Retrieving and reusing qualitative cases: An application in humanoid-robot soccer

Abstract

This paper proposes a new Case-Based Reasoning (CBR) approach, named Q-CBR, that uses a Qualitative Spatial Reasoning theory to model, retrieve and reuse cases by means of spatial relations. Qualitative relations between objects, represented in terms of the EOPRA formalism, are stored as qualitative cases that are applied in the definition of new retrieval and reuse algorithms. The retrieval algorithm uses a Conceptual Neighborhood Diagram to compute the similarity between a new problem and the cases in the case base, and to select the most similar case. The reuse algorithm uses a composition algorithm to calculate the adapted position of the agents based on their frame of reference. The proposed approach was evaluated on simulation and on real humanoid robots. Results suggest that this proposal is faster than using a quantitative model with numerical similarity measurement such as the Euclidean distance. As a result of running Q-CBR, the robots obtained a higher average number of goals than those obtained when running a metric CBR approach.

Qualitative case-based reasoning and learning

“…Due to the large variety of distinct QSR formalisms, these methods find applications in a number of domains, such as robot navigation and self-localization [57], geographic information systems [20], cognitive linguistics [66,53] among others [13,70]. The present paper builds upon our previous work where an extension of the OPRA, the Elevated Point Relation Algebra EOPRA [42], was applied to represent the agents in the robot-soccer domain [27,28]. EOPRA assigns an intrinsic orientation to objects and defines a qualitative distance based on an elevated point in the domain, that could be defined as the height of the observer [42].…”

“…3 B represents that both A and B are discretised into 16 orientation relations (4m) and 8 distance relations (2m). For the relative orientation, A is in the sector 1 of B and B is in the sector 13 of A; and for the relative distance, A is in the sector 5 of B and B is in the sector 3 of A [28]. In this work, these sectors are grouped into regions, reducing the number of boundaries.…”

“…In our previous work, the cases were hand-coded and the focus was on the retrieval and reuse processes [27,28]. So, in order to verify the learning phase, we have analysed the behaviour of the RL method with discretised states [29].…”

“…QCBRL creates a compact world model representation and solves the exponential space complexity problem described in Albrecht and Stone [3]. Case retrieval and reuse are executed assuming Conceptual Neighbourhood Diagrams (CND) [23] and a qualitative similarity function, that computes the similarity between a new problem and an element of the case base, retrieving the most similar case to a given situation and reusing its solution to solve a new problem [28]. New cases are learned by executing a partial RL method; that is, when no similar case is retrieved, QCBRL iterates the RL method until the occurrence of the first successful episode.…”

Guided Navigation from Multiple Viewpoints using Qualitative Spatial Reasoning

“…However, in contrast to the work presented in this paper, the authors did not develop a way of guiding an autonomous agent using the qualitative information embedded in the rules. OPRA was also used in Homem et al (2017) and Bianchi et al (2018) in a case-based reasoning system for retrieving and reusing qualitative cases for situations involving multiple robots. That work, however, did not assume any probabilistic filtering method and, thus, is prone to sensor and motor noise.…”