[paper] proof of concept pendulum joyrider

With no further fuss, you can skip to the files.


The archive attached includes SketchUp 8 & SkP 3.2 schematic model of the sim, and an HTML5 proof of concept interactive demo, which shows the principle behind the idea. The demo only have very basic formula for computing the motor drive speeds and it doesn't have a heave drive, but it's only there to give you an idea how it should work. The SketchUp model doesn't even have that, but it's attachable to the SimTools, so you can attach to it roll, pitch, surge, sway and heave directly and it will roughly move around, giving you basic idea.

First things first, that kind of build is not for noobs (i.e. me) because it takes ridiculous amounts of power to drive it. With this specific model, it would take about 5 kW per pitch and roll motors and 1.5 kW for heave motor. There's immediately obvious solution of decreasing the power requirements by making the pendulum arm shorter, but it should still be sufficiently long in order to make the difference from standard joyrider, which is the whole point.

This rig attempts to simulate car motion by simulating forces acting on the driver. All the forces, accelerations, bumps etc., all the car motions are ultimately creating forces on the seat and the driver experiences forces going in opposite direction due to his body's inertia. One cunning thing about it is that all the multitude of simultaneously acting forces add up together to a single superposition force-vector that actually acts on the driver. So the motor controller attempts to carefully calculate that vector and attempts to re-create it. Doing it properly would result in rather accurate simulation of all possible car motions. This doesn't accounts for rotational accelerations, though, and with this kind of rig it will such equation (rotaitonal acceleration immediately creates forces which is undesired) has no solutions. But since rotational accelerations in the car are typically very small, they shouldn't pay too much role, and my experience with 401 motion simulator rig (with yawwing) serves a proof to that.

I'm not exactly sure why I should post this because it's lightyears away from "fully baked", but I decided to share the idea and hear suggestions.

This is a wise decision on your part since many of us have already built rigs and have experienced what works and what doesn’t.

Your theory here is correct except for one thing! While yaw may not provide much of a noticeable acceleration on the 401, your redesigned joyrider will create very undesirable roll and pitch rotations with your pivots being near or above the head. Your legs and feet will be swung around intensely, which completely dissolves the immersion imo. I’m not sure how you came to the conclusion that this would be a good idea after you posted this about the FD401:

What makes you think having the pivot points above the center of mass of the driver is going to be any different than having them below? The best method imho to practically eliminate the feel of roll rotation is to have the roll axis run from your belly button to your feet if possible. This allows for the least rotational acceleration to be felt by the driver. Pitch should rotate around your belly button or there about to best eliminate its rotational accelerations.

Btw, motion intensity in mostly sensed by variations in the inner ear so having your pivots around your head practically eliminates the feeling of motion where you want it the most.

@BlazinH Thanks for your reply.

>What makes you think having the pivot points above the center of mass of the driver is going to be any different than having them below?
When the pivot is under the seat, the rotation of the rig creates a sense force in opposite direction and it then abruptly switches to the right direction as rig stops moving and your back fully rests on the seat. Regular joyrider (as well as 401) makes pivot point in such position that rotational acceleration doesn't per se creates sense of force. Taking that a step forward, the very acceleration already creates a sense force in right direction even before gravity kicks in. This same principle is used in Toyota Simulator, but since it actually uses linear accelerations it gives infinitely more accurate representation of car G-forces.

>This allows for the least rotational acceleration to be felt by the driver.
Sorry but that needs heavy correction. Rotational acceleration depends only on actual angular acceleration of the rig, and it absolutely does not matter exactly where the driver is located. The driver's position relatively to the pivot point impacts his sense of tangential acceleration, and as the speed increases it also impacts sense of centripetal acceleration. By putting driver right on the pivot you can eliminate (un)desired tangential forces during angular acceleration, but rotational acceleration is still there. However, if the driver's belly is at the pivot point, lower body will experience tangential forces in direction of acceleration, and head will experience tangential forces in opposite direction. I'm trying to eliminate this by putting the whole body below pivot point, but seeing as it cranks up power requirement I'm considering putting pivot point closer to head level. This would diminish vestibular sense of acceleration from rig motion, only leaving gravitational component, but it reduces power requirement and also eliminates tangential acceleration vector direction mismatch between lower body and the head (direction discrepancy will actually increase but magnitude will drop a lot). In this configuration however, sense of force from acceleration is less evenly distributed between different parts of the body, most of it is at the bottom. Legs indeed will receive more thrust than the rest of the body, but that's a side effect of putting centre of masses directly underneath pivot point.

>having your pivots around your head practically eliminates the feeling of motion where you want it the most
Yes that's true. This rig primairly designed to give "butt feel" of the car, and given it's specifics, that is rather literal wording. However it's still worth mentioning that the other options include having vestibular sense of acceleration from rig acceleraiton going in complete opposite direction, so having it nullified is not exactly wrong by comparison. The vestibular sense of continious acceleration will still kick when the rig would reach the right angle, which is what pretty much all rigs that try to simulate continious forces are trying to do.