One of the obsessions was concerning how to incorporate positional tracking in an inherently limited environment. Given that your feet are essentially glued to the foot pedals, some of the challenges related to tracking are greatly reduced. I'm guessing some of you might be wondering what good is position tracking if you're not actually free to physically walk around. It's true that with a foot controller, you're essentially tied to one spot, but your body can still sway a bit from side to side, and forwards and back. Those small motions are significant enough due to parallax effects, which can give you a greater sense of "being there".
My original thought was using a mechanism like the GameTrak in the pedal assembly, or perhaps mounted a bit higher perpendicular to the floor with the cables connected to a chest strap with "wings" mounted behind. The GameTrak's simplicity is quite ingenious but it has some limitations. More importantly, I just wanted to avoid patent issues, so I tried figuring out some variant to side-step it. While I was moving around on a make-believe pedal board trying to visualize the motions and how to sense them, I realized that I was shifting weight from one foot to another, or towards the front or rear of my feet - the very thing that wii balance board, or more generally, a platform with independent weight sensors in each corner is designed to detect.
Now on its own, it's only of limited use - it can't provide all the movement controls we need, or even sense some of the body motions for positional tracking (example: it can't differentiate between leaning your torso to the side and shifting your body to the side). The simplest solution to the latter problem seems to be a chest-mounted orientation sensor. There are additional benefits to the chest sensor as well - such as being a reference point for skeletal modelling in conjunction with arm mounted sensors.
Ok, so the balance board is essentially a 2 axis sensor. That means we can incorporate forwards/reverse, and either strafing or turning. If we want to enable both of those, the board needs an extra axis. My thought is a twistable pedal under each foot. Realistically, only one would probably be needed, but two adds a lot more flexibility in terms of control schemes which is what I want to get into next.
So if a weight-shift sensing board is used for position tracking, then how can it be used to control movements? Well, there's actually quite a few ways - none of them more correct than the other, but more a matter of preference - not unlike how there's several control schemes available with a single game pad. It should offer enough flexibility to find various methods that suit most people. Here are a few options:
1) Moving side to side and front to back evenly is only for slight position tracking such as looking around a corner or off the edge of a cliff without controlling player movement. In order to actually move your character forwards or backwards, weight is shifted towards the front or back of a specific foot. It would take some getting used to in order to prevent actual movements when all you want to do is look around, but I think it's workable. In order to turn and strafe, each twist pedal would be assigned to one of the tasks.
2) Enable or disable movement with the push of a button on the hand controls. For example, while not pressing a button on a hand controller, any swaying on the platform is used to detect body position, though twisting either of the pedals can still turn your character. When you press the button, and lean forward or backward it will move your character accordingly, while shifting your weight side to side causes you to strafe (or turn). The remaining motion can be controlled with the twist pedals.
Those essentially cover the premise. There are several variations or combinations to the two examples above which makes the system quite flexible. So while a weight shift sensing board can be effective for character control, it provides almost everything needed for basic position sensing too at no additional cost (other than the chest sensor, which as I mentioned before offers other potential benefits).
I think such a system including arm tracking and hand controllers (with just buttons - no analog sticks needed) could be sold for under $150 based on my estimates.
