I’m representing a research group at our uni doing a proof-of-concept of a telepresence robot. Something sturdier than a Roomba and selfie-stick but nothing huge.
We have bought a Ricoh Theta V camera for 360 degree view of the environment. However, we are having trouble getting live streaming to work from the camera to a client across the interwebs.
The Ricoh Developer Console website is broken, so we can’t get client credential for any of the web apis and examples
Ricoh won’t answer my support requests.
Ricoh seems to have discontinued a lot of their APIs regarding live streaming, which seems to complicate traversing the NAT quite a lot.
I tried to use the USB SDK to connect to the camera by wire to atleast get some stream working, but the ready binaries don’t find the connected device and the SDK is using such old build tools that wxWidgets won’t compile on my machine.
How can we get a simple video stream from the robot (running Windows 10 on Intel NUC mini-PC) to a user interface over the internet?
We need to support only one-way video and voice, just data channel commands back from the user interface. Preferably, the camera could live-stream for 1-2 hours at a time, with minimum setup. So no manual button presses and phone configurations if possible. It can be powered by usb and a fan/heatsink can be arranged, but booting this up should be able to be automated via software as much as possible.
Any help on how to get started would be great. We have experience in programming embedded devices, linux and JavaScript, but not that much in Android, so if that’s what it takes, examples and resources there would be nice too. It feels like we are at a dead-end with this device now that all we get out of Ricoh’s site is broken links and still-photos when live-streaming was supposed to be supported with examples.
As you’re using Windows 10, first try get it working on a laptop as it might be easier to debug than the NUC on the robot (as a first step).
You likely still need the THETA V driver for Windows 10 at the link below. It seems like 1,500 people went to get the driver from our forum alone, so it’s likely still needed.
Other ideas for implementation
–
The Lockheed Martin Amelia Drone was also done by an undergraduate group at the Rochester Institute of Technology in NY. There is a lot of great info linked to from that thread. I really think you’ll find some great stuff there.
Unfortunately, due to COVID-19 in NY, the group cannot meet. Glad to hear that your research group is still active.