An Anchor-Based Pedestrian Navigation Approach Using Only Inertial Sensors

Abstract

In inertial-based pedestrian navigation, anchors can effectively compensate the positioning errors originating from deviations of Inertial Measurement Units (IMUs), by putting constraints on pedestrians' motions. However, these anchors often need to be deployed beforehand, which can greatly increase system complexity, rendering it unsuitable for emergency response missions. In this paper, we propose an anchor-based pedestrian navigation approach without any additional sensors. The anchors are defined as the intersection points of perpendicular corridors and are considered characteristics of building structures. In contrast to these real anchors, virtual anchors are extracted from the pedestrian's trajectory and are considered as observations of real anchors, which can accordingly be regarded as inferred building structure characteristics. Then a Rao-Blackwellized particle filter (RBPF) is used to solve the joint estimation of positions (trajectory) and maps (anchors) problem. Compared with other building structure-based methods, our method has two advantages. The assumption on building structure is minimum and valid in most cases. Even if the assumption does not stand, the method will not lead to positioning failure. Several real-scenario experiments are conducted to validate the effectiveness and robustness of the proposed method.

Keywords: Rao-Blackwellized particle filter; anchor; building structure; pedestrian navigation.

Figure 1

Overall framework of the proposed method.

Figure 2

Anchor definition. (a) The real-scenario Perpendicular Intersectant Corridors (PIC) from a typical office building; (b) the floorplan of a PIC.

Figure 3

(a) A trajectory sample for MDL principle formation. $p_{c1}$ and $p_{c2}$ are characteristic points; (b) definition of the perpendicular distance $d_{\perp }$ and angle distances $d_{\theta }$ between two lines; (c) real-scenario trajectory partitioning result.

Figure 4

An example of anchor extraction from corridor-walks.

Figure 5

Anchor extraction in different turning patterns.

Figure 6

The Probability of Detection (POD) and the False Alarm Rate (FAR) curves. The x-axis denotes the threshold over which a line segment is identified as a corridor-walk.

Figure 7

The generative probabilistic model for filtering.

Figure 8

Noise assumption in the observation model.

Figure 9

(a) The different values for the parameter $k$ in the rewarding function correspond to different extents of convex; (b) the curve of positioning errors against different values of $k$.

Figure 10

The chosen rewarding function for particle weight updating.

Figure 11

The insole-shaped MIMU (Multiple Inertial Measurement Unit) module.

Figure 12

Test one: Office building scenario with abundant anchor revisits (a) Raw trajectory with multiple iterations; (b) calibrated trajectory with better consistency than raw trajectory.

Figure 13

Test two: Library scenario with few anchor revisits (a) Raw trajectory with only one anchor revisit; (b) calibrated trajectory with better consistency in the circled area.

Figure 14

Test three: Library scenario with no anchor revisits. The accuracy of our method degenerates to the raw trajectory.

Figure 15

Trajectories for our method, the HAH method and raw trajectory. Our method has better accuracy than HAH and raw trajectory.

Figure 16

Raw trajectory and extracted anchors. The trajectory is designed to have intensive extracted anchors to test the robustness of the method when wrong anchor association is unavoidable.