Robotic Ubiquitous Cognitive Ecology for Smart Homes

Abstract

Robotic ecologies are networks of heterogeneous robotic devices pervasively embedded in everyday environments, where they cooperate to perform complex tasks. While their potential makes them increasingly popular, one fundamental problem is how to make them both autonomous and adaptive, so as to reduce the amount of preparation, pre-programming and human supervision that they require in real world applications. The project RUBICON develops learning solutions which yield cheaper, adaptive and efficient coordination of robotic ecologies. The approach we pursue builds upon a unique combination of methods from cognitive robotics, machine learning, planning and agentbased control, and wireless sensor networks. This paper illustrates the innovations advanced by RUBICON in each of these fronts before describing how the resulting techniques have been integrated and applied to a smart home scenario. The resulting system is able to provide useful services and pro-actively assist the users in their activities. RUBICON learns through an incremental and progressive approach driven by the feedback received from its own activities and from the user, while also self-organizing the manner in which it uses available sensors, actuators and other functional components in the process. This paper summarises some of the lessons learned by adopting such an approach and outlines promising directions for future work.

Unsupervised feature selection for sensor time-series in pervasive computing applications

“…In particular, we compare the performance of ICF with respect to the CleVer method, a state-of-the-art unsupervised feature filter for time-series, for a varying proportion of irrelevant features in the original MTS. 1 To this end, we have employed real-world data collected in two experimental deployments, one associated to mobile robot navigation and one related to human activity recognition (HAR).…”

“…This work has been developed in the context of the RUBICON project [1], which proposes a vision of a wireless sensor and actuator network where data analysis and learning capabilities are spread in all components of the systems, depending on the capabilities of the hosting devices. To this end, it defines a pervasive learning system that consists of a network of learning modules distributed on devices characterized by limited computational and communication capabilities.…”

Concentric ESN: Assessing the Effect of Modularity in Cycle Reservoirs

“…Reservoir Computing (RC) [1] is a paradigm for recurrent neural network design that, in the latter years, has found wide application thanks to its (relative) simplicity and effectiveness in dealing with sequential data processing tasks, such as with sensor network data [2], [3],mobile robot navigational data [4], [5], telephone load forecasting [6], ambient assisted living [7], [8], biomedical data [9], [10], etc. RC is based on a conceptual separation between the recurrent part of the network, i.e.…”

A Benchmark Dataset for Human Activity Recognition and Ambient Assisted Living

“…The EU project RUBICON [4,21] develops cognitive robotic ecologies, i.e. systems made out of wireless sensors, actuators and robots, each equipped with perception, attention, memory, action, learning, and planning capabilities.…”