Page 49 - Read Online
P. 49

Page 16 of 27                   Wang et al. Intell Robot 2023;3(4):538-64  I http://dx.doi.org/10.20517/ir.2023.30




                                                       
                                                       parent(      ),     =      
                                                       
                                                       
                                                       
                                             child(      ) = parent(      ),     <   
                                                                                                      (36)
                                                       
                                                       parent(     +1 ),      ℎ      
                                                       
                                                       
                                                       = 1, 2, ,      
                    d. Shift: Select two different genes,       and      , by Equation (29), then shift the destination cell of the first
                      gene towards that of the second gene by one cell to adjust the location directionally.
                                               {
                                                     (                   )
                                                                         ) ,
                                                 shift parent(          ), parent(             =      
                                      child(      ) =
                                                 parent(      ),                  ℎ                    (37)
                                               = 1, 2, ,      
                      where the shift function has the definition as follows:



                                                                                (⌊      ⌋  ⌊      ⌋)
                                                                                        − 1        − 1
                  shift(      ,       ) =       + sgn ((      − 1 mod       ) − (      − 1 mod       )) +       · sgn  −  (38)
                                                                                                   
               (iii) Mutation
                    Mutation is mainly used to attempt to access a new destination cell to increase the adequacy of access
                    to the solution space in this problem. In advantageous circumstances, UAVs may approach new cells
                    with high search benefits in the new search path. In this algorithm, the mutated gene is selected by
                    Equation (29) and the mutation result is selected based on the probability distribution in Equation (32).

                (iv) Selection
                    Similar to the traditional selection in GA, a binary tournament selection method is used to select the
                    chromosomes in the population.

               5.1.4. Procedures of the priority-encoded IGAFA
               The procedures of the priority-encoded IGAFA, i.e., the procedures to obtain the optimal set of control input
               sequences, are as shown in Table 1.


               5.2. Complete search process
               Basedonthe previoussections, theproceduresof a completesearchprocessformoving targetsusingCSMTPE,
               are as shown in Table 2.



               6. RESULTS
               Simulation experiments for the multi-UAV regional cooperative search problem in this paper are performed in
               Matlab. The computer configuration is as follows: Windows 10, CPU AMD(R) Ryzen(R) 7 4800HS of 2.9GHz
               up to 4.2GHz, Memory 16GB of 3200MHz.

               The experiment assumes a military scenario where UAVs are used to search for targets of concealed enemy
               soldiers in a mission area. The UAVs search for enemies through airborne human detection equipment, while
               the enemies actively evade the UAVs’ search. Set the size of the mission area as       =       = 4     and divide the
               area into 40×40 grid cells with       = 100  . Assuming that the UAV moves at a speed of 15  /   and the target
               moves at a maximum speed of 6  /  , set the motion configuration for moving targets to be:             = 40  ,
                     = 1,       = 600  ,       = 0.8,      1 = 500  , and      2 = 300  , which implies       = 50   and       = 2. Set the
   44   45   46   47   48   49   50   51   52   53   54