Researchers from RMIT in Melbourne, Australia have developed a flying running companion called Joggobot. The system uses the built-in camera on a commercially-available Parrot AR Drone quadrocopter to track the position of a jogger, and fly a few feet out in front. While the current version has some serious limitations, there is huge potential for the development of a fully interactive training partner or coach in the very near future.

Joggobot is the work of Eberhard Gräther and Florian Mueller, with assistance from Wouter Walmink, Chad Toprak, Josh Platt, Conor O’Kane, Jennifer Lade, Jonathan Duckworth, Wendy Ju and Wolfgang Gräther, and has been developed and built as part of Mueller's Exertion Games Lab project within the Games Program at the Royal Melbourne Institute of Technology (RMIT). Mueller is one of a team of researchers who have already developed a distance jogging solution which allows runners from different geographical locations to run together.

Runners using this spatial audio system each wear headphones linked to mobile phones and other gadgetry that determines the performance of each partnered jogger. If the system detects that one of the pair is falling behind the pace set by the other, the voice of the that runner actually sounds like it's coming from somewhere behind the one out in front. When both runners are keeping pace with each other, the audio is adjusted appropriately.

Heading out with a running partner has been found to be far more engaging and often more challenging than jogging alone, but a partner or trainer may not always be available when your timetable dictates. Joggobot gives the runner a physical object to interact with while out on a solo run, and is perhaps a much more engaging proposition than merely following a route on a smartphone/in-shoe-based training app, or so-called ghost runners on wrist-based solutions like the Forerunner fitness watch from Garmin, where you race against an avatar version of yourself from an earlier run.

"Our aim is to explore how robots should be designed to create engaging jogging experiences in order to increase our understanding of how to design robots for exertion activities," explained the research team.

Joggobot is capable of tracking the jogger using an integrated, front-mounted camera and some tag detection software that looks out for special markers printed on the runner's t-shirt. The quadrocopter takes off when the camera registers the markers and rises to about the same height as the tags on the t-shirt, its altitude being determined by an internal sensor.

Instead of being controlled by a wireless remote, the flying companion is able to autonomously react to the actions of the jogger courtesy of some custom control software based on the AR Drone open API. A relative distance of about three meters (9.84 feet) from the runner is maintained while in flight, and if Joggobot loses sight of the markers, it safely lands.

For the moment, the Joggobot system is reported to be limited to a straight line run but the researchers are planning on looking at varied and complex jogging patterns in the future, and have already started testing the use of special gestural movements and signals to control or instruct Joggobot during a run.

I can think of a number of useful applications for this development beyond giving lone joggers a partner or coach. Pairing the technology with a smartphone (via Bluetooth, for instance) would not only cater for streaming or sharing of training footage but could also greatly enhance the safety of the runner by alerting emergency services with location details and a live video feed in the event of an accident or if the jogger should run into other trouble.

Source: Exertion Games Lab