The research improves the ability of robots in their perceptions in terms of robotic arms by developing a torque-aware multi-station alignment approach to help tackle the problem of obtaining accurate multi-station alignment. The kinematics model of the robot arm is defined using a better Denavit–Hartenberg (D-H) approach for the establishment of mapping the trajectories from the Cartesian space to the joint space. This study proposes an enhanced rapid random tree (RRT) algorithm for conducting path planning of the rotary manipulator by taking advantage of the benefits of bidirectional and target-biased RRT methods. A simulation control system for the robot arm is designed using computational torque approaches. Based on the above, this study develops an advanced nine-point calibration approach. Using multiple pose data obtained from the same spatial point, the length of the robot sixth axis is calculated; thus, the sixth axis can be determined after the end-effector installation. Experiments showed that using the enhanced RRT algorithm resulted in the robot achieving minimum path costs, runtimes, and nodes of 1701.7 mm, 7.8 seconds, and 366, respectively. Nonlinear moment controller achieved improved performance in trajectory errors as well as control speeds. Using the three-stage alignment approach, angular and positional errors could be less than 0.06° and 0.20 mm, respectively.
