Advanced KSOM based Redundancy Resolution of a Mobile Manipulator System for Motion on an Uneven Terrain

Redundancy Resolution of a Mobile Manipulator Using Learning Based Algorithm

“…In this investigation, we have proposed an advanced KSOM-based network with a WLN matrix to address redundancy in the MM. The MM training data were generated using forward kinematics [14,28]. The complete workspace can be learned using the inverse kinematics of the KSOM-based learning approach.…”

“…Each neuron is associated with a weight vector w n , which discretizes the input space, and a joint angle vector θ n , which discretizes the output space . A n is a proportional linear map or an approximated inverse Jacobian mapping between the input and output spaces [14,28]. The suggested system's 3-D KSOM lattice structure is shown in Fig.…”

“…Redundancy resolution has been studied in the past using various approaches. For addressing inverse kinematics resolution problems, the commonly used methods are include neural network-based solutions and the pseudo-inverse Jacobian matrix (PIJM) [11][12][13][14]. Different researchers, such as Wiedmeyer et al [16], Zlajpah et al [17], Basso et al [18], Chembuly et al [19], and Aldas et al [4], respectively, studied redundancy resolution using optimization techniques, task decomposition, traditional task priority, and kinematic and dynamics models of trajectory tracking.…”

“…The mobile platform's linear and angular velocities are calculated using the left and right wheels' angular displacements. In this study, those angular displacements are determined using the KSOM network, whereas, in previous literature, the linear and angular velocities were determined from the joint angles after determining those angles from the network [14,15]. This study's main contribution is as follows: 1) Design of KSOM-based WLN for the inverse kinematics of the seven-DOF MM redundancy resolution.…”