Page 15 - Read Online
P. 15

Lei et al. Intell Robot 2022;2(4):313­32  I http://dx.doi.org/10.20517/ir.2022.18   Page 321

               The procedure of the proposed DCPP in a broiler barn is described in Algorithm 1. It is worth noting that the
               DCPP does not provide the robot with an actual coverage path, which could vary based on the IPP.


               Algorithm 1: Directed coverage path planning procedure in a broiler barn
               Initialize the left turn times    = 0;
               Initialize the next step state as null, S = [];
               Initialize the robot heading to the longer side of the barn;
               while (the robot does not reach target position) do
                   if (   < 2) then
                       if (no obstacle on the right) then
                           Turn right and move forward;
                       else if (no obstacle in the front) then
                           Move forward;
                       else if (no obstacle on the left) then
                           Turn left and move forward;
                              =    + 1;
                       end
                   end
                   if (   ≥ 3 and the robot can move forward) then
                       if (no obstacle in the front) then
                           Set S = [Move forward];
                       else if (no obstacle on the left) then
                           Set S = [Turn left and move forward];
                           Turn left;
                       else if (no obstacle on the right) then
                           Set S = [Turn right and move forward];
                           Turn right;
                       end
                       Perform S;
                   end
               end



               3.3. Informative planning protocol
               The informative planning protocol (IPP) further refines and optimizes the coverage path based on historical
               data on broiler mortality distribution in a broiler barn. In the general case, the IPP incorporates the multi-
               objective informative planning and uses DCPP and historical data to guide the robots to more frequently
               visit those places with a higher probability of broiler mortality appearance. This multi-objective optimization
               problem is stated in Equation (11).


                                                   {                            }
                                           = argmax   (  ),I 1 (  ),I 2 (  ), · · · ,I    (  ) ,
                                         ∗
                                                                                                      (11)
                                                ∈M
                                        s.t. C    ≤ B
               where    is a sequence of robot motion planning,    is the selected robot motion with maximum information
                                                          ∗
               gain, M is the list of possible robot motion sequences, C    is the cost budget, and B is a budget with regard
               to time, memory, energy, maximum iteration times, etc. I    (  ),    = 1, 2, ...,   , represents the information
               gathered by the robot motions and    is the number of the information source in the grid [34] .   (  ) represents
               the information from planned DCPP directions.
   10   11   12   13   14   15   16   17   18   19   20