Niu Jing, Shen Chuanyan, Zhang Lipeng, Li Qijun, Ma Haohao
School of Mechatronics and Automotive Engineering, Tianshui Normal University, Tianshui, China.
School of Vehicle and Energy, Yanshan University, Qinhuangdao, China.
PeerJ Comput Sci. 2024 Dec 20;10:e2620. doi: 10.7717/peerj-cs.2620. eCollection 2024.
The widespread adoption of plant protection robots has brought intelligent technology and agricultural machinery into deep integration. However, with advances in robotic autonomy, the energy that robots can carry remains limited due to constraints on battery capacity and weight. This limitation restricts the robots' ability to perform tasks continuously over extended periods.
To address the challenges of achieving low energy consumption and efficiency in path planning for plant protection robots operating in mountainous environments, a multi-objective path optimization approach was developed. This approach combines the improved A* algorithm with the Improved Whale Optimization Algorithm (A*-IWOA), utilizing a 2.5D elevation grid map. First, an energy consumption model was created to account for the robot's energy use on slopes, based on its kinematic and dynamic models. Then, an improved A* search method was established by expanding to an 8-domain diagonal distance search and introducing a cost function influenced by cross-product decision values. Using the robot's motion trajectory as a constraint, the IWOA algorithm was applied to optimize the vector cross-product factor (p) by dynamically adjusting population positions and inertia weights, to minimize both energy consumption and path curvature. Finally, in simulation and orchard scenarios, the application effects of the proposed algorithm were evaluated and compared against notable variants of the A* algorithm using the robot ROS 2 operating system.
The experimental results show that the proposed algorithm substantially reduces the travel distance and enhances both path planning and computational efficiency. The improved approach meets the driving accuracy and energy consumption requirements for plant protection robots operating in mountainous environments.
This algorithm offers significant advantages in terms of computational accuracy, convergence speed, and efficiency. Moreover, the resulting paths satisfy the stringent energy consumption and path planning requirements of robots in unstructured mountain terrain. This improved algorithm could also be replicated and applied to other fields, such as picking robots, factory inspection robots, and complex industrial environments, where robust and efficient path planning is required.
植保机器人的广泛应用使智能技术与农业机械实现了深度融合。然而，随着机器人自主性的提高，由于电池容量和重量的限制，机器人能够携带的能量仍然有限。这种限制制约了机器人长时间连续执行任务的能力。
为应对山区环境中植保机器人路径规划实现低能耗和高效率的挑战，开发了一种多目标路径优化方法。该方法将改进的A算法与改进的鲸鱼优化算法（A-IWOA）相结合，利用2.5D高程网格地图。首先，基于机器人的运动学和动力学模型，创建了一个能耗模型，以计算机器人在斜坡上的能量消耗。然后，通过扩展到8邻域对角距离搜索并引入受叉积决策值影响的成本函数，建立了一种改进的A搜索方法。以机器人的运动轨迹为约束，应用IWOA算法通过动态调整种群位置和惯性权重来优化向量叉积因子（p），以最小化能耗和路径曲率。最后，在仿真和果园场景中，使用机器人ROS 2操作系统评估了该算法的应用效果，并与A算法的显著变体进行了比较。
实验结果表明，所提出的算法大幅缩短了行进距离，提高了路径规划和计算效率。改进后的方法满足了山区环境中植保机器人的行驶精度和能耗要求。
该算法在计算精度、收敛速度和效率方面具有显著优势。此外，生成的路径满足了非结构化山区地形中机器人严格的能耗和路径规划要求。这种改进算法还可以复制并应用于其他领域，如采摘机器人、工厂巡检机器人以及需要稳健高效路径规划的复杂工业环境。
