Page 150 - Innovations in Intelligent Machines

Evolution-based Dynamic Path Planning for Autonomous Vehicles 141
250
Expected value of loss function 150
200

100

50

0
0 5 10 15 20 25 30 35 40 45
Time sample
Fig. 22. Evolution of dynamic path planning with a moving target

4
step = 0
3.5 time = 0
3
Latitude (deg) 1.5 2 1 1
2.5

1
0.5
0
10 10.5 11 11.5 12 12.5 13 13.5 14 14.5 15
Longitude (deg)
Fig. 23. Off-line path planning result. The planner does not have the knowledge
that the obstacles will move in the future

The second example is a dynamic planning problem with moving obstacles.
During the off-line planning, the planner does not have the knowledge that
the obstacles will move in the future. The off-line planning result shown in
Figure 23 illustrates that the planner is able to find a near-optimal path to go
almost directly to the target and the goal location. Almost immediately after
the vehicle starts moving from its initial location, the obstacles begin moving
north and south to block the vehicle from getting in and out of the area where
the target is located. Figure 24 shows snapshots of the dynamically generated
path during the simulation. These snapshots show that the vehicle is able