1: \begin{abstract}

2: Two novel nonlinear pose (\textit{i.e}, attitude and position) filters developed directly on the Special

3: Euclidean Group $\mathbb{SE}\left(3\right)$ able to guarantee prescribed

4: characteristics of transient and steady-state performance are proposed.

5: The position error and normalized Euclidean distance of attitude error

6: are trapped to arbitrarily start within a given large set and converge

7: systematically and asymptotically to the origin from almost any initial

8: condition. The transient error is guaranteed not to exceed a prescribed

9: value while the steady-state error is bounded by a predefined small

10: value. %

11: \begin{comment}

12: The constrained error is converted to its unconstrained form, termed

13: transformed error.

14: \end{comment}

15: {} The first pose filter operates based on a set of vectorial measurements

16: coupled with a group of velocity vectors and requires preliminary

17: pose reconstruction. The second filter, on the contrary, is able to

18: perform its function using a set of vectorial measurements and a group

19: of velocity vectors directly. Both proposed filters provide reasonable

20: pose estimates with superior convergence properties while being able

21: to use measurements obtained from low-cost inertial measurement, landmark

22: measurement, and velocity measurement units. The equivalent quaternion representation and complete implementation

23: steps of the proposed filters are presented.  Simulation results demonstrate

24: effectiveness and robustness of the proposed filters considering large

25: error in initialization and high level of uncertainties in velocity

26: vectors as well as in the set of vector measurements.

27: \end{abstract}

28: