Augmenting Virtual Environments with Physical Stats
When you are engaged in a virtual recreation of a physical event, particularly sporting events, and you think you are getting quite good. It could be a huge boost to the ego, or a crippling slap, to find yourself suddenly competing with the world's greatest in that sport, using their latest data.
For those who consider themselves the best, it could be a wonderful experience. For everyone else, to have their nose rubbed in the dirt, and showcased just how bad they really are, would be like a wet slap.
Yet, that is what one company is actively doing. New start-up iOpener Media, with Andy Lurling at the helm, is working to record real-time positional data from F1 races, and telecast that to virtual gameworlds, so that gamers can race in real-time against their heroes.
"It's clear that the next trend in gaming is going to be bringing real objects into the virtual world; playing not against other gamers but people doing the real thing," Andy stated.
The European Space Agency (Esa) was so impressed with his proposal, it gave the company a grant to develop a proof of concept. A German venture capitalist has stumped up cash to develop it further.
The firm is currently in talks with six developers about using the technology.
Whilst F1 racing is the initial use, the concept would work with any data stream.
Differential GPS is used, with a network of fixed base stations to correct GPS signal from each tracked object (car), to narrow down accuracy to centimetre precision.
Other tweaks include fitting cars with an inertial measurement unit (IMU), commonly used in guided missile systems, which measure acceleration, angle and yaw of the object.
From the track side, the data is sent over the net to a server farm, where it is saved before being pumped out - or "mediacast" - to eager gamers.
The delay between collecting the data and the gamer being immersed in it is up to five seconds, similar to the lag on a TV broadcast. Race data is also stored so a race can be replayed at a later date. This in particular makes it a wonderful tool for training.
Finally, iOpener Media provides AI code for the local rendering, so that drivers from the physical world, do not pass straight through your car in the simulation. Instead, they flow round your car, with the AI taking over when they get close to you, and steering the car as close as possible to its ideal position. In groups, the cars react to one another. Once at least one nearby car is under AI control, and heading towards the others, those others will also assume AI control long enough to avert a collision. In this way, realistic flow control is maintained.