Page 185 - Autonomous Mobile Robots
P. 185

Landmarks and Triangulation in Navigation                  169

                              and dual landmark case. The method proposed here is the extension of Boley’s
                              LSE [24].

                              4.4.3.1 Single-landmark LSE (SLSE)
                              The single-landmark LSE algorithm is based on the coordinates shown in
                              Figure 4.9a. The origin point is placed at the landmark, and the x-axis is set
                              parallel to the robot moving trajectories. From each sampling point on a robot
                              trajectory, the bearings to the landmark are measured as α i (i = 0, 1, ... , n),
                              by using the methods in Section 4.2. The position of each sample is noted in
                                                T
                              the vector z i = (x i , y i ) , the distances between each position, and z 0 , which
                              can be obtained from the readings of odometry, are noted as d i = z i − z 0 =
                                       T
                              (x i − x 0 , y 0 ) . It is easy to observe that:
                                                              x 0 + d i
                                                      tan(α i ) =                      (4.29)
                                                                y 0
                              Rewriting Equation (4.29), we have:

                                              x 0 cos(α i ) − y 0 sin(α i ) =−d i cos(α i )  (4.30)

                              Row-by-row collecting all the equations for i = 1, ... , n, wehaveover
                              determined equations which can be expressed as:

                                                          Az 0 = b                     (4.31)

                              where
                                                                                    
                                       cos(α 1 )  − sin(α 1 )                −d 1 cos(α 1 )
                                       cos(α 2 ) − sin(α 2 )                 −d 2 cos(α 2 )
                                                                                    
                                          .        .    ,  z 0 =    ,           .
                                                               x 0                  
                                          .        .             y 0             .
                                 A =                                  b =            
                                         .        .                           .     
                                       cos(α n ) − sin(α n )                 −d n cos(α n )
                                 Using the Least Square method, we can estimate z as follows:
                                                            T  −1 T
                                                      z 0 = (A A)  A b                 (4.32)
                              In this method, we adopt samples at the positions z i (i = 1, ... , n)asthe
                              reference sample (RS) to estimate the robot position z 0 .

                              4.4.3.2 Dual-landmark LSE (DLSE)
                              In this case, the coordinates are built on two landmarks. The original point is set
                              to one landmark and the x-axis point to the other one. The coordinate values of




                              © 2006 by Taylor & Francis Group, LLC



                                FRANKL: “dk6033_c004” — 2006/3/31 — 16:42 — page 169 — #21
   180   181   182   183   184   185   186   187   188   189   190