larger imagePhase II system prototyping began in February of 1995 with the above artist's concept of the system design. The sketch was prepared during the Phase I feasibility study.
The following sequence of photos shows the system under construction. The finished system will have safety shields installed around the ends and back of the frame, the cables will be neatly routed, and various other beautifications made.
The user sits facing a frame in which a payload of switches is moved horizontally and vertically. The horizontal and vertical drives are servomotors mounted at the base and rear on either side of the frame. Cables (nylon-covered Kevlar "strings") couple the motors to move the payload.
The user sees computer generated imagery of the cockpit in the headmounted display. The image appears in stereo and includes a generated image of the user's hand. (This being an Army project, the entire user model is painted green.) In operation, terrain and airport images appear out the cockpit windows.
This photo of the end of the frame shows one of the servo motors coupled to a drum which drives the motive cables. The frame is welded of two inch square stainless steel tube. 300 series stainless steel and other non-magnetic materials are used for the construction to minimize errors induced in the magnetic tracking.
This detail of the frame construction shows the installation of limit switches. Should a software error cause the payload to overshoot its intended range of travel, a hardwired switch causes the servo system to shut down. The motors are shorted to quickly halt travel when a limit switch or emergency stop is sensed.
The headmounted display, dataglove, and primary controls await a user. Note that the throttle and joystick are provided as real controls, but 3D models of them must be included in the visual database. The user sees nothing but imagery generated for the head-mounted display.
Magnetic trackers yield inherently noisy position measurements. The tracker data can be smoothed by filtering, but the time-averaging process results in lags between the tracker motion and the tracker output. We minimize lags by using accelerometer and angular rate sensor data along with the magnetic tracker data, resulting in effectively lag-free tracking. The inertial sensors are packaged in a plastic box about two and a half inches square, which is large compared to the magnetic tracker on top of the box. The inertial package is lightweight, however, and is mounted to the forearm rather than the hand.
Toggle switches and rotary controls must be returned to the positions of each control being simulated. Servomotors drive groups of four switches or rotary controls to put them into position. For toggle switches the entire switch is inverted as needed. Inside the payload, some of the stainless steel gears have been replaced by plastic versions as an experiment in saving weight.
A VME rack holds the servo controllers, servo amplifiers, power sequencing, and emergency stop hardware.
Return to the virtual cockpit project page.
