*4.2. On the Efficacy of the Proposed Planner*

The planning results of two typical simulation cases are depicted in Figure 8, which shows the efficiency of each layer. In each of the two cases, the finally derived path is collision free and kinematically feasible. Both properties can be reflected by the footprints and curvature profiles plotted in Figure 9.

## *4.3. On the Occlusion Awareness of the Proposed Planner*

This subsection investigates the occlusion awareness of the paths planned by the proposed method. Figure 10 shows the results with/without the positioning-related cost *J*positioning in layer three. When the cost term *J*positioning is discarded, the rate of good positioning distance along the entire path is 86.5% and 92.5% in the aforementioned two typical simulation cases, respectively. By contrast, with the cost term included, the rate grows to 97.0% and 96.5%. This comparative result clearly shows that our proposed planner can efficiently reduce the positioning inaccuracy caused by occlusions.

**Figure 8.** Path planning results of typical simulation cases in side view and bird-eye view: (**a**) Case 1; (**b**) Case 2 (zoom in to see more clearly).

**Figure 9.** Path planning performance, w.r.t. collision avoidance, and kinematic feasibility: (**a**) Case 1; (**b**) Case 2.

**Figure 10.** Comparative path planning results, w.r.t. occlusion awareness: (**a**) Case 1; (**b**) Case 2.
