On February 20, 1996, at 1040 eastern standard time, an Advanced Technologies Ultra Sport 254, N254AS, collided with terrain during hovering maneuvers near West Point, Virginia. The certificated commercial pilot, the sole occupant, received serious injuries. The aircraft, an experimental helicopter, was destroyed. Visual meteorological conditions prevailed at the time of the accident, no flight plan was filed. The local test flight was conducted under 14 CFR 91, and originated from West Point, Virginia, at approximately 1020. A witness reported that the pilot was performing hovering maneuvers during a "...data gathering flight." The witness stated that the pilot performed pedal turns and sideward flight in both directions with no difficulty, and that the engine appeared to be running smoothly. According to the witness, the helicopter was hovering backward at a pace he described as "...a brisk walk...[then stopped its rearward flight and] pitched the nose down to transition to forward flight." The witness reported that the helicopter continued to pitch over into a progressively steeper nose down attitude, and descended into the terrain. According to the pilot, "The purpose of the engineering flight test was to obtain aircraft state information with respect to attitudes, rates, coupling, etc., in all four axis for FLAC-H software development and simulation verification." The pilot reported that as he "...was coming to a stable hover at approx. 20 ft AGL [above ground level] following the aft cyclic pulse, the aircraft developed an uncommanded pitch down of approx. 30 to 40 deg/sec." The pilot said he "...initially applied forward cyclic and positive collective in order to fly through translational lift and enter stable forward flight...I realized I had no control in the forward longitudinal axis...I impacted the ground approximately two seconds after the uncommanded pitch acceleration... ." The helicopter was equipped with the Fuzzy Logic Adaptive Controller-Helicopter (FLAC-H) system. The FLAC-H system is designed to allow for remote piloting of the helicopter upon which it is installed, or it can be used to gather flight data from the helicopter and transmit the data to a remote location. A flight test engineer stated that personnel monitoring the data stream from the helicopter "...confirmed that forward cyclic had been applied to initiate the nose-down pitch." The aircraft manufacturer provided investigators with a video tape of the accident, which was examined by the National Transportation Safety Board's engineering staff. The video showed the helicopter performing maneuvers as described by witnesses. According to the Safety Board's engineering staff, the video evidence did not indicate any failure or malfunction of the helicopter flight control system nor the helicopter's power train system. Additionally, the video showed flight control rates that are consistent with nominal helicopter control systems. There was no evidence that would reveal binding or stuck flight control systems. Personnel from the Safety Board's engineering division examined the flight control and power train systems at the aircraft manufacturer's headquarters. The examination revealed a broken bracket that was attached to the helicopter's main rotor hub. According to the Safety Board's engineering staff, the fracture was similar to overload failure. Video evidence indicated that the bracket was subject to ground impact forces during the crash sequence. There was no other damage to the flight control system. Continuity was established throughout the cyclic flight control system. There was no evidence of preimpact flight control malfunction or anomaly. The operator's written statement indicated that the pilot had 3,554 hours total flight time, of which 2,032 hours was in rotorcraft/helicopter. The written statement indicated that the pilot had 2.2 hours of flight experience in the accident make and model helicopter.