*4.3. Temporary Response Effect of the New Risk Area*

During the process of drone logistics transportation in urban environments, the obstacles and risk areas in the environment can basically be examined and commanded to go around in the global path pre-planning stage before starting the mission due to the more comprehensive network coverage. However, due to the complexity of the urban environment, it is still challenging to avoid unknown obstacles in advance, such as flocks of birds, which require commanding the drone to change its route to avoid them.

In the path planning algorithm of this study, the drone scans the global environment at each step. Once there are new risk areas that affect the original flight path of the drone, the subsequent path is replanned to ensure that the drone adapts to the dynamic urban low-altitude environment. This section focuses on analysing the effect of the avoidance strategy proposed by the algorithm.

As shown in Figure 10 and Table 5, when a new risk area appears at the location of the point (0.2, 0.9), drone 1 moves one step according to the original path and finds that the subsequent original path passes through the new high risk cost area, so a local path replanning is performed to avoid the new risk area. The red dashed line in Figure 10 represents the locally replanned path of drone 1, and the solid red line indicates the original path. The solid blue line indicates the path of drone 2, the solid green line indicates the path of drone 3. As the new risk zone does not affect the original paths of drone 2 and drone 3, the paths of these two drones do not change. For analysing the impact on drone 1, which was replanned to avoid the new risk zone, we further compared and analysed the path parameters. The path length of drone 1 increased from 1.62 of the original path to 1.69, and the growth rate was 4.32%. The path risk cost was affected by the new risk zone, which increased from 5.20 to 5.22, with a growth rate of 0.35%. The service completion was always 100%, indicating that the path length increase was negligible. It is clear that the avoidance strategy proposed by the algorithm allows the drone to change the original path before entering the new risk zone. It could ensure that the risk cost from the new risk zone is mitigated and the increase in path length is minimal.

**Figure 10.** Path replanning due to new risk zones.

**Table 5.** Results of path replanning.


In order to further study new risk zones, this paper investigates the effect of the number of new risk zones on the path planning results. As shown in Figure 11, the length of path 1 increased by about 3.99% on average with the addition of each new risk area, the service completion always remained at 1, and the path risk cost increased by about 0.30% on average.

**Figure 11.** Path replanning due to new risk zones.

In summary, the temporary response of the algorithm to new risk zones can reduce new risk costs on the basis of service completion. The avoidance strategy is influenced by the time when the risk zone is discovered. The above discussed that a new risk zone is discovered before the drone enters that risk zone. Because the avoidance strategy requires the drone to scan and judge whether there is a new risk zone once in each step, it can guide the drone to update the next path point in time to avoid the risk zone. In the case that the drone has already entered the new risk zone when it is found, it is obvious that the drone will change the next path point according to the strategy, thus leaving the risk zone with the shortest distance. This case does not need to be discussed.
