*4.3. Experiments on the Manipulator*

To demonstrate the feasibility and effectiveness of the algorithm proposed in this paper, an experimental platform for path planning was built in a laboratory environment using the IRB1410 robot arm. A random position (1.289 m, 0.013 m, 0.873 m) was selected as the position of an obstacle in the manipulator workspace. In this scene, the path point of the manipulator end-effector is obtained by the PPRO-RRT algorithm, and then the manipulator is controlled by the program procedure from the starting point (1.241 m, 0.156 m, 0.703 m) around the obstacle to the target point (1.166 m, −0.087 m, 0.921 m), and the sequence of the path is shown in Figure 6.

**Figure 6.** *Cont*.

(**a**) (**b**)

**Figure 6.** Path sequence with the PPRO-RRT algorithm in obstacle avoidance. (**a**–**d**) represent different states of the manipulator and obstacle at each moment.

In this project, the individual joint angles of the robot arm change, as shown in Figure 7, from which we can see that the individual joint angles change smoothly.

**Figure 7.** Variation of each joint angle during the process of obstacle avoidance.

#### **5. Discussion**

In this paper, simple and complex environments were chosen to verify the feasibility of the proposed algorithm. The experimental results showed that the PPRO-RRT algorithm had better performance in both environments than the RRT algorithm, the RRT\* algorithm, and the RRT-Connect algorithm. However, in the field conditions, the position of the obstacles may change or there may be narrow gaps between the obstacles. The algorithm in this paper does not fit in the above scene, which creates some limitations on the path planning of the manipulator. Therefore, the following research will focus on enhancing the applicability of the algorithm to a wider range of scene.

In addition, in the real environment, the manipulator may collide with the obstacle, which is because the actual size of the manipulator needs to be considered during the

experiment in the real scene, which would make the collision detection algorithm more complicated and increase the running time of the algorithm. In order to simplify the collision detection algorithm, the links of the manipulator are abstracted as lines and the obstacle is inflated. The actual size of the manipulator is added to the obstacle so that the complex process of collision detection is transformed into a problem of the position relationship between a line and a sphere. Experimental results showed that with the proposed algorithm in this paper, the manipulator could safely reach the target point from the starting point.
