<![CDATA[Hexapod robots offer good mobility on uneven terrain but are susceptible to instability due to uneven load distribution among the legs when the body becomes tilted. This study evaluates a quaternion-based feedforward stabilization method that directly compensates for body posture using absolute orientation data, without rule-based inference or iterative error convergence. A BNO08x IMU provides real-time orientation quaternions, while touch sensors at the end of each leg are used to detect contact with the ground. Posture correction is calculated using quaternion conjugate rotation and applied to the body control point before inverse kinematics is performed. Experiments were conducted at platform tilt angles of 0°, 5°, 10°, 15°, 20°, and 25° using a 30-s tilt-and-recovery protocol with five replications at each angle. The stabilization residual ranged from 0.048–0.065° at tilt angles of 0–10° and was approximately 0.29° at 15°, before increasing to 2.09–5.69° at 20–25°. Joint-angle monitoring indicated that the increased residual at higher tilt angles was associated with femur–tibia workspace saturation rather than a failure of the stabilization algorithm. Following the recovery phase, the residual returned to 0.05–0.11° under all test conditions. An independent t-test showed no significant difference between pitch and roll up to 20° (p > 0.05), whereas a significant difference was observed at 25° (p = 0.021). The proposed method directly employs quaternion operations without fuzzification, rule-based inference, or defuzzification, providing a computationally simpler approach to hexapod balance stabilization under low-to-moderate tilt conditions.]]>
