Learning user models of mobility-related activities through instrumented walking aids

Abstract

Abstract-We present a robotic walking aid capable of learning models of users' walking-related activities. Our walker is instrumented to provide guidance to elderly people when navigating their environments; however, such guidance is difficult to provide without knowing what activity a person is engaged in (e.g., where a person wants to go). The main contribution of this paper is an algorithm for learning models of users of the walker. These models are defined at multiple levels of abstractions, and learned from actual usage data using statistical techniques. We demonstrate that our approach succeeds in determining the specific activity in which a user engages when using the walker. One of our proto-type walkers was tested in an assisted living facility near Pittsburgh, PA; a more recent model was extensively evaluated in a university environment.

A probabilistic approach to learn activities of daily living of a mobility aid device user

“…al. [3] to learn users' high level activities which were based on the topological location of the user and the most likely activity a user could perform at the respective location. The HSMM framework operated at three different levels: at the lowest level the metric motion is described by metric coordinates, at the mid-level the framework uses topological regions as its element and at the highest level the person's activities are divided into logically broader activities.…”

“…[20], [3], [21]), propositions have been mostly aimed at predicting either the fundamental user activities broadly aligned with what we categorise in our work as APs, or more abstract long term activities (as in the case of Glover et. al.…”

“…al. [3]). In the approach suggested in this work, an HHMMbased model is used not only to infer APs, but also exploits these APs in a more sophisticated hierarchical structure to construct and identify overall activities of daily living (e.g.…”

“…Automated assistive devices like guiding canes [1], robotic wheelchairs [2], walkers [3] [4] and manipulators [5] are all designed to make life of elderly and disabled people more comfortable by helping them in their everyday activities, enabling them to live more independently and ultimately improving their overall quality of life. With a marked demographic shift at a global level indicating that the worldwide proportion of people aged 60 and over is expected to double between 2000 and 2050 [6], providing the elderly and frail with assistive devices that can enhance their mobility needs in particular has been a strong motivation for researchers to develop more advanced technological aids through alternative or augmentative aid devices [7].…”

User-adapted plan recognition and user-adapted shared control: A Bayesian approach to semi-autonomous wheelchair driving

“…The goal of the framework is to find the Maximum a Posteriori motion pattern given a trajectory in order to predict the motion of people. Glover et al (2004) estimate the activity a person is engaged in when using a walking aid. They adopt a Hidden Markov Model (HMM) that integrates metric, topological and temporal information into a probabilistic estimate over activities.…”