Contact type dependency of texture classification in a whiskered mobile robot

Abstract

Actuated artificial whiskers modeled on rat macrovibrissae can provide effective tactile sensor systems for autonomous robots. This article focuses on texture classification using artificial whiskers and addresses a limitation of previous studies, namely, their use of whisker deflection signals obtained under relatively constrained experimental conditions. Here we consider the classification of signals obtained from a whiskered robot required to explore different surface textures from a range of orientations and distances. This procedure resulted in a variety of deflection signals for any given texture. Using a standard Gaussian classifier we show, using both hand-picked features and ones derived from studies of rat vibrissal processing, that a robust rough-smooth discrimination is achievable without any knowledge of how the whisker interacts with the investigated object. On the other hand, finer discriminations appear to require knowledge of the target's relative position and/or of the manner in which the whisker contact its surface. (146 words)

Biomimetic tactile target acquisition, tracking and capture

“…(rotation around its base, leading to 'protraction' or forward movement of its tip), and is instrumented for measuring deflection in two axes (denoted x and y) using a hall-effect sensor and a small powerful magnet glued to the base of the whisker shaft. We can assume, here, that all software components are running at the same rate of f = 20Hz, leading to response times measured in increments of T = 50ms (the much higher frequency data available from the platform are not relevant to this study since we are not discriminating fine spatial or temporal detail such as texture or movement [21]). Sample number is denoted n ∈ {1, 2, ...}.…”

“…Previous implementations of artificial whisker sensing, reviewed in [14,15], have shown them to be useful for navigation [16,17] as well as for object identification/localisation [5,[18][19][20]. At the same time, robots with whiskers are helping biologists to understand the nature of the task facing biological whisker specialists [15,21].…”

Towards hierarchical blackboard mapping on a whiskered robot

“…Though reliable features can be extracted for radial distance estimation in this paper, and contact speed on a stationary robot [13], it is unclear what other features can be extracted from whisker deflection signals for discriminating different kinds of object properties. In our own lab we are developing features for whisker based tactile sensing of contact geometry [14] and texture [17]. In future we hope to be able to combine features for diverse tactile properties in rich environments into a coherent system onboard a mobile robot, which in turn would provide reports that could be used as inputs to hierarchical object models as presented in this study.…”

“…In contrast, shaft contacts are less informative. For example, an unknown distance to an object along the shaft can confuse attempts to classify surface orientation and texture [17]. Shaft contacts are rare in practice in both rodents and mobile robots, occurring only when small objects enter the field of multi-whisker arrays between the whisker tip points.…”

“…Individual components of such a system have previously been investigated in isolation, including whiskered texture recognition [15], [26], [11], [17], [32], surface shape recognition [30], [24], [21], [13], and object recognition [19]. These components have previously been tested under ideal laboratory conditions or in individual mobile settings [41]; here we present steps integrating them into a single platform for hierarchical object recognition, along with results and observations on their performance 'in the wild' in a common arena environment.…”