First order Sliding Mode Control (SMC1) is used to control robotic manipulators deployed in industrial applications. However, the tracking performance of SMC1 is affected by modelling uncertainties, disturbance, and noisy measurement signals. In addition, SMC1 introduces an undesirable chattering effect that causes vibrations and reduces the manipulator positioning accuracy. This work investigates deployment of a Quasi-Continuous second order Sliding Mode Control (QCSMC) and its stability, over a nonlinear multilink Revolute-Revolute (RR) manipulator, to enhance the manipulator tracking performance, and reduce the chattering effect under the presence of modelling uncertainties, disturbance, and noisy state measurement. This investigation work introduces the manipulator’s dynamics modelling, forward and inverse kinematic solutions. The study then shows the desired joint space trajectory solution, whereby modelling uncertainties are considered. The work presents an online uncertainty quantification analysis to investigate the controller performance and stability with parameter change. SMC1 and QCSMC are then deployed to follow extracted trajectories under the effect of model-parameter-change, noisy measurement, and disturbance signals. Simulation results show that performance of QCSMC, overcomes the performance of SMC1, under the considered simulation conditions. Simulation results were further endorsed through experimentation methods by an NXT-based RR manipulator prototype. Simulation results demonstrate that applying QCSMC have reduced the associated joint-space trajectory root mean square errors (RMSE) when compared to SMC1 associated RMSE results. The enhancement reached an average of (84% to 97%) with infused model change and (64% to 79%) enhancement with infused disturbance. However, QCSMC shows a 2% RMSE value enhancement for the case of infused measurement noise when compared to SMC1. Experimental results demonstrate that the associated RMSE have also reduced by applying QCSMC when compared to SMC1. The experimental scenario nature includes a combined effect of modelling uncertainties, disturbances and measurement noise all applied together. Experimental scenario 1 shows that applying QCSMC have reduced the tracking RMSE for the first revolute joint angular position by 10%, and for the second joint angular position by 5% when compared to SMC 1 associated tracking RMSE values. Experimental scenario 2 shows a similar trend with 3% RMSE tracking enhancement for the second joint angular position when QCSMC is applied. Simulation and Experimental results agree, QCSMC introduces a better performance than SMC1, implying an enhanced performance, with less tracking errors, reduced chattering effect and higher robustness with reduced vibrations. In addition, the provided results signify robustness of both controllers against uncertainties and noise with better performance for QCSMC over SMC1.

一阶滑模控制 (SMC1) 用于控制工业应用中部署的机器人操纵器。然而，SMC1 的跟踪性能受建模不确定性、干扰和噪声测量信号的影响。此外，SMC1 会引入不良的颤振效应，导致振动并降低机械手定位精度。这项工作研究了在非线性多链路 Revolute-Revolute (RR) 机械臂上部署准连续二阶滑模控制 (QCSMC) 及其稳定性，以提高机械臂跟踪性能，并减少建模存在下的抖动效应不确定性、干扰和噪声状态测量。这项调查工作介绍了机械手的动力学建模、正向和反向运动学解决方案。然后，该研究显示了所需的联合空间轨迹解决方案，其中考虑了建模的不确定性。这项工作提出了一种在线不确定性量化分析，以研究参数变化时的控制器性能和稳定性。然后部署 SMC1 和 QCSMC 以在模型参数变化、噪声测量和干扰信号的影响下遵循提取的轨迹。仿真结果表明，在所考虑的仿真条件下，QCSMC 的性能优于 SMC1 的性能。基于 NXT 的 RR 机械手原型通过实验方法进一步证实了模拟结果。仿真结果表明，与 SMC1 相关的 RMSE 结果相比，应用 QCSMC 降低了相关的关节空间轨迹均方根误差 (RMSE)。注入模型改变时增强达到平均（84% 到 97%），注入干扰时增强达到平均（64% 到 79%）。然而，与 SMC1 相比，QCSMC 在注入测量噪声的情况下显示出 2% 的 RMSE 值增强。实验结果表明，与 SMC1 相比，应用 QCSMC 也降低了相关的 RMSE。实验场景性质包括所有一起应用的建模不确定性、干扰和测量噪声的综合影响。实验场景 1 表明，与 SMC 1 相关的跟踪 RMSE 值相比，应用 QCSMC 已将第一个旋转关节角位置的跟踪 RMSE 减少了 10%，将第二个关节角位置的跟踪 RMSE 减少了 5%。实验场景 2 显示了类似的趋势，当应用 QCSMC 时，第二个关节角位置的 RMSE 跟踪增强为 3%。仿真和实验结果一致，QCSMC 引入了比 SMC1 更好的性能，这意味着性能增强，跟踪错误更少，颤振效应减少，并且在减少振动的情况下具有更高的鲁棒性。此外，所提供的结果表明两个控制器对不确定性和噪声的鲁棒性，QCSMC 的性能优于 SMC1。