Multi-Hypothesis SLAM for Non-Static Environments With Reoccurring Landmarks

Abstract

A static world assumption is often used when considering the simultaneous localization and mapping (SLAM) problem. In reality, especially when long-term autonomy is the objective, this is not valid. This paper studies a scenario where uniquely identifiable landmarks can attend multiple discrete positions, not known a priori. Based on a feature based multi-hypothesis map representation, a multi-hypothesis SLAM algorithm is developed inspired by target tracking theory. The creation of such a map is merged into the SLAM framework allowing any available SLAM method to solve the underlying mapping and localization problem for each hypothesis. A recursively updated hypothesis score allows for hypothesis rejection and prevents exponential growth in the number of hypotheses. The developed method is evaluated in an underground mine application, where physical barriers can be moved in between multiple distinct positions. Simulations are conducted in this environment showing the benefits of the multi-hypothesis approach compared to executing a standard SLAM algorithm. Practical considerations as well as suitable approximations are elaborated upon and experiments on real data further validates the simulated results and show that the multi-hypothesis approach has similar performance in reality as in simulation.

Enhanced Berth Mapping and Clothoid Trajectory Prediction Aided Intelligent Underground Localization

“…Therefore, before conducting the experiments on retrieving the nearest map nodes, it was essential to determine a reasonable value for K through experimentation. In this paper, Group 1 from each dataset was selected for experimentation, and an optimal value for K was chosen within the threshold range [2,8]. The experimental results are presented in Table 4.…”

“…Localization is vital for connected and automated vehicles (CAVs) [1,2]. Satellite-based localization, reliant on global navigation satellite system (GNSSs), encounters difficulties in satellite-denied areas like underground parking scenes [3].…”

Hypothesis selection with Monte Carlo tree search for feature-based simultaneous localization and mapping in non-static environments

“…In our previous work multiple hypotheses in the map representation are introduced and the exponential growth in the number of hypotheses is limited by taking decisions based on statistical tests (Nielsen and Hendeby, 2022b). This however results in a somewhat overwhelming book-keeping of old hypotheses and a solution that does not scale well in the number of hypotheses since each hypothesis requires a dedicated SLAM solver.…”

“…Changes in an environment can be categorized into two main categories (Nielsen and Hendeby, 2022b):• Dynamic changes, with objects moving relatively fast so that the perception of the object changes while the object is in field-of-view.• Non-static changes (sometimes also referred to as semi-static changes (Rosen et al, 2016)), that are much slower and typically not captured by sensors in real time, but rather discovered when an area is revisited.…”

“…Also outdoor scenarios have possible applications where construction works can temporarily alter the environment or parking lots can be occupied or not. In Nielsen and Hendeby (2022b) it also described how this type of changes have high relevance in an underground mine environment, and it is shown that it is beneficial to somehow incorporate this type non-static changes into the SLAM model. This paper uses, the map representation presented in Nielsen and Hendeby (2022a), and further developed in Nielsen and Hendeby (2022b), where each landmark is modeled by discrete modes where only one can be true at each time instance.…”