Does close proximity robot motion tracking alter gait? John D. Yamokoski, Scott A. Banks * Department of Mechanical & Aerospace Engineering, University of Florida, Gainesville, FL, United States 1. Introduction A wide variety of measurement techniques have been employed to measure the dynamic motions of human joints [41]. These techniques have included computed tomography (CT) (e.g. Ref. [12]) and magnetic resonance (MR) (e.g. Ref. [14]) imaging, radiostereometric analysis (RSA) [22], in vitro cadaver tests (e.g. Ref. [8]), mechanical linkages (e.g. Ref. [10]), and skin [9] and bone-pin (e.g. Ref. [26]) mounted markers (both electromag- netic and optical). While all of these technologies have greatly contributed to our understanding of human (and non-human) kinematics, each method has distinct limitations with regard to accurate measurement of skeletal kinematics. RSA and CT/MR methods provide excellent spatial resolution, but limited dynamic capability. In vitro methods suffer from uncertainty of muscle and applied loads. Marker-based motion capture suffers inaccuracy from skin movement relative to the underlying bones [17]. Fluoroscopic imaging and model-image registration techni- ques were introduced almost twenty years ago to provide direct observation of dynamic joint motion [6]. These techniques provide a measurement tool to quantify in vivo three-dimen- sional (3D) joint motion during dynamic, weight-bearing activities that is unaffected by skin and soft-tissue motion (Fig. 1). These methods have been adopted and evolved by numerous groups worldwide. Searching peer-reviewed publica- tion databases with the keywords ‘‘kinematic* and fluoroscop*’’ returns over 350 results and one of the author’s first papers in this area now has over 170 citations [3]. Despite the utility of these methods, they still rely upon imaging equipment designed for use in surgery or interventional radiology (e.g. Refs. [7,25,16,24,37]). Thus, activities are limited to those that can be observed within a small field of view within the workspace of a typically immobile piece of equipment. Many other papers have reported fluoroscopic analysis of knee kinematics during ‘gait’, but none of these was conducted during free-speed overground normal walking. Banks’ group has reported the use of a treadmill and multiple positionings of the stationary fluoroscope [4]. Komistek and Mahfouz’s group has reported on the initiation of gait (the first step) using a fixed imaging system [43,23]. Li’s group reported on treadmill or gait initiation steps, but not overground normal gait [28]. Gait & Posture 34 (2011) 508–513 A R T I C L E I N F O Article history: Received 18 June 2010 Received in revised form 7 July 2011 Accepted 9 July 2011 Keywords: Gait analysis Dynamic radiographic imaging Fluoroscopy Kinematics Kinetics A B S T R A C T Dynamic fluoroscopic imaging and three-dimensional model-image registration techniques provide detailed joint kinematic measurements for motions constrained to small volumes of space. Several groups are working to mount radiographic imaging hardware onto mobile platforms to provide these same imaging capabilities for observation of unrestricted activities. These dynamic radiographic imaging systems could provide accurate skeletal kinematics during a wide range of clinically relevant, daily activities. However, the premise that people move naturally when followed by a dynamic imaging system has not been evaluated. The goal of this study was to determine if a close-up robot tracking system affects natural free-speed gait. 14 healthy adults were recruited to walk through the workspace of a dynamic radiographic imaging system. Randomized walking trials were performed with and without the dynamic imaging system tracking the motions of the subject’s left knee. With- and without-robot trials were compared using detailed temporal–spatial and frequency analysis of kinematic and kinetic parameters. On average, participants increase their stride length by 0.9 cm. There also were slight increases in unexplained variation in ankle flexion/extension and ground reaction forces compared to baseline measurements. The statistically significant differences indicate that, on average, people tried to move faster through the workspace of the dynamic radiographic imaging system while it was actively tracking their motion. These differences are small and potentially clinically irrelevant. ß 2011 Elsevier B.V. All rights reserved. * Corresponding author at. Department of Mechanical & Aerospace Engineering, University of Florida, MAE-A 318, Gainesville, FL 32611-6250, United States. Tel.: +1 352 392 6109; fax: +1 352 392 7303. E-mail address: banks@ufl.edu (S.A. Banks). Contents lists available at ScienceDirect Gait & Posture jo u rn al h om ep age: ww w.els evier.c o m/lo c ate/g aitp os t 0966-6362/$ – see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2011.07.002