APF method stands out due to its advantages, including low computational complexity, real-time capabilities, ease of control, robustness, and excellent obstacle avoidance performance [12]. However, traditional APF methods suffer from drawbacks such as getting trapped in local minima, encountering unreachable target situations, and generating suboptimal paths in multi-target scenarios [13].

To address these limitations, researchers have explored various improvements to APF. Some have introduced modifications to the potential field function itself, enhancing its ability to navigate around static obstacles and find optimal paths [14]. Others have proposed combining APF with complementary algorithms like ACO [17] or RRT [18] to mitigate the local minima issue. Machine learning techniques, such as integrating black hole potential fields with reinforcement learning [19] or using APF as guidance for DQN action selection [20], have also shown promise in overcoming APF's limitations.

This research delves into the problem of UAV trajectory planning within cave environments, focusing on two key challenges:

1. Closed 3D Environment Modeling: Establishing a suitable model that captures the constraints of the cave's 3D space, including obstacles on both the ceiling and floor, as well as complex passage shapes, to enable optimal path planning to a single target point (shortest path under flight constraints).2. Local Obstacle Avoidance: Ensuring the UAV can effectively adjust its trajectory to avoid unforeseen static obstacles within the cave environment without violating the overall constraints of the modeled 3D space.

Contributions:

This study makes significant contributions to the field of UAV trajectory planning in cave environments:

1. Improved APF with Rotating Attractive Potential: We introduce a novel approach that leverages a rotating attractive potential to effectively address the problem of local minima often encountered by traditional APF methods.2. ROA-Based Parameter Optimization: We incorporate the raccoon optimization algorithm (ROA) to automatically fine-tune the APF parameters. This adaptation allows the algorithm to better suit the unique characteristics of various cave environments, leading to enhanced efficiency and performance in path planning.

Paper Organization:

The rest of this paper is structured as follows: Section 2 details the problem formulation, encompassing cave environment modeling and the optimization problem's mathematical description. Section 3 presents the proposed methodology, including the traditional APF, its limitations, and the proposed improvements using rotating attractive potential and ROA-based parameter optimization. Section 4 presents simulation results and analysis, showcasing the effectiveness of the proposed method. Finally, Section 5 concludes the paper and outlines potential future research directions.