is introduced to avoid singular configurations given by det(T(a)W-1TT (a)) = 0. To avoid division by zero, s > 0, is chosen as a small number, while p > 0 is scalar weight. A large p ensures high maneuverability at the cost of higher power consumption and vice versa. The optimization problem (1.44) is a non-convex nonlinear program and it requires a significant amount of computations at each sample (Nocedal and Wright, 1999). Consequently, the following two implementation strategies are attractive alternatives to nonlinear program efforts.
Was this article helpful?