In order to real-time adjust the humanoid robot gait, a humanoid robot gait generation method was proposed. The robot motion was simplified to the inverted pendulum motion of the three-dimensional linear model, and passed the pre-planned Zero Moment Point (ZMP) trajectory. According to the Center of Mass (CoM) and ZMP relations, the CoM trajectory was obtained. The front side gait and lateral gait were simplified into seven-link and five-link structures, the triangle theorem was used to calculate the angle of each joint, and the ZMP equation was introduced to discuss the stability of the walking process. The system simulation was done according to given conditions and combined with the actual system and its operation conditions were analyzed to verify the validity of the proposed planning method.

To study the problem of network threat assessment, a new method based on attribute recognition theory was proposed. This method extracted indicators of network attacking events in a period and modeled with entropy value to determine weights. Threat degree was quantified after the assessment and classification of attacking events in the network. The example's results indicate that this method is simple, practical and able to assess network threats effectively.