Research Article

Autonomous Navigation Based on SEIF with Consistency Constraint for C-Ranger AUV

Algorithm 1

The SEIF-CC SLAM algorithm.
Require: waypoints
Global         //global variables
= Initialization (0);  //initialize global variables
Threshold = 8;  //The limited number of active features
WHILE waypoints not 0DO       //main loop
, , , ); //new vehicle pose
 Add random noise to nominal control values:
//Motion model
 Predict_FEJ(, , , )      //prediction
   Predict state
   Compute Equation (37)  //estimate at
   Compute   Ψ   Equation (12) with
   Predict information matrix Equation (10)
   Predict information vector Equation (11)
 Mean recovery:
 Get observations : set of rang-bearing observation
 Add random measurement noise
 //Data association
 //: index tags for each landmark;
;
 //Update step
 IF not empty THEN // is not a new feature
   Observe_Update(, , )
   FOR from 1 to size(, 2) DO
     Compute Equation (38)
     Compute Equation (15)
     Compute Equation (13)
     Compute Equation (14)
   END FOR
 END IF
//Augmentation step
 IF not empty THEN //new feature
  SEIF_Augment()
    Compute at
    Compute Equation (21)
    Compute Equation (19)
    Compute Equation (20)
 END IF
 Recover mean
 IF the number of active features > Threshold THEN
  SEIF_Sparse();  //sparsification
 END IF
END WHILE