Page 22 - Read Online
P. 22
Page 328 Lei et al. Intell Robot 2022;2(4):31332 I http://dx.doi.org/10.20517/ir.2022.18
Figure 9. Illustration of robot global path planning in the tested workspace: (A and C) examples of informative trajectory between two
feeding/drinking lines in various DCPP directions and historical dead bird distribution data; and (B) the overall DCPP trajectory.
Figure 10. Informative planning protocol avoids a random obstacle in the grid: (A) the final planned informative directed coverage trajectory
in grid; and (B) heat map with historical data of dead broilers and DCPP direction information.
imperialistic competition in humans.
Then, the reactive local navigator was utilized to generate real-time commands (e.g., acceleration, deceleration,
and turning) for the robot arriving at a dead bird. Figure 12 shows three instances of grids for local naviga-
tion. Based on our simulation studies, using the dynamic window approach (DWA) local navigator, the robot
successfully reached the existing dead birds and simultaneously avoided moving/still broilers. Robot running
time and local trajectories could vary with grid locations, dead bird number in a grid, and distribution of live
birds. As shown in Figure 12A, the two connected grids were between feeding and drinking lines surrounded
