Remote maintenance consultation is challenging. It is hard to identify problems without seeing the target machine and difficult to describe the needed procedures verbally. Visual interaction would help, but video stream requires good networks that are not always available.
In our low-bandwidth remote maintenance assistance system there is no video stream, instead we transfer still images and hand gesture coordinates. Service support uses virtual reality view and gesture recognition to remotely help maintenance operation. These gestures are seen in real-time on-site using augmented reality overlaid on the video view in a mobile phone.
This demo was implemented by a three person team from VTT in a Hackathon 2016 organized by Ponsse. The work was supported by FIMECC S-Step project.