Abstract: In addressing the trajectory tracking control for existing mobile robots, the predominant has been focused on mitigating inherent self-pose errors, while the influence of trajectory curvature on tracking control has been overlooked. In order to further enhance the tracking accuracy of intelligent mobile wire-launching robots, a control approach is introduced herein, denoted as the trajectory evaluation model controller（TEMC）, which is predicated upon a trajectory evaluation model. Firstly, the establishment of kinematic models pertinent to mobile robots is undertaken. Secondly, to articulate the geometric interrelation between the reference trajectory and the robot, a trajectory evaluation model（TEM） is formulated. The model incorporates curvature and its rate of change to quantify trajectory complexity. A comprehensive trajectory evaluation function is devised, amalgamating trajectory complexity with factors of self-pose errors. Additionally, utilizing a back propagation neural network, a controller founded upon the trajectory evaluation model is introduced, accompanied by a presentation of the stability proof. Finally, the efficacy of the trajectory evaluation model is substantiated through rigorous simulation, and the essential parameter range within the model is determined empirically via experimental methods. It is demonstrated that the TEMC yields a tracking precision enhancement of over 48% compared to conventional adaptive backstepping controllers.

Key words: trajectory tracking, trajectory evaluation model, trajectory complexity, trajectory evaluation function

摘要： 针对现有移动机器人轨迹跟踪控制主要考虑自身位姿误差，未考虑路径曲率对跟踪控制的影响，为了进一步提高智能移动放线机器人的跟踪精度，提出了一种基于路径评价模型的控制方法（trajectory evaluation model controller，TEMC）。建立移动机器人运动学相关模型；为了描述参考路径与机器人之间的几何关系建立了路径评价模型（trajectory evaluation model，TEM），并引入曲率及曲率变化率对路径复杂度进行定义，同时综合路径复杂度与位姿误差因素设计了路径评价函数；借助BP神经网络，提出了一种基于路径评价模型的控制器，并给出了稳定性证明；通过仿真实验证明了路径评价模型的有效性，并通过实验法给出了路径评价模型中核心参数的取值范围，同时TEMC跟踪精度相较于传统自适应反演控制器提升了48%以上。

关键词: 轨迹跟踪, 路径评价模型, 路径复杂度, 路径评价函数